Neuromechanics of the lower limb - Implications for ACL injury ...

Neuromechanics of the Lower Limb: Implications for ACL Injury Prevention

Aaron S. Fox Bachelor of Exercise and Sports Science (Honours)

Submitted in fulfilment of the requirements for the degree of Doctor of Philosophy

Deakin University June, 2016

parisr

Redacted stamp

i

Acknowledgements

There are a number of people I wish to thank and acknowledge for their

contribution to the undertaking and completion of this thesis.

Special thanks must go to the coaches and players from clubs in the Geelong

Netball League (Geelong West St. Peters, Grovedale, Newtown and Chilwell,

and St. Mary’s) and Geelong District Netball League (Bannockburn, Bell Post

Hill, Geelong West and Thomson) who willingly gave their time to participate

in this research.

I would like to thank our laboratory technicians Mr. David Gleadon and Mr.

Ryan Pane for their technical assistance during data collection. The phrases

“Hey Dave / Ryan …” coupled with “X is broken” or “X isn’t working” became

consistent parts of my vocabulary over the last four years. Without your

assistance on a number of technical issues I would likely still be stuck

collecting data rather than acknowledging you here. I would also like to thank

Mr. Daniel Brown, Ms. Emily Gleeson and Ms. Alex Bauer for giving up their

time to help with data collection.

The time spent completing this thesis would not have been the same without

the group of fellow postgraduate students at Deakin. Even during stressful

times or setbacks, a simple chat over coffee or lunch was enough to reset and

keep going. Thank you for giving me a number of memorable moments over

the last four years.

I would like to express my sincerest gratitude towards my supervisors for their

expertise, guidance and support over the last four years. I have been lucky

enough to be surrounded by a supervisory team that has provided numerous

opportunities to continually learn and improve.

ii

To Dr. Scott McLean, thank you for the opportunity to visit and work with the

team in the Human Performance Laboratory at the University of Michigan.

The knowledge and skills I gained in the four months I spent working at

Michigan was vital in shaping this thesis.

To Dr. Jason Bonacci, your assistance during the last four years has been

invaluable. Thank you for sharing your knowledge and expertise in all things

biomechanics. Your persistence for perfection in a number of areas has made

me a better researcher, and for that I will always be grateful. Most of all, I will

never again write a cover letter without hearing your voice in the back of my

mind.

Lastly, to Dr. Natalie Saunders, had I not made the decision to work with you

on a simple Honours project years ago I would not be the same person I am

today. You have always encouraged me to meet challenges head on and given

me a sense of belief in what I can achieve. I am certain this thesis would not

have been the same without your supervision and guidance. I cannot thank

you enough for the investment you have made in me over the years.

iii

List of Publications

Fox AS, Bonacci J, McLean SG, Saunders N. Efficacy of ACL injury risk

screening methods in identifying high-risk landing patterns during a sport-

specific task. Scand J Med Sci Sports. doi: 10.1111/sms.12715.

Fox AS, Bonacci J, McLean SG, Spittle M, Saunders N. A systematic evaluation

of field-based screening methods for the assessment of anterior cruciate

ligament (ACL) injury risk. Sports Med. 2016; 46(5): 715-35.

Fox AS, Bonacci J, McLean SG, Spittle M, Saunders N. What is normal? Female

lower limb kinematic profiles during athletic tasks used to examine ACL

injury risk: A systematic review. Sports Med. 2014; 44(6): 815-32.

iv

List of Conference Presentations Presenter Underlined

Fox A, Bonacci J, McLean S, Saunders N. Exploring individual adaptations to

an anterior cruciate ligament injury prevention program. Paper accepted for

presentation at the 2016 Sports Medicine Australia Conference, 12th – 15th October

2015, Melbourne, Victoria, Australia.

Fox A, Bonacci J, McLean S, Spittle M, Saunders N. Detecting biomechanical

adaptations in sport-specific landing patterns using field-based ACL injury

risk screening methods. Paper presented at the 2015 Australian Conference of

Science and Medicine in Sport, 21st – 24th October 2015, Sanctuary Cove,

Queensland, Australia.

Fox A, Bonacci J, McLean S, Spittle M, Saunders N. The relationship between

performance on a single-leg squat and leap landing task: Moving towards a

netball-specific ACL injury risk screening method. Paper presented at the XXV

Congress of the International Society of Biomechanics, 12th – 16th July 2015,

Glasgow, Scotland.

Fox A, Bonacci J, McLean S, Spittle M, Saunders N. Assessing the effectiveness

of field-based ACL injury risk screening methods in identifying high-risk

landing patterns during a sport-specific task. Paper presented at the XXV

Congress of the International Society of Biomechanics, 12th – 16th July 2015,

Glasgow, Scotland.

v

Fox A, Bonacci J, McLean S, Spittle M, Saunders N. The influence of muscular

activation profiles on lower limb biomechanics during a sport-specific task.

Paper presented at the 33rd International Conference on Biomechanics in Sports, 29th

June– 3rd July 2015, Poitiers, France.

Fox A, Bonacci J, Spittle M, Saunders N. A systematic evaluation of field-based

screening tools for anterior cruciate ligament injury risk. Paper presented at the

Asics Conference of Science and Medicine in Sport, 22nd – 25th October 2013, Phuket,

Thailand.

vi

Preface

The major components of this thesis have been presented as a compilation of

studies arranged into chapters (Chapters Four to Nine). A number of the

methods used across these studies are similar. Detailed descriptions of the

methods used can be found in the Methodology Overview chapter (Chapter

Three). To improve the overall readability and avoid repeating redundant

information, the individual study chapters only include a brief description of

the experimental procedures, and data collection and analysis methods used.

Where these brief descriptions are given, readers will be referred back to the

relevant section of the Methodology Overview chapter where the detailed

description can be found.

Readers viewing the electronic version of this thesis may take advantage of the

hyperlinks provided throughout. Where a section, page, table or figure is

referred to, this reference will contain a link to the relevant page.

Across the study chapters, limitations that apply specifically to the present

study will be discussed at the end of the respective chapter. Conversely,

limitations that are relevant across the majority of studies will be discussed

within the General Discussion chapter (Chapter Ten).

vii

Table of Contents

Acknowledgements ................................................................................................... i

List of Publications ................................................................................................ iii

List of Conference Presentations ......................................................................... iv

Preface ....................................................................................................................... vi

List of Tables ............................................................................................................. x

List of Figures ........................................................................................................ xiv

List of Abbreviations and Acronyms ................................................................. xxi

Abstract ................................................................................................................. xxiv

Chapter One: Introduction ..................................................................................... 1

1.1 Purpose ............................................................................................................. 5

1.2 Specific Aims and Hypotheses ...................................................................... 5

1.3 Significance of Research ................................................................................. 8

Chapter Two: Review of Literature ....................................................................... 9

2.1 Anterior Cruciate Ligament (ACL) Injuries................................................. 9

2.2 ACL Injury Risk Factors ............................................................................... 14

2.3 ACL Injury Prevention Programs ............................................................... 22

2.4 ACL Injury Risk Screening ........................................................................... 26

2.5 Netball ............................................................................................................. 27

2.6 Summary ......................................................................................................... 30

Chapter Three: Methodology Overview ............................................................ 33

3.1 Ethical Approval ............................................................................................ 33

3.2 Participants ..................................................................................................... 33

3.3 Experimental Procedures ............................................................................. 35

3.4 Data Collection and Analysis ...................................................................... 53

3.5 Statistical Analysis ......................................................................................... 79

viii

Chapter Four: Intra-Individual Anomalies in Muscle Activation during a

Landing Task ........................................................................................................... 89

4.1 Introduction .................................................................................................... 89

4.2 Methodology .................................................................................................. 91

4.3 Results ............................................................................................................. 96

4.4 Discussion ..................................................................................................... 103

4.5 Conclusions .................................................................................................. 114

Chapter Five: The Influence of Muscle Activation Profiles on Lower Limb

Biomechanics during a Landing Task .............................................................. 115

5.1 Introduction .................................................................................................. 115

5.2 Methodology ................................................................................................ 117

5.3 Results ........................................................................................................... 123

5.4 Discussion ..................................................................................................... 149

5.5 Conclusions .................................................................................................. 159

Chapter Six: Exploring Individual Adaptations to an Anterior Cruciate

Ligament Injury Prevention Program ............................................................... 160

6.1 Introduction .................................................................................................. 160

6.2 Methodology ................................................................................................ 163

6.3 Results ........................................................................................................... 173

6.4 Discussion ..................................................................................................... 199

6.5 Conclusions .................................................................................................. 214

Chapter Seven: A Systematic Evaluation of Field-Based Screening Methods

for the Assessment of Anterior Cruciate Ligament Injury Risk ................. 215

7.1 Introduction .................................................................................................. 215

7.2 Methodology ................................................................................................ 218

7.3 Results ........................................................................................................... 223

7.4 Discussion ..................................................................................................... 239

7.5 Conclusions .................................................................................................. 258

ix

Chapter Eight: Efficacy of Field-Based Anterior Cruciate Ligament Injury

Risk Screening Methods in Identifying High-Risk Landing Postures during

a Sport-Specific Task ........................................................................................... 260

8.1 Introduction .................................................................................................. 260

8.2 Methodology ................................................................................................ 263

8.3 Results ........................................................................................................... 273

8.4 Discussion ..................................................................................................... 282

8.5 Conclusions .................................................................................................. 290

Chapter Nine: The Relationship between Performance of a Single-Leg Squat

and Leap Landing Task: Moving Towards a Netball-Specific Anterior

Cruciate Ligament Injury Risk Screening Method ........................................ 291

9.1 Introduction .................................................................................................. 291

9.2 Methodology ................................................................................................ 295

9.3 Results ........................................................................................................... 300

9.4 Discussion ..................................................................................................... 307

9.5 Conclusions .................................................................................................. 313

Chapter Ten: General Discussion ..................................................................... 314

10.1 Summary ..................................................................................................... 314

10.2 Neuromuscular Contributions to ACL Injury Risk .............................. 317

10.3 ACL Injury Prevention Programs ........................................................... 318

10.4 Field-Based Screening Methods for ACL Injury Risk .......................... 320

10.5 Limitations .................................................................................................. 321

10.6 Conclusions ................................................................................................ 324

References .............................................................................................................. 326

Appendices ............................................................................................................ 368

x

List of Tables

Table 2.1 Internal knee moments produced by the major muscles

surrounding the knee joint at flexion angles between 20 and 50

degrees. ............................................................................................... 17

Table 3.1 Guidelines used for surface electrode placement. ........................ 59

Table 4.1 Summary of the number of 'normal' and anomalous trials across

individual participants. .................................................................... 99

Table 5.1 Information criterion indices used for model selection in the latent

profile analysis. ................................................................................ 121

Table 5.2 Summary of muscle activation profiles employed by individual

participants across trials. ................................................................ 125

Table 5.3 Effect sizes for statistically significant overall differences in

activation between muscles within muscle synergy one. .......... 128


activation between muscles within muscle synergy two. ......... 129


activation between muscles within muscle synergy three. ....... 130

Table 5.6 Mean (± SD) muscle synergy weighing coefficient values for the

three muscle activation profiles identified. ................................. 140

Table 5.7 Summary of differences identified by SPM1D post-hoc analyses

for hip joint rotations between the three muscle activation

profiles. .............................................................................................. 143

xi


for knee joint rotations between the three muscle activation

profiles. .............................................................................................. 144


for hip joint moments between the three muscle activation

profiles. .............................................................................................. 147


for knee joint moments between the three muscle activation

profiles. .............................................................................................. 148

Table 6.1 Characteristics of participants allocated to control and training

groups. .............................................................................................. 164

Table 6.2 Reported attendance to the injury prevention program for

participants in the training group. ................................................ 174

Table 6.3 Description of principal patterns (PPs) extracted from hip and

knee joint rotation data from the leap landing task. .................. 178

Table 6.4 Description of principal patterns (PPs) extracted from hip and

knee joint moment data from the leap landing task. .................. 179

Table 6.5 Mean (± SD) muscle synergy weighing coefficient values from the

leap landing task for the control and training groups from

baseline to six-weeks. ...................................................................... 182

Table 6.6 Effect sizes (Cohen’s d) for individual changes of participants in

the training group from baseline to six-weeks for muscle

weighing coefficient values in muscle synergy one. .................. 189

xii



weighing coefficient values in muscle synergy two. .................. 190



weighing coefficient values in muscle synergy three. ............... 191


the training group from baseline to six-weeks for hip and knee

joint rotation principal pattern (PP) scores. ................................. 192


the training group from baseline to six-weeks for hip and knee

joint moment principal pattern (PP) scores. ................................ 193

Table 6.11 Standard deviations of individual change scores from baseline to

six-weeks for the control and training groups across muscle

synergies one, two and three. ........................................................ 194


six-weeks for the control and training groups for joint rotation

principal pattern scores. ................................................................. 195


six-weeks for the control and training groups for joint moment

principal pattern scores. ................................................................. 196

Table 6.14 Effect sizes (Cohen’s d) for changes of ‘high-’ and ‘low-risk’ sub-

groups within the control and training groups from baseline to

six-weeks for knee abduction/adduction moment principal

pattern (PP) scores. .......................................................................... 198

xiii

Table 7.1 Considerations and questions for evaluating identified screening

methods. ............................................................................................ 222

Table 7.2 Search terms used and search results of the secondary literature

search. ................................................................................................ 224

Table 7.3 Extracted data for the identified field-based screening

methods. ............................................................................................ 234

Table 8.1 Operational definitions for Landing Error Scoring System

items. ................................................................................................. 268

Table 8.2 Operational definitions for Tuck Jump Assessment items. ....... 269

Table 8.3 Characteristics of participants allocated to ‘high-’ and ‘low-risk’

groups based on Landing Error Scoring System (LESS)

score. .................................................................................................. 275

Table 8.4 Characteristics of participants distributed to ‘high-’ and ‘low-risk’

groups based on Tuck Jump Assessment score. ......................... 279

Table 9.1 Descriptions of principal patterns (PPs) extracted from single-leg

squat hip, knee and ankle joint rotation data. ............................. 302

Table 9.2 Descriptions of principal patterns (PPs) extracted from leap

landing hip and knee joint rotation data. ..................................... 303

Table 9.3 Descriptions of principal patterns (PPs) extracted from leap

landing hip and knee joint moment data. .................................... 304

Table 9.4 Summary of correlations between single-leg squat joint rotation

and leap landing joint rotation principal patterns (PPs). .......... 305

Table 9.5 Summary of correlations between single-leg squat joint rotation

and leap landing joint moment principal patterns (PPs). .......... 306

xiv

List of Figures

Figure 3.1 Diagrammatic representation of the experimental flow of

recruited participants. ..................................................................... 35

Figure 3.2 Sagittal view of the leap landing task. .......................................... 36

Figure 3.3 Diagrammatic representation of the laboratory set-up and

instrumentation used for the leap landing task. ......................... 39

Figure 3.4 Sagittal view of the bilateral drop vertical jump-landing

task. .................................................................................................... 40


instrumentation used for the drop vertical jump-landing task. 41

Figure 3.6 Sagittal view of the tuck jump task. .............................................. 42


instrumentation used for the tuck jump task. ............................. 43

Figure 3.8 Starting position for the single-leg squat task. ............................ 43

Figure 3.9 Sagittal view of the single-leg squat task. .................................... 44


instrumentation used for the single-leg squat task. .................... 45

Figure 3.11 Visual feedback guidelines for safe and effective landing

technique used within the ‘Down 2 Earth’ injury prevention

program. ............................................................................................ 47

Figure 3.12 Variations in stationary jump-landing exercises used within the

‘Down 2 Earth’ injury prevention program. ................................ 49

Figure 3.13 Variations in directional jump-landing exercises used within the


xv

Figure 3.14 Variations in lunging exercises used within the ‘Down 2 Earth’

injury prevention program. ............................................................ 51

Figure 3.15 Variations in sport-specific landing exercises used within the


Figure 3.16 Image of Delsys Trigno parallel-bar surface electrode with

specially tailored adhesive skin interface attached. .................... 58

Figure 3.17 Locations of EMG surface electrodes for an example

participant. ........................................................................................ 58

Figure 3.18 Example processing framework for EMG data. .......................... 62

Figure 3.19 Illustration of muscle synergies. .................................................... 64

Figure 3.20 Marker locations and the stationary calibration pose used to

define a kinematic model comprised of seven skeletal segments

(pelvis; right and left thigh; right and left shank; and right and

left feet). ............................................................................................. 69

Figure 3.21 Illustration of sign conventions used to represent hip, knee and

ankle joint rotations about three rotational degrees of

freedom. ............................................................................................ 73

Figure 3.22 Illustration of the DBSCAN cluster analysis algorithm using a

minimum points value of one. ....................................................... 82

Figure 3.23 Examples of SPM curves. ................................................................ 86

Figure 4.1 Sagittal view of the leap landing task. .......................................... 92

Figure 4.2 A) Mean total variance accounted for ( ) relative to the

number of extracted muscle synergies. B) relative to

the number of extracted muscle synergies. .................................. 97

xvi

Figure 4.3 Normalised activation coefficients of the three muscle synergies

extracted from the factorisation of EMG data. ............................ 98

Figure 4.4 Mean and standard deviation of approach speeds across 'normal'

and anomalous trials. .................................................................... 100

Figure 4.5 Mean hip and knee joint rotations of ‘normal’ and anomalous

trials from initial contact (IC) to 150 ms post IC during the leap

landing task. ................................................................................... 101

Figure 4.6 Mean hip and knee joint moments of ‘normal’ and anomalous


landing task. ................................................................................... 102

Figure 4.7 Individual participant example of muscle synergy weighing

coefficients from ‘normal’ and anomalous trials. ...................... 108

Figure 4.8 Individual participant example of muscle activation waveform

data from ‘normal’ and anomalous trials. .................................. 109

Figure 5.1 Sagittal view of the leap landing task. ........................................ 118


extracted from the factorisation of EMG data. .......................... 124

Figure 5.3 Mean electromyography (EMG) waveform data of the three

muscle activation profiles from 150 milliseconds (ms) prior to

initial contact (IC) through to 300ms post IC during the leap

landing task. ................................................................................... 126

Figure 5.4 Muscle synergy weighing coefficient values for MAP-1. ........ 131



xvii

Figure 5.7 Muscle synergy weighing coefficient values for the three muscle

activation profiles identified. ....................................................... 139

Figure 5.8 Mean and standard deviation of approach speed for the three

muscle activation profiles. ............................................................ 141

Figure 5.9 Mean hip and knee joint rotations of the three muscle activation

profiles from initial contact (IC) through to 150 milliseconds (ms)

post IC during the leap landing task........................................... 142

Figure 5.10 Mean hip and knee joint moments of the three muscle activation

profiles from initial contact (IC) through to 150 milliseconds (ms)

post IC during the leap landing task........................................... 146




program. .......................................................................................... 166


number of extracted muscle synergies. B) relative to

the number of extracted muscle synergies. ................................ 175


extracted from the factorisation of EMG data. .......................... 176

Figure 6.5 Mean (± SD) muscle synergy weighing coefficient values from

the leap landing task for the control and training groups from

baseline to six-weeks. .................................................................... 181

Figure 6.6 Mean EMG waveform data from 150 ms prior through to 300 ms

post initial contact (IC) during the leap landing task for the

control and training groups at baseline and six-weeks. ........... 183

xviii

Figure 6.7 Mean hip and knee joint rotations from initial contact (IC)

through to 150 ms post IC during the leap landing task for the


Figure 6.8 Mean hip and knee joint moments from initial contact (IC)



Figure 6.9 Individual changes of participants in the training group from

baseline to six-weeks for gastrocnemius weighing coefficient

values from muscle synergy one. ................................................ 203


baseline to six-weeks for medial hamstrings weighing coefficient

values from muscle synergy three. ............................................. 205


baseline to six-weeks for hip internal (IR) / external (ER) rotation

principal pattern (PP) two PP-Scores. ......................................... 207

Figure 6.12 Mean knee abduction / adduction moment from initial contact

(IC) through to 150 ms post IC during the leap landing task for

the ‘high-’ and ‘low-risk’ sub-groups from the training group at

baseline and six-weeks. ................................................................. 209

Figure 7.1 Diagrammatic representation of search strategy results and

study selection process. ................................................................. 226

Figure 7.2 Demonstration of the jump-landing task used for the Landing

Error Scoring System. .................................................................... 227

Figure 7.3 A) Specifically developed nomogram sheet for use with the

Clinic-Based Algorithm. B) Completed example of nomogram to

predict probability of high knee loads. ....................................... 230

xix

Figure 7.4 Example of trials deemed as “high-risk” (left) and “low-risk” as

per the Observational Screening of Dynamic Knee Valgus

(OSDKV) method guidelines. ...................................................... 231

Figure 7.5 Demonstration of the jump-landing task used for the Tuck Jump

Assessment. .................................................................................... 232

Figure 8.1 Sagittal view of bilateral drop vertical jump-landing task. ..... 265

Figure 8.2 Sagittal view of the tuck jump task. ............................................ 265


Figure 8.4 Histogram illustrating the distribution of Landing Error Scoring

System (LESS) scores across the participants examined. ......... 273

Figure 8.5 Relationship between Landing Error Scoring System (LESS)

score and knee flexion during the leap landing task. ............... 274


through to 150 milliseconds (ms) post IC during the leap landing

task of participants allocated as ‘high-’ and ‘low-risk’ based on

Landing Error Scoring System (LESS) score. ............................. 276

Figure 8.7 Mean normalised hip and knee joint moments from initial

contact (IC) through to 150 milliseconds (ms) post IC during the

leap landing task of participants allocated as ‘high-‘ and ‘low-

risk’ based on Landing Error Scoring System (LESS) score. ... 277

Figure 8.8 Histogram illustrating the distribution of Tuck Jump Assessment

scores across the participants examined..................................... 278

xx


through to 150 milliseconds (ms) post IC during the leap landing

task of participants allocated as ‘high-’ and ‘low-risk’ based on

Tuck Jump Assessment (TJA) score. ........................................... 280


contact (IC) through to 150 milliseconds (ms) post IC during the

leap landing task of participants allocated as ‘high-’ and ‘low-

risk’ based on Tuck Jump Assessment (TJA) score. .................. 281

Figure 8.11 Histogram illustrating the distribution of values for A) knee

flexion at initial contact; B) peak knee flexion; C) knee abduction

at initial contact; D) peak knee abduction; E) peak knee

abduction moment; and E) peak knee internal rotation moment

during the leap landing task. ....................................................... 286

Figure 9.1 Sagittal view of the single-leg squat task. .................................. 296


xxi

List of Abbreviations and Acronyms

1D One-dimensional

2D Two-dimensional

3D Three-dimensional

6W Six-weeks

% MVC Percentage of maximal voluntary contraction

ABD Abduction

ACL Anterior cruciate ligament

ADD Adduction

AIC Akaike information criterion

ANOVA Analysis of variance

ATS Anterior tibial shear

BIC Bayesian information criterion

BL Baseline

CCA Canonical correlations analysis

d Cohen’s d effect size

D2E Down 2 Earth

DBSCAN Density-based spatial clustering of applications with noise

DoF Degrees of freedom

EMG Electromyography

ER External rotation

EXT Extension

FLEX Flexion

xxii

FMS Functional Movement Screen

FPPA Frontal plane projection angle

GAS Gastrocnemius

GMAX Gluteus maximus

GMED Gluteus medius

GRF Ground reaction force

HSD Honest Significant Difference test

IC Initial contact

ICC Intra-class correlation coefficient

iLESS i-Landing Error Scoring System

IR Internal Rotation

JCS Joint coordinate system

κ Kappa statistic

LESS Landing Error Scoring System

LESS-RT Landing Error Scoring System Real-Time

LHAM Lateral hamstrings

LPA Latent profile analysis

MANOVA Multivariate analysis of variance

MAP Muscle activation profile

MHAM Medial hamstrings

MVICs Maximal voluntary isometric contractions

η2 Eta-squared effect size

NMF Non-negative matrix factorisation

NPV Negative predictive value

xxiii

OA Osteoarthritis

OSDKV Observational Screening of Dynamic Knee Valgus

PCA Principal component analysis

PEA Percentage of exact agreement

pKAM Peak knee abduction moment

PPV Positive predictive value

QuadHam Quadriceps-to-hamstrings strength ratio

r Pearson’s correlation coefficient

RF Rectus femoris

RMSE Root mean square error

ROM Range of motion

SEM Standard error of the mean

SPM Statistical Parametric Mapping

SPM1D One-dimensional Statistical Parametric Mapping

VAF Variance accounted for

VL Vastus lateralis

VM Vastus medialis

xxiv

Abstract

Injuries to the anterior cruciate ligament (ACL) occur most frequently during

sports which require the performance of high-risk tasks, such as landing and

side-step cutting. These injuries occur at a disproportionate rate in females

compared to males. ACL injuries are associated with a number of adverse

short- and long-term consequences. Understanding the underlying

mechanisms and modifiable risk factors surrounding ACL injury, with the

ultimate objective to promote prevention, is imperative. The overall purpose

of this thesis was to explore three inter-related areas relating to ACL injury

prevention in a cohort of sub-elite female netball players. Neuromuscular

contributions to ACL injury risk during the performance of a high-risk landing

task, the effect of an ACL injury prevention program on individual

neuromuscular and biomechanical adaptations, and the efficacy of field-based

ACL injury risk screening methods were explored.

Poor or inadequate neuromuscular control (i.e. the interaction between the

neural and muscle systems which coordinate and control movement) is often

considered a primary risk factor for ACL injury. Neuromuscular control can

be modified via training, hence an improved understanding of neuromuscular

contributions to ACL injury risk can assist in the design of targeted injury

prevention programs. The first two studies of this thesis focused on this area.

Electromyography and biomechanical data from the lower limb were collected

during repeat performances of a high-risk landing task. The findings from

these studies further the notion that neuromuscular control can vary from one

performance to the next. Anomalies in neuromuscular control (i.e. deviations

from an individual’s typical or usual muscle activation profile) were found

across repeat trials of the high-risk landing task, while the use of different

Abstract

xxv

muscle activation profiles across trials was also identified. In general, the

presence of neuromuscular anomalies did not influence lower limb

biomechanics during landing. Specific activation characteristics within each of

the profiles identified were linked to biomechanical deficits associated with

ACL loading and injury risk. Tailored injury prevention programs targeting

these deficits in muscle activation may induce more beneficial neuromuscular

adaptations for reducing ACL injury risk.

Variable neuromuscular and biomechanical adaptations have been shown to

occur across individuals subsequent to neuromuscular training. Despite this,

individual responses to ACL injury prevention programs have received little

attention. The third study of this thesis focused on the individual

neuromuscular and biomechanical responses to an ACL injury prevention

program. Electromyography and biomechanical data from the lower limb

were collected during repeat performances of a high-risk landing task at two

testing sessions, separated by a six-week period. During the six-week period,

a control group continued their regular training practices while a training

group undertook a six-week injury prevention program. The training program

induced a number of neuromuscular and biomechanical adaptations during

the high-risk landing task. However, examination of individual data revealed

dissimilar and inconsistent responses to the training. Despite the inconsistent

responses, individuals demonstrating a pre-existing high-risk movement

strategy (large peak knee abduction moments) demonstrated the largest

changes in frontal plane knee moments during landing. This likely stemmed

from the injury prevention program specifically targeting movement

strategies associated with frontal plane knee moments. These findings

provided further evidence that tailored injury prevention programs targeting

Abstract

xxvi

high-risk neuromuscular or biomechanical strategies may be the most effective

in reducing ACL injury risk.

The ability to identify athletes at-risk of ACL injury appears to be an important

step in maximising the effectiveness of targeted ACL injury prevention

programs. Screening methods that can be completed in field or clinical settings

offer the most viable option for ACL injury risk screening in the wider

community. The final three studies of this thesis focused on identifying the

optimal method(s) for field-based evaluation of ACL injury risk. A systematic

review conducted as part of this thesis found minimal evidence supporting

the use of field-based ACL injury risk screening methods for predicting future

ACL injuries. In addition, it was noted that the ability of field-based screening

methods to detect high-risk lower limb biomechanics during sport-specific

tasks known to cause ACL injuries was yet to be explored. The following study

subsequently explored the efficacy of two field-based ACL injury risk

screening methods (the Landing Error Scoring System and Tuck Jump

Assessment) in identifying lower limb biomechanics consistent with increased

ACL loading or injury risk during a high-risk sport-specific landing task.

Athletes were screened using the two aforementioned methods, while lower

limb biomechanical data were collected during the high-risk sport-specific

landing task. Minimal relationships were evident between scores obtained

from the screening methods and lower limb biomechanics from the sport-

specific landing task. These findings suggested that the screening methods

examined may not be effective in identifying ACL injury risk in the sport-

specific context (i.e. netball) examined. It was proposed that the bilateral

nature of the athletic tasks used within the screening methods contributed to

these findings. Therefore, the final study of this thesis explored the

relationship between performance of a unilateral movement and a high-risk

Abstract

xxvii

netball-specific landing task. For this study, a single-leg squat was used as the

unilateral screening movement in order to examine its utility within netball-

specific ACL injury risk screening protocols. Lower limb biomechanical data

were collected during performance of a single-leg squat and netball-specific

landing task (leap landing). Performance during the single-leg squat was

linked to a number of biomechanical factors proposed to increase ACL loads

or injury risk during the netball-specific landing task. These findings

highlighted the potential of a single-leg squat to be incorporated within field-

based netball-specific ACL injury risk screening protocols. Identifying

movements that have better associations to sport-specific tasks associated with

ACL injuries may further aid in the development of appropriate ACL injury

risk screening methods across different sports and sporting populations.

The experimental and review studies undertaken as part of this thesis have

produced novel findings that can contribute to ACL injury prevention

practices. A number of findings across this thesis highlight the need for

tailored ACL injury prevention programs targeting neuromuscular and

biomechanical deficits that are relevant to the individual. Identifying

screening methods that are capable of detecting these deficits during sporting

tasks known to cause ACL injuries is a vital step in improving the design of

targeted ACL injury prevention programs. The single-leg squat was identified

as a potentially useful movement to be incorporated within netball-specific

ACL injury risk screening protocols. Further examination of simplistic

unilateral movements against other sport-specific tasks associated with ACL

injury may aid in ascertaining the optimal method(s) for ACL injury risk

screening across different sports or sporting populations.

1

Chapter One: Introduction

Anterior cruciate ligament (ACL) injury is a traumatic sports-related injury

that often occurs in the absence of contact with another player (i.e. noncontact).

Injuries to the ACL occur more frequently in sports which require frequent

high-risk sporting tasks, such as landing and cutting movements.[1] Further,

adolescent (e.g. high school) and adult females are frequently shown to incur

a higher rate of ACL injury than male counterparts.[2-7] A number of short- and

long-term consequences are associated with ACL injuries, including surgical

intervention, lengthy absence from competitive sport, and reduced

participation in or premature retirement from high-risk sports.[8-11] Even more

concerning are the risks of re-injury[12-14] and osteoarthritis (OA)

development[15-17] subsequent to sustaining an ACL injury. Previous injury

dramatically increases the risk of future ACL injury in both the previously

injured and uninjured limb.[12-14] Early onset OA (athletes aged 30-50 years)

often presents in athletes who have sustained an injury to the ACL, with

associated pain, functional limitations and decreased quality of life.[16-19]

Understanding the underlying mechanisms and risk factors surrounding ACL

injury, with the ultimate objective to promote prevention, is imperative.

Extensive studies[20-29] have elucidated biomechanical mechanisms that load

the ACL and subsequently increase the potential for ACL injury during high-

risk sporting tasks.[30] Collectively, these works demonstrate that increases in

ACL loads can occur from biomechanical deficits in the sagittal, frontal and

transverse planes. However, multi-planar loading of the knee results in the

largest increases in ACL loads.[20-22, 31, 32] ACL injuries are therefore considered

to occur from altered biomechanics across multiple planes rather than

Introduction

2

biomechanical deficiencies in a singular plane alone.[33] In contrast, less is

known regarding the underlying neuromuscular contributions to ACL injury

risk or injury occurrences. The extensive knowledge of lower limb

biomechanics that contribute to increased ACL loading and injury provides a

useful framework to investigate neuromuscular control and ACL injury.

Neuromuscular control refers to the interaction between the neural and

muscle systems which coordinate and control movement. Detecting muscle

activation characteristics that result in high-risk lower limb biomechanics

during sporting movements can aid in identifying potential neuromuscular

risk factors for ACL injury. Given muscle activation characteristics can be

altered via neuromuscular training,[34-45] the identification of neuromuscular

risk factors provides a basis for targeted ACL injury prevention strategies.[15]

A perplexing factor associated with noncontact ACL injuries is the ‘one-off’

nature in which they occur, where injuries are often sustained during sporting

tasks athletes have replicated safely countless times.[46] Given the

neuromuscular system’s role in controlling lower limb motion,[15, 47] it is

conceivable that a random intrinsic dysfunction within the neuromuscular

system contributes to the phenomenon of ‘one-off’ ACL injuries. The inherent

variability of the neuromuscular system[48-53] suggests that anomalies in

neuromuscular control may occur. The presence of these neuromuscular

anomalies during a high-risk task could render an athlete at-risk of ACL

injury. The concept of neuromuscular anomalies and their relation to lower

limb biomechanics and ACL injury risk is yet to be explored. Further, little is

known about the combined muscle activation characteristics (i.e. the overall

muscle activation profile) that contribute to the presence of lower limb

biomechanics linked to increased ACL loading and injury risk. Previous

studies attempting to link neuromuscular factors to lower limb biomechanics

Introduction

3

associated with ACL loading have tended to focus on discrete muscle

activation characteristics of specific muscles.[54-64] Further understanding of the

link between neuromuscular control and movement strategies consistent with

increase ACL loads or injury risk may aid in identifying neuromuscular

targets for ACL injury prevention programs.

Current injury prevention practice has proven successful in reducing ACL

injury rates[65-67] and altering biomechanical and neuromuscular factors

associated with ACL injury.[36, 39, 40, 43, 68-80] Despite the repeated success of injury

prevention programs, overall ACL injury rates have remained relatively

unchanged over time.[1, 2] ACL injury prevention programs appear effective in

reducing injury risk in the short-term, but seem to lack effectiveness in the

long-term.[81] Current consensus is that further research is required to enhance

our understanding and maximise the effectiveness of ACL injury prevention

practices.[30] While a plethora of studies have examined neuromuscular and

biomechanical responses to ACL injury prevention programs, the effect of

these programs on individual adaptations is often overlooked. The success of

ACL injury prevention programs is generally defined by an overall change in

the training group, in isolation or relative to a control group, with little

attention paid to individual responses. Non-uniform neuromuscular and

biomechanical adaptations have been found across individuals in response to

the same training stimulus.[71, 82, 83] Failure to examine responses at the

individual level may mask beneficial or detrimental adaptations obtained by

certain individuals,[82] impacting our knowledge of effective injury prevention

practice. Understanding how different individuals respond to ACL injury

prevention programs will maximise our ability to implement interventions

that provide a benefit to all involved.

Introduction

4

Identifying athletes’ at-risk of ACL injury appears to be an important step in

maximising the effectiveness of ACL injury prevention programs. Female

athletes deemed as ‘high-risk’ have been found to receive a greater

prophylactic effect from neuromuscular training compared to ‘low-risk’

athletes.[84] Laboratory-based measures provide an accurate method to explore

biomechanical and neuromuscular deficiencies that may predispose an athlete

to ACL injury. This approach, however, requires extensive laboratory-based

equipment and training to complete, and is likely prohibitive to the majority

of the wider community. Field-based screening methods provide a more

accessible option for identifying dysfunctions that may lead to ACL injury.

Various field-based screening methods have been proposed for identifying

those at-risk of future ACL injury.[85-92] These methods vary in complexity from

the visual assessment of athletic tasks,[85, 86, 91, 92] to the use of a combination of

measures and algorithms[87, 88, 90] to predict ACL injury risk. While these

methods are designed for use by clinicians or the wider community, their

applicability for use in these settings is yet to be examined. The majority of

field-based screening methods show good agreement with laboratory-based

measures in identifying movement strategies that may contribute to increased

ACL injury risk.[87-91, 93, 94] However, generic bilateral athletic tasks (e.g. drop

vertical jump landings) have been used to validate these methods.[88, 90, 91, 93, 94]

The specificity of these movements to ACL injury risk could be considered

questionable. Previous work has shown generic athletic tasks do not

adequately represent those seen in a competitive sporting environment,[95, 96]

and nearly all noncontact ACL injuries occur during unilateral landing or

plant-and-cut manoeuvres.[97-100] In addition, ACL injuries appear to occur

during varying game situations and tasks across different sports.[5, 101-103]

Considering these factors, it is important that screening methods can identify

athletes’ who utilise high-risk movement strategies during unilateral sport-

Introduction

5

specific manoeuvres associated with ACL injuries. Examination of current

field-based screening methods against sport-specific movements known to

cause ACL injuries is therefore required. Further, there is evidence to suggest

that the predictive validity of screening methods varies across different

sporting populations.[104, 105] The development of sport-specific screening

methods may be an important consideration in improving ACL injury risk

screening protocols. Investigation of these areas will improve our

understanding of the optimal field-based methods for identifying athletes’ at-

risk of ACL injury in wider community settings.

1.1 Purpose

The overall purpose of this thesis was to explore three inter-related areas

relating to ACL injury prevention. First, the concepts of neuromuscular

anomalies and muscle activation profiles during the performance of a high-

risk sporting task were investigated. Second, the effect of an ACL injury

prevention program on individual neuromuscular and biomechanical

responses was explored. Finally, examination of existing field-based ACL

injury risk screening methods and a proposed sport-specific screening

movement was undertaken to determine their applicability for use in wider

community and sport-specific settings.

1.2 Specific Aims and Hypotheses

Specifically, the aims of each study included:

i. Determining whether intra-individual anomalies in muscle

activation profiles occurred during repeat performance of a high-risk

landing task, and if so, whether these neuromuscular anomalies

influenced lower limb biomechanics (Chapter Four).

Introduction

6

ii. Identifying the muscle activation profiles employed during a high-

risk landing task and their subsequent impact on lower limb

biomechanics linked to ACL loading or injury risk (Chapter Five).

iii. a. Investigating the effects of an ACL injury prevention program on

neuromuscular control and lower limb biomechanics during a

high-risk landing task at the group and individual level (Chapter

Six); and

b. Determining whether individuals’ deemed as ‘high-risk’ prior to

commencing training received a greater prophylactic benefit

from the ACL injury prevention program (Chapter Six).

iv. Systematically evaluating and comparing current field-based

screening methods for ACL injury risk to determine their efficacy of

use in wider community settings (Chapter Seven).

v. Examining the efficacy of two field-based screening methods for

ACL injury risk in identifying lower limb biomechanics consistent

with increased ACL loading or injury risk during a high-risk sport-

specific landing task (Chapter Eight).

vi. Determining the relationship between performance of a unilateral

task and a netball-specific landing task that has been associated with

ACL injuries, in order to examine the utility of including a single-leg

squat movement within netball-specific ACL injury risk screening

protocols (Chapter Nine).

It was hypothesised that:

i. Intra-individual anomalies in muscle activation profiles would

occur, and that anomalous activation profiles would result in lower

limb biomechanics consistent with increased ACL loads or injury

risk when compared to the remaining typical (i.e. ‘normal’) trials.

Introduction

7

ii. A number of different muscle activation profiles would be identified,

and that certain profiles would result in lower limb biomechanics

consistent with increased ACL loading and injury risk.

iii. a. The training group overall would benefit from the injury

prevention program as indicated by improvements in

neuromuscular and biomechanical risk factors associated with

ACL injury. However, variable responses to the program would

be observed across individuals within the training group; and

b. Individuals deemed as ‘high-risk’ for ACL injury would receive

a greater prophylactic benefit from the training.

iv. Current screening methods would include criteria appropriate for

identifying movement strategies associated with ACL injury risk,

however, would have a limited ability to predict future ACL injuries.

Some of the screening methods would be more applicable for use in

wider community settings than others due to having minimal

equipment and time requirements to complete.

v. Due to the generic bilateral nature of the screening movements,

scores obtained from the screening methods examined would have

no relationship to the lower limb biomechanics from the sport-

specific high-risk landing task.

vi. Lower limb biomechanical patterns from the unilateral task would

be associated with hip and knee biomechanical patterns from the

netball-specific landing task.

Introduction

8

1.3 Significance of Research

This research will further knowledge surrounding the neuromuscular

contributions to ACL injury risk by identifying the muscle activation

characteristics that contribute to the presence of lower limb biomechanics

known to increase ACL loads and injury risk. Identification of these

neuromuscular factors will aid in the design of ACL injury prevention

programs that target deficits in neuromuscular control. Currently, the effect of

ACL injury prevention programs at the individual level is not well

understood, particularly with regard to adaptations in muscle activation

characteristics. It is possible that athletes do not respond similarly to injury

prevention programs and analysis of group data may mask this.

Understanding how different individuals respond to an identical ACL injury

prevention program will enhance our ability to design programs that provide

a benefit for all involved. While laboratory-based measures provide the ability

to identify biomechanical and neuromuscular risk factors for ACL injury, these

are not as applicable or accessible to the wider community. It is important that

screening methods applicable for wider community use can identify athletes

who employ movement strategies that increase ACL loads or injury risk

during high-risk sporting tasks. Identifying optimal field-based methods for

detecting high-risk movement strategies during sporting tasks associated with

ACL injury will provide evidence for the use of screening methods in wider

community settings.

9

Chapter Two: Review of Literature

2.1 Anterior Cruciate Ligament (ACL) Injuries

2.1.1 Consequences of ACL Injury

Injury to the ACL results in adverse short- and long-term consequences.

Surgical repair, comprehensive rehabilitation, and a lengthy absence from

sport are common short-term consequences experienced by athletes.[16, 106, 107]

Proposed long-term consequences of an ACL injury include reduced

participation rates in high-risk sports and a high rate of early athletic

retirement.[11] More concerning, however, is the elevated risk of developing

early onset osteoarthritis (OA) following an ACL injury.[16, 17, 108, 109] While sports

participation in itself increases the risk of developing OA later in life, previous

joint injury further enhances this risk.[110] Injury to the ACL increases joint

instability and alters movement at the knee, placing the individual at risk for

development of OA.[111]

A 12-year follow-up of 67 ACL-injured female soccer players found that 82%

of players had developed osteoarthritic changes.[17] Similarly, Maletius and

Mesner[112] reported that 84% of patients presented with radiographic changes

equivalent to OA 20-years post ACL rupture. Recent reviews by Oiestad et

al.[113] and Claes et al.[19] suggest these percentages may overestimate the

prevalence of knee OA following ACL injury or reconstruction. Oiestad et

al.[113] reported a prevalence of 26% for patellofemoral osteoarthritis 12-years

after ACL injury. This value, however, may underestimate the prevalence of

OA as only patellofemoral OA was reported. The inclusion of tibiofemoral OA,

which has also been shown to present in ACL-injured individuals,[17, 114] may

Review of Literature

10

increase this prevalence. A meta-analysis including a total of 1554 subjects

found knee OA in 16.4% of subjects who underwent ACL reconstruction, and

increased to 50.4% in subjects who underwent combined ACL reconstruction

and meniscectomy.[19] These values, again, may underestimate the overall

prevalence of OA as only individuals treated surgically were included.

Conservative, non-surgical treatment is also an option following ACL

injury,[115] and those who opt for this treatment may also experience OA later

in life.[114] Whether or not past reported values have overestimated the

prevalence of OA following ACL injury, the development of knee OA has been

identified in athletes as young as 30 years of age, and by the age of 50 many

athletes require knee osteotomy or arthroplasty to repair the damage.[17] These

long-term consequences can lead to pain and functional limitations, resulting

in a decreased quality of life.[16-18, 116, 117]

An additional consequence surrounding ACL injuries is the associated risk of

re-injury.[14, 118] Paterno et al.[12-14] examined the incidence of ACL re-injury and

found the risk was 15 times greater in those with a history of ACL

reconstruction compared to matched controls. This effect appeared to be

magnified in females, with the risk of a second ACL injury approximately four

times greater for females compared to males.[14] The increased risk of re-injury

is not only limited to the initially affected limb. After an initial injury, the

contralateral limb appears to be at an even greater risk than the previously

injured limb.[13, 14] The risk of suffering a contralateral ACL injury has been

reported as being greater than that of sustaining a first time ACL injury.[12]

Primary prevention of injury is the most effective means for avoiding the

future disabilities associated with ACL injuries.[15] Given the adverse

consequences and high-risk of re-injury associated with injuries to the ACL,


11

there is a need to understand and develop effective injury prevention

strategies.

2.1.2 Sex Bias in ACL Injury Rates

A sex bias exists in the rate of noncontact ACL injuries, with adolescent (e.g.

high school) and adult females consistently experiencing this type of injury at

a higher rate than male counterparts.[1, 2, 4-7, 119] Despite the vast amount of

targeted research and injury prevention initiatives, ACL injury rates and the

associated sex disparity have not changed in recent times.[1, 2] Agel et al.[2]

found the rate of noncontact ACL injuries experienced by female collegiate

basketball and soccer athletes remained consistent over a 13-year period.

Similarly, Hootman et al.[1] found the yearly rate of ACL injuries to be

consistent across 16-years of collegiate injury data from 15 sports.

A number of studies have compared sexes during high-risk sporting tasks in

an attempt to understand why females are at greater risk of ACL injury.[59, 120-

135] Compared to males, females often exhibit different neuromuscular and

biomechanical strategies during cutting and landing tasks, including; greater

activation of the quadriceps relative to the hamstrings,[129, 130] greater lateral to

medial quadriceps activation,[131] reduced preparatory activation of the

hamstrings,[120] reduced muscle activity durations,[132] reduced hip and knee

flexion,[121, 133, 134, 136-138] greater knee abduction angles[122, 135, 137-148] and moments,[59,

122, 149] greater energy absorption in the frontal plane,[150] and greater peak tibial

rotations.[129, 151] While it is tempting to hypothesise that the sex-bias in ACL

injury risk could be diminished by shifting females towards male

neuromuscular and biomechanical strategies, the underlying assumption that

the male strategy is ‘ideal’ for protection against injury may be erroneous.[152]

Various structural differences exist between males and females, particularly


12

through the lower limb, and females may need to employ adapted

neuromuscular or biomechanical strategies to accommodate these structural

differences.[152] While altered neuromuscular control and movement strategies

in females may contribute to their increased risk of injury, it is plausible that

they represent a compensatory mechanism to accommodate for joint

mechanical variations.[152] Rather than drawing conclusions from sex

comparisons, future research should have an isolated focus on sex to identify

the sex-specific factors underpinning ACL injury risk. While both sexes

experience ACL injuries, an emphasis on females may be more imperative due

to their higher rate of ACL injury.[1, 2, 4-7, 119] Increasing our understanding of

sex-specific factors relating to ACL injury risk and prevention may aid in

bridging the gap in ACL injury rates between sexes.

2.1.3 Sporting Movements Associated with ACL Injury

Landing and/or performing a directional change have been identified as the

primary sporting movements associated with noncontact ACL injury.[5, 97-100, 102,

103] Visual observation of ACL injuries have identified side-step cutting or

landing following a jump as the predominant tasks being performed when

injuries occur.[5, 98-100, 102, 103] The rapid deceleration associated with these high-

risk tasks is thought to be a contributing factor to the injury.[55, 153] Taking this

into consideration, it is not surprising that the highest incidence of ACL

injuries are seen in sports that incorporate these movements, such as American

football, Australian Rules football, basketball, handball, and netball.

While the mechanism for noncontact ACL injuries appear similar across

sports,[98, 99] the tasks being performed in the lead up or at the injury event may

differ. Myklebust et al.[5] reported 19 of 23 noncontact ACL injuries observed

over three consecutive handball seasons occurred during faking or cutting


13

movements while in possession of the ball. In contrast, ACL injuries in netball

and basketball have been more frequently observed during landing rather

than cutting manoeuvres.[98, 103] Within Australian Rules football, a

combination of landing and cutting movements contribute to the majority of

noncontact ACL injuries.[102] The different tasks that result in ACL injuries

across sports likely share similar characteristics regarding the position and

joint loading of the lower limb, however, the goals of the athlete and lead up

to the task being performed are inherently different. Sport-specific variations

in athletic manoeuvres can have a modulating effect on lower limb

biomechanics.[154, 155] In addition, simplistic athletic tasks (such as bilateral drop

landings and drop vertical jumps) do not reflect the lower limb biomechanical

demands of sport-specific tasks.[95, 96] Considering these factors, examining

sport-specific tasks that contribute to ACL injuries appears pertinent to

increasing our understanding of ACL injury mechanisms and risk factors

across a range of sports.

2.1.4 The ‘One-Off’ Phenomenon

A perplexing factor associated with noncontact ACL injuries is the ‘one-off’

nature in which they occur, where injuries are sustained during sporting tasks

athletes have replicated safely countless times.[46] High-risk sporting tasks

associated with ACL injury are frequently performed during sports in which

ACL injuries are known to occur,[156, 157] yet ACL injuries occur at a much lower

rate. It is conceivable that a random intrinsic dysfunction is the inciting event

for these ‘one-off’ ACL injury occurrences. Considering the neuromuscular

system’s role in controlling the lower limb,[15, 47] anomalies in neuromuscular

control may be a contributing factor to these ‘one-off’ injury occurrences.

Understanding whether anomalies in neuromuscular control occur, and if so,

whether they influence lower limb biomechanics during high-risk sporting


14

tasks may yield valuable information regarding neuromuscular contributions

to ACL injury risk.

2.2 ACL Injury Risk Factors

There is consensus that ACL injuries are multifactorial in nature,[30, 153] with risk

factors generally categorised as non-modifiable (e.g. anatomical/structural,

hormonal, and genetic risk factors) or modifiable (skill-level, shoe-surface

interaction, biomechanical, and neuromuscular risk factors). Although ACL

injuries likely result from an interaction between multiple factors, research

should be directed towards those that have the potential to be modified.[15] The

following section will focus on the biomechanical and neuromuscular factors

associated with ACL injury risk, as these offer the most modifiable means for

injury prevention strategies.[15, 158]

2.2.1 Biomechanical Risk Factors

Injuries to the ACL occur when forces applied to the ligament are greater than

the loads it can withstand.[158] A number of biomechanical factors related to

lower limb motions and loads have been linked to increases in ACL loading

or noncontact ACL injury risk.[20-29]. ACL injuries have been observed to occur

with a forceful valgus collapse and tibial rotation with the knee in a relatively

extended position.[98, 99, 159] Experimental studies support these observations,

where knee valgus and internal rotation moments,[20-23] reduced knee flexion,[20,

24, 25] and large anterior tibial shear forces[26] have all been shown to increase

ACL loads.

‘Dynamic valgus’ and peak knee abduction moments during landing have

been identified as predictors of ACL injury in adolescent female athletes.[160] In

a prospective study of 205 athletes, those who went on to suffer an ACL injury


15

exhibited greater knee valgus angles and peak knee abduction moments

during a bilateral drop vertical jump.[160] Positions of ‘dynamic valgus’ stem

from knee abduction (i.e. abduction of the tibia relative to the femur) combined

with hip adduction and internal rotation,[160-162] all of which have been

associated with peak knee moments during athletic tasks.[149, 163, 164] The

contribution of the hip to dynamic valgus positions suggests proximal joint

biomechanics also play an important role in the ACL injury mechanism. A

more erect lower limb posture (i.e. reduced hip and knee flexion) reduces the

capacity to attenuate or absorb forces at the knee.[165-167] Where insufficient

flexion is present, the passive joint restraints (e.g. the ACL) are required to take

up a greater portion of counteracting and stabilising joint loads.[20] During

jump landings, peak ACL strain occurs when knee flexion angles are lowest,[25]

and cadaveric data and simulations show greater increases in ACL loads in

response to external moments when knee flexion angles are reduced.[20, 24]

Large increases in ACL strain have also been reported with internal tibial

rotation.[20, 21] Although external tibial rotation results in minimal increases in

ACL strain,[20, 21] rotation in this direction has also been observed during ACL

injury occurrences.[99] Although distal mechanics may also have the potential

to influence motion at the knee, there is minimal evidence[168, 169] linking specific

ankle motions to increased ACL loading or injury risk.

While isolated sagittal, frontal and transverse plane factors are suggested to

increase ACL injury risk, combined knee loading across multiple planes

results in the largest ACL loads.[20-22, 31, 32] Due to this, ACL injuries are thought

to occur via a multi-planar mechanism.[33] A focus on this multi-planar

mechanism is required for the development of optimal injury prevention

strategies.[32] It is also important to acknowledge that while certain

biomechanical risk factors have been implicated in ACL injury, they are not


16

the underlying causative factor. The neuromuscular system is fundamental in

controlling lower limb biomechanics,[15, 47] whereby the muscles provide

support against external loads[170-172] and contribute to joint stability during

dynamic tasks.[47, 173] Therefore, adequate muscle activation is essential in

counteracting potentially injurious loads being applied to the lower limb. The

presence of altered or poor neuromuscular control is likely to contribute to

lower limb biomechanics that increase ACL loads or injury risk during high-

risk sporting tasks. These resultant lower limb biomechanics should not be

considered as the underlying cause of ACL injury, but as a consequence of

neuromuscular dysfunction. Considering lower limb biomechanics are the

outcome of a given movement, they provide a means for which to link

neuromuscular dysfunction with potential increased ACL injury risk. Linking

neuromuscular dysfunctions to lower limb biomechanics known to increase

ACL loads or injury can aid in identifying neuromuscular contributions to

ACL injury and provide a practical target for injury prevention strategies.

2.2.2 Neuromuscular Risk Factors

As highlighted in the previous section, there is extensive knowledge

surrounding the movement strategies that contribute to ACL injury risk.

Considerable effort has also been applied to identifying the underlying

neuromuscular factors that contribute to the presence of potentially injurious

biomechanics during athletic tasks.[44, 56, 131, 134, 171, 174-181] Neuromuscular control

refers to the interaction between the neural and muscle systems which

coordinate and control movement. Altered or poor neuromuscular control of

the lower limb is recognised as a primary risk factor for ACL injury,

particularly in females.[15, 30, 47] The musculature surrounding the hip and knee

act to stabilise the lower extremity joints and provide support against external

loads.[171, 172] Based on their orientation and attachment sites, muscles in the


17

lower limb are capable of producing moments in certain directions as a means

to counteract external loads (see Table 2.1).[171] When external loads are applied

to the lower limb during high-risk sporting tasks, the central nervous system

must adjust muscle activation to oppose these forces.[171] Therefore, the

selective activation strategies of these muscles play an important role in

influencing ACL injury risk.

Table 2.1 Internal knee moments produced by the major muscles

surrounding the knee joint at flexion angles between 20 and 50 degrees. Adapted and reproduced with permission from Besier

et al.[171]

Adduction Abduction Int. Rotation Ext. Rotation

Flexion

Semimembranosus Semitendinosus

M. Gastrocnemius Gracilis

Sartorius

Biceps Femoris L. Gastrocnemius

Semimembranosus Semitendinosus

Gracilis Sartorius Lateral

Gastrocnemius

Biceps Femoris M. Gastrocnemius

Extension Vastus Medialis* Rectus Femoris*

Vastus Lateralis* Rectus Femoris* Vastus Medialis Vastus Lateralis

Int. – internal; Ext. – external; M. – medial; L. - lateral * - The quadriceps muscles are capable of producing both adduction and abduction moments due to

their common insertion into the patella tendon

Muscle activation strategies that counteract external knee loads and reduce

ligament loading have been identified.[170] Lloyd and Buchanan[170] found that

co-contraction of the quadriceps and hamstrings provided the majority of

support against varus/valgus isometric loads. Co-contraction of the

quadriceps and hamstring muscle groups can also reduce ACL strain during

knee flexion by resisting the anterior displacement of the tibia relative to the

femur.[182, 183] Due to their respective moment arms, medial and lateral knee

muscles are capable of counteracting knee abduction and internal rotation

moments.[171] Increases in activation of medial thigh muscles and the lateral


18

hamstrings have been observed during pre-planned side-step cutting to

counteract the large abduction and internal rotation moments, respectively,

experienced at the knee.[171] As the gastrocnemii are biarticular and span the

knee joint, selective activation of these muscles may also play a role in

protecting the knee and ACL.[184, 185] During unilateral landing simulations,

Morgan et al.[185] found elevated gastrocnemius forces were associated with

increased joint compression and reduced ACL forces. While a primary

function of the gastrocnemii is to produce plantarflexion, these findings

suggest its secondary function may be to co-contract with the quadriceps to

elevate joint compression and protect the ACL from external joint loading.[185]

The presence of selective activation strategies suggests the central nervous

system can alter muscle activation patterns as a means to protect the knee joint

from high external loads.[171]

Various elements of neuromuscular control have also been implicated in

increased ACL injury risk or linked to biomechanical factors associated with

ACL loading.[131, 174, 177, 178, 186-191] The preparatory activation strategies of lower

limb muscles appear important in modulating ACL injury risk, as dysfunction

in this area has been linked to ACL injuries[177, 186] and lower limb postures

associated with increased ACL loads.[174, 178] Individuals’ who demonstrate an

activation pattern of elevated preparatory activation of the vastus lateralis in

conjunction with reduced preparatory activation of the semitendinosus during

a side-step cutting manoeuvre are at increased risk of sustaining an ACL

injury.[177] Elevated preparatory activation of the lateral quadriceps and

hamstrings has also been linked to greater knee abduction angles during

landing.[187] Brown et al.[174] found during single-leg landings, greater

preparatory activation of the rectus femoris resulted in increased peak anterior

tibial shear force, while greater preparatory activation of the lateral hamstrings


19

was associated with a reduction in knee flexion angle. A muscle activation

strategy of greater vastus medialis and gluteus maximus preparatory

activation, in conjunction with reduced biceps femoris and gastrocnemius

preparatory activation has also been linked to smaller knee flexion angles

during landing.[178] Greater preparatory and peak activations of the quadriceps

have been associated with increases in anterior tibial translation and shear

forces.[26, 174, 188] However, increased quadriceps forces in the preparatory phase

of landing may increase joint stability and limit ACL strains during the impact

phase.[192] Quadriceps preparatory activation could therefore be vital in

enhancing knee joint stability during dynamic single-leg tasks, and reducing

quadriceps activation may be an erroneous training goal within injury

prevention programs.[174] Targeting the degree and timing of hamstring muscle

activation may be a more appropriate method for reducing anterior tibial shear

forces and ACL strain during landing. Co-contraction of the hamstrings with

the quadriceps can reduce the forces experienced in the ACL by limiting

anterior tibial translations relative to the femur across a range of knee flexion

angles commonly experience during cutting and landing tasks.[182, 183] Reduced

preparatory activation of the hamstrings may result in lower hamstrings

contraction forces during the initial portion of ground contact, reducing knee

stability.[120] Hamstring muscle recruitment closer to initial landing contact

may also allow peak muscle activity to better coincide with peak tibial shear

forces, thereby protecting the ACL.[55]

In addition to preparatory activation strategies, muscle activation during the

landing or cutting phases of athletic tasks have also been linked to high-risk

lower limb biomechanics.[131, 189-191] A lateral or medial imbalance of the vastii

can generate frontal plane knee moments, owing to their respective frontal

plane moment arms.[170, 171] Compared to males, females have been shown to


20

activate their vastus lateralis to a greater extent than their vastus medialis

during unanticipated cutting manoeuvres.[131] This activation strategy may be

a factor in the greater knee abduction angles and moments experienced by

females during cutting manoeuvres.[59, 131, 193] In addition; female athletes who

exhibit greater peak activity of the vastus lateralis, combined with a lower

medial to lateral quadriceps activity ratio during the landing phase of a drop

vertical jump have been found to exhibit reduced hip and knee flexion, and

higher peak knee abduction moments.[189] The proximal muscles of the lower

limb also play an important role in protecting the ACL during athletic tasks.

Hip muscle strength and neuromuscular control are cited as important factors

in limiting knee abduction angles and moments during cutting and landing

tasks.[57, 194-198] Lower hip abduction and external rotation isometric strength

have also been prospectively linked to an increased risk of ACL injury.[198]

These findings have relevant links to biomechanical factors associated with

ACL loading and injury risk. An inability to limit hip adduction and internal

rotation can result in a position of ‘dynamic valgus’ during landing.[162] The

gluteus maximus and medius play a role in controlling hip adduction and

internal rotation, therefore inadequate activation of these muscles may

contribute to a lower limb posture consistent with increased ACL injury

risk.[199] This is supported by experimental observations, where reduced levels

of gluteus maximus activation have been associated with greater knee

abduction during athletic tasks.[190, 191]

The combined activation strategies of lower limb muscles appears critical in

protecting the ACL from external loading during high-risk sporting tasks. The

muscles of the lower limb must therefore function in an appropriate manner

each time these tasks are performed to prevent ACL injuries from occurring.

High-risk tasks, such as landing and side-step cutting, are regularly performed


21

during sports in which ACL injuries are known to occur.[156, 157] Muscle

activation may not be consistent across all of these performances, as a number

of activation patterns can be used when performing the same task.[200]

Considering the role of the neuromuscular system in protecting the ACL,

variations in an individual’s muscle activation profile may leave them

susceptible to injury. It is possible that unexpected alterations or ‘anomalies’

in muscle activation may be a contributing factor to the ‘one-off’ nature in

which noncontact ACL injuries occur. The presence of anomalies in muscle

activation profiles and their subsequent impact on lower limb biomechanics

during high-risk sporting tasks is yet to be investigated. Detecting whether

these neuromuscular anomalies present within repeat trials of a high-risk

sporting task and their influence on lower limb biomechanics could reveal

important information surrounding neuromuscular contributions to ACL

injury risk. Should these anomalies in muscle activation occur and contribute

to the presence of lower limb biomechanics linked to increased ACL loads, this

would provide a target for neuromuscular training protocols within ACL

injury prevention programs.

Furthermore, the link between the overall muscle activation profiles employed

during high-risk sporting tasks and their relation to lower limb biomechanics

is not well understood. Previous studies that have linked neuromuscular

factors to lower limb biomechanics associated with ACL loading have tended

to focus on discrete muscle activation characteristics. These studies examine

selective temporal and/or amplitude features of muscle activation data; such

as muscle onset timing,[54-56, 58] the timing and/or amplitude of peak muscle

activity,[55, 59, 63, 64] and muscle burst duration;[54, 55] with these data points often

isolated to individual muscles or muscle groups. An isolated approach on

specific activation characteristics of individual muscles or muscle groups fails


22

to acknowledge the relationship between multiple muscles and their

combined activation characteristics. As there is no single muscle crossing the

knee capable of supporting the joint against injurious loads, activation

strategies comprising multiple muscles are required to reduce ACL injury risk

during athletic tasks.[161] Understanding the activation characteristics of

multiple muscles is therefore a key factor in preventing ACL injuries during

high-risk sporting tasks.[161] A comprehensive analysis of multiple lower limb

muscles will aid in identifying the specific muscle activation profiles

employed by athletes’ during high-risk sporting tasks. Concurrent

examination of how different activation profiles impact on lower limb

biomechanics known to increase ACL loads may provide valuable knowledge

regarding neuromuscular contributions to ACL injury risk. Multiple studies

have demonstrated the capacity of neuromuscular control to be modified via

neuromuscular training.[34-45] Therefore, an improved understanding of

neuromuscular contributions to ACL injury risk can assist in the design of

targeted and effective injury prevention programs.

2.3 ACL Injury Prevention Programs

Injury prevention programs aimed at reducing the incidence of ACL injuries

or modifying neuromuscular and biomechanical risk factors are prominent

throughout the literature.[5, 34-45, 65-69, 71-75, 79, 83, 166, 201-217] There is some evidence to

suggest that injury prevention programs can reduce ACL injury rates[5, 65-67] and

substantial support for the use of these programs in inducing neuromuscular

and biomechanical adaptations proposed to reduce ACL injury

risk.[34-45, 68, 69, 71-75, 79, 166, 201, 213-217] A recent meta-analysis[218] incorporating 14

studies[5, 65-67, 202-211] supported the use of neuromuscular training and

educational interventions, suggesting these programs have the capacity to


23

reduce ACL injuries in female athletes by 51%. However, the estimated effect

of programs largely varied across studies, with the authors unable to explain

the variability observed.[218] Less conclusive results were reported in a review

by Stevenson et al.[219], where only two[65, 206] of ten studies reviewed

demonstrated a statistically significant reduction in ACL injuries following a

neuromuscular training program. While certain studies showed a significant

decrease in injuries within certain subgroups[5, 204] or a trend towards reduced

ACL injuries following training,[67, 205, 207, 212] two studies[208, 209] found an increase

in ACL injuries within the groups receiving the prevention program.[219] While

injury prevention programs can have a beneficial effect on reducing ACL

injury rates, this effect appears to be inconsistent across the literature.

Athlete compliance,[209, 212, 220, 221] intervention exposure,[209, 210] and the specificity

of the training program[212, 222] are cited as factors that may impact on the

effectiveness of injury prevention programs in reducing ACL injury rates.

Compliance with training protocols appears to be a key factor in determining

the success of an ACL injury prevention program.[205-207, 221] Sugimoto et al.[223]

found a potential inverse dose-response relationship between athlete

compliance and program success, where attending and completing prescribed

training sessions was directly related to the reduction of ACL injuries in young

female athletes. Supervision of training sessions may also be a key factor in

determining a program’s success.[223, 224] Fortington et al.[225] found many

players participating in a coach-led injury prevention program did not

perform exercises as prescribed, and this may have an impact on the benefits

received from the training. Without supervision and an emphasis on proper

technique, limited or poor adaptations to training exercises becomes a

possibility and may result in no decrease or an increased rate of injury in

program participants.[223]


24

The inconsistent results surrounding injury prevention programs and ACL

injury rates highlights the need to examine changes in lower limb

biomechanics and neuromuscular control following training.[161] This

approach may allow us to ascertain the biomechanical or neuromuscular

mechanisms by which training influences ACL injury risk.[161] A number of

injury prevention programs incorporating plyometric, balance, resistance

and/or technique focused training have proven effective in increasing

hip[39, 213, 214] and knee flexion[39, 41, 68-70, 72, 74, 75, 79, 83, 213-215]; reducing hip

adduction,[71, 74, 80] hip internal rotation,[71] and knee abduction angles;[70, 72, 80, 214]

decreasing anterior tibial shear forces;[213] decreasing knee

abduction[37, 68, 69, 166, 216] and internal rotation[216, 217] moments; increasing

hamstrings strength;[72] increasing preparatory activation of the hip[36, 39] and

hamstrings[40, 41] muscles; increasing overall activation of the hip[39] and

hamstrings[40] muscles; and increasing hamstrings-to-quadriceps co-

contraction[69] during cutting and landing tasks. These results demonstrate

injury prevention programs can be effective in altering an individuals’

movement or neuromuscular strategies during high-risk landing or cutting

tasks.

Despite the apparent success of injury prevention programs in recent times,

ACL injury rates have remained relatively unchanged[1, 2] with the associated

sex bias remaining a concern in high-risk sports.[2-7] In addition, the effect of

injury prevention programs on ACL injury rates is inconsistent, with certain

studies finding no effect on injury rates with training.[208, 209] A limited

understanding of how different individuals respond to ACL injury prevention

programs may be a factor that is restricting our understanding of the

mechanisms underpinning success or failure. Success of ACL injury

prevention programs are often dictated by an overall improvement in the


25

training group, in isolation or relative to a control group, with little attention

paid to individual responses. However, it may be erroneous to expect that

everyone will respond in the same manner when given the same training

stimulus. There has been limited examination of individual responses to ACL

injury prevention programs. Of those studies that have considered the

individual, variations have been observed in the responses to training.[71, 82, 83]

Cowling et al.[82] found individuals displayed non-uniform adaptations in

quadriceps and hamstrings muscle activation characteristics during a sport-

specific landing task, despite completing identical training programs. No

differences were found from pre- to post-training when examining group data,

highlighting the potential of individual responses to be masked when

examining data in this manner. Similarly, Dempsey et al.[83] found no

differences in torso postures during a side-step cutting manoeuvre following

training, however examination of individual changes found certain

individuals had large favourable responses with regard to torso rotation. The

authors hypothesised that these individuals may have displayed a more at-

risk technique at baseline, and hence obtained a greater benefit from the

training.[83] In contrast to these two studies, Pollard et al.[71] found a beneficial

effect of training on frontal and transverse plane hip motion when performing

a jump-landing task. When examining individual data, however, certain

individuals showed greater improvements in one plane versus the other.[71] It

was suggested that individuals may have exhibited greater changes in the

plane in which they demonstrated a pre-existing at-risk pattern.[71]

Interestingly, detrimental adaptations following training were observed in

certain individuals suggesting that not all responded favourably to the

training program.[71] These studies highlight how examining group data in

isolation can vary our interpretation as to how individuals gain a benefit, or

lack thereof, from ACL injury prevention programs. Failure to consider


26

individual responses may therefore be impacting our current understanding

of what constitutes effective ACL injury prevention practice.

2.4 ACL Injury Risk Screening

Laboratory-based measures (e.g. motion analysis, force plates,

electromyography) provide an accurate method to explore segment motions,

joint kinetics and muscle activation patterns during sporting movements.

These tools provide an option to evaluate neuromuscular or biomechanical

strategies that may predispose an athlete to ACL injury. However, the use of

laboratory-based equipment for ACL injury risk screening comes with high

financial and time costs. Extensive training is required to operate this

equipment, and the cost of three-dimensional motion analysis has been

reported in the range of $1000 per athlete per test.[87] A recent simulation study

found incorporating current injury risk screening practices within ACL injury

prevention efforts is a less cost-effective strategy than universal

neuromuscular training.[226] However, potential barriers to implementing

universal neuromuscular training exist. Although ACL injury prevention

programs are often associated with successful outcomes, the level of adoption

and compliance is low,[227, 228] and their deployment may not be effective in

“real-world” situations.[229-231] Current injury prevention programs often

involve the whole team, require considerable time investment and may detract

from sport-specific skill training.[84] These factors may deter coaches from

implementing such programs within their pre-season or in-season training.[84,

229] Screening methods that identify high-risk athletes may improve coach

intent to implement prevention programs for these athletes.[84] In addition, the

ability to identify individuals ‘at-risk’ of ACL injury appears to be an

important step in maximising the effectiveness of injury prevention programs.


27

Myer et al.[84] found female athletes deemed as ‘high-risk’ (knee abduction

moment > 25.25 Nm during a drop vertical jump) received a greater

prophylactic effect from neuromuscular training compared to ‘low-risk’

athletes. Considering these factors, there is an important place for accessible,

cost-effective ACL injury risk screening protocols in ACL injury prevention

efforts.

Similar to injury prevention programs, injury risk screening also requires time

investment from coaches and athletes. However, there may be a role for

coupling screening with more targeted and intensive injury prevention

programs if the screening methods used have high accuracy with minimal

cost.[226] Screening methods that can be completed in a field or clinical setting

may meet these demands, and also be more applicable for wider community

use. Identification of screening methods that are predictive of injury, valid in

identifying high-risk movement strategies, while remaining cost-effective

with regard to equipment and time may provide further support for the

integration of screening within ACL injury prevention efforts. Field-based

screening methods for identifying ACL injury risk are emerging,[85-88, 90-92]

however assessment of their applicability for use in wider community and

sport-specific settings is yet to be undertaken. Evaluating the capacity of

current field-based ACL injury risk screening methods to be used within wider

community and sport-specific contexts will assist in identifying the most

appropriate method(s) to be used in these settings.

2.5 Netball

Netball is a sport played among Commonwealth countries, and has one of the

highest participation rates for team sports in Australia.[232] It is a sport that

covers all ages and skill levels, but is predominantly played by females.[232, 233]


28

Netball requires sudden changes of direction and rapid decelerations, in

combination with leaps to perform attacking or defensive manoeuvres.[234]

These manoeuvres place considerable loads on the lower limb, particularly at

the knee joint, with netball readily acknowledged as a high-risk sport for lower

limb and ACL injuries.[234-236] Netball provides a convenient cohort, particularly

in Australia, for research examining ACL injury risk and prevention strategies.

2.5.1 Injury and Netball

Injury surveillance data reveals lower limb injuries, particularly to the knee

and ankle, are the most commonly injured areas in netball.[237-243] Hopper et

al.[238] found a total of 608 injuries occurred during competitive netball

tournaments over a five-year period, with 92.6% of these injuries involving the

lower limb. Similarly, Otago and Peake[244] found that 85.3% of netball related

injuries from 12-months of insurance claims data were to the lower limb.

Netball has also gained a reputation for the occurrence of ACL injuries.[245]

While acknowledged as a problem injury within the sport, the rate of ACL

injury in netball is quite low. Only 1.8% (n = 11) of all injuries across the five-

year period examined by Hopper et al.[238] involved the ACL. A low-rate of

ACL injuries relative to the total number of injuries is not uncommon across

high-risk sports. In other female high-risk sports (such as basketball, lacrosse,

soccer and volleyball), Hootman et al.[1] found that the contribution of ACL

injuries to total injuries ranged from 2.0 – 5.0%. Despite the low relative rate

of ACL injuries in netball, these injuries tend to be the most severe experienced

by players and often require a lengthy rehabilitation process.[238] There is

currently no data available examining the long-term consequences specifically

for netball-related ACL injuries. However, it is likely that players suffer the

same debilitating conditions later in life as athletes’ from other sports.


29

2.5.2 High-Risk Sporting Tasks in Netball

Side-step cutting and landing are the primary sporting tasks observed during

noncontact ACL injuries.[5, 98, 99, 102, 103, 159] Due to the specific demands of netball,

players perform a high frequency of cutting and landing manoeuvres

throughout a match.[156, 157, 246] While both tasks are frequently performed,

landings are often implicated as the high-risk task for lower limb injuries

within the sport.[103, 245] Landing is commonly cited as the inciting event when

lower limb and ACL injuries occur in netball.[103, 238-240, 243] In a video analysis of

ACL injuries from elite female netball competition; 13 of the 16 ACL injuries

observed occurred during landing manoeuvres, while only one was observed

during a side-step cutting movement.[103] In particular, the ‘leap’ landing

technique employed by netball players is a known mechanism of ACL injuries

within the sport.[245] Although unplanned sporting tasks have been found to

incur lower limb biomechanics linked to greater ACL loading,[247] these do not

play a major role for ACL injuries within netball. Even within competitive

environments leap landings appear highly planned, often used as part of

attacking manoeuvres in order to receive passes after breaking free from

defenders.[157] During unplanned tasks, unexpected movements or

perturbations are proposed to contribute to the higher risk of ACL injury.[247]

In contrast, it is likely that the specific motions and lower limb postures

associated with leap landings contribute to a higher risk of ACL injuries

during this manoeuvre. A leap landing involves a run-up, followed by a single

limb take-off and land on the contralateral limb.[157, 245] In netball, players are

often required to decelerate abruptly upon landing. Deceleration with one step

and the leg near full extension are characteristic of leap landing performances,

and places a high load on the knee and ACL.[245] The leap landing technique

also results in distinct peak knee abduction moments early after initial landing


30

contact.[248] Laboratory analysis examining the risk of ACL injury in netball

should focus on this landing technique due to its relationship with movement

strategies known to increase ACL loads and capacity to cause injury within

the sport.

2.6 Summary

Noncontact ACL injuries are a traumatic sports related injury that occur most

frequently during landing and cutting tasks.[5, 98, 99, 102, 103, 159] A perplexing factor

associated with noncontact ACL injuries is the ‘one-off’ nature in which they

occur, where injuries are sustained during sporting tasks athletes have

replicated safely countless times.[46] A number of debilitating short- and long-

term consequences are associated with ACL injuries, including an increased

risk of early onset osteoarthritis and re-injury.[9, 12, 14, 16, 17, 109, 112] Primary

prevention of injury is the most effective means for avoiding these future

disabilities.[15] Therefore, there is a need to understand and develop effective

injury prevention strategies.

The biomechanical factors that contribute to increases in ACL load and injury

risk are well understood. It is accepted that ACL injuries occur through a

multi-planar mechanism[33] and a focus on this is required for the development

of optimal injury prevention strategies.[32] Altered or poor neuromuscular

control of the lower limb is recognised as a primary risk factor for ACL injury,

particularly in females.[15, 47] The selective activation strategies of lower limb

muscles play an important role in influencing ACL injury risk. It is possible

that ‘anomalies’ in muscle activation may be a contributing factor to the ‘one-

off’ nature in which noncontact ACL injuries occur. The overall muscle

activation profile employed during high-risk sporting tasks may also influence

lower limb biomechanics, and subsequently, ACL injury risk. Examining


31

whether neuromuscular anomalies and different muscle activation profiles

present during high-risk sporting tasks and their subsequent impact on lower

limb biomechanics could reveal important information surrounding

neuromuscular contributions to ACL injury risk.

There is some evidence to suggest that injury prevention programs can reduce

ACL injury rates[5, 65-67] and substantial support for the use of these programs

in inducing neuromuscular and biomechanical adaptations proposed to

reduce ACL injury risk.[34-45, 68, 69, 71-75, 79, 166, 201, 213-217] Despite this, ACL injury rates

have remained relatively unchanged in recent years.[1, 2] Few studies[71, 82, 83]

have explored the individual neuromuscular and biomechanical responses to

ACL injury prevention programs, and this may be limiting our capacity to

understand what constitutes effective ACL injury prevention practice.

Laboratory-based measures frequently used to identify athletes’ at-risk of ACL

injury are not cost-effective and prohibitive to the majority of the wider

community.[87] Accurate and cost-effective screening methods that can be

coupled with more targeted and intensive injury prevention programs are

likely to be beneficial in reducing the rate of ACL injuries.[226] Field-based

screening methods are a more cost-effective and applicable option for wider

community use. A number of ACL injury risk screening methods that can be

completed in field or clinical settings have been developed,[85-88, 91-93] however

assessment of their applicability for use in wider community and sport-

specific settings is yet to be undertaken. Evaluating the capacity of current

field-based ACL injury risk screening methods to be used within wider

community and sport-specific contexts will assist in identifying the most

appropriate method(s) to be used in these settings


32

In Australia, netball provides an ideal cohort to study noncontact ACL

injuries. Netball is predominantly played by females and involves the

performance of high-risk sporting tasks that are associated with ACL injuries.

In particular, the leap landing technique is often associated with ACL injuries

within the sport. Research focusing on this cohort and high-risk task can yield

valuable information relating to ACL injury risk and prevention.

33

Chapter Three: Methodology Overview

This chapter serves as an overview of the general methods employed within

this thesis. Details regarding the general methodology used to examine

neuromuscular control and lower limb biomechanics, and experimental

procedures that are common across studies are presented herein. Within

subsequent chapters, readers will be referred back to the appropriate sections

of this chapter where applicable.

3.1 Ethical Approval

Project approval was given by the Deakin University Human Research Ethics

Committee (Project No. 2013-007; see Appendix A, Page 369). Prior to testing,

all participants were informed of the research and potential hazards via a plain

language statement. Participants volunteered and provided written consent

prior to any involvement in the research or laboratory testing sessions (see

Appendix B, Page 370).

3.2 Participants

Sub-elite female netball players were recruited for all studies within this thesis.

The players included in this research were from a community-level open-age

female netball competition which teams play one match per week. In addition

to matches, players on average trained 1.74 ± 0.85 times per week (range = 1 –

4) for an average of 71.71 ± 14.14 minutes (range = 45 – 100). Players had on

average 12.40 ± 4.42 years of netball experience (range = 2 – 21). Prior to testing,

participants completed an injury history questionnaire (see Appendix C, Page

378) and were excluded if they had: (i) a current lower limb/back injury; (ii)

Methodology

34

previous history of serious knee injury or lower limb/back surgery; or (iii)

suffered an injury to the lower limb/back in the past six weeks. Participants

were also required to have had no previous exposure to an ACL injury

prevention program as part of their training history. These exclusion criteria

were in line with similar studies in this area.[125, 249-252] Participants were also

required to be free of any neuromuscular or musculoskeletal disorders that

affected the lower limb. The details of specific participant demographics for

each study are presented in their respective chapters.

Upon recruitment, participants were asked whether any foreseeable

circumstances would limit their ability to complete the follow-up testing

session (i.e. testing session two) six-weeks after baseline testing. Where this

was the case, these participants were allocated to a ‘singular’ testing group

where their only commitment was attending a single testing session. As there

were single-time point studies included in this thesis, an inability to attend the

follow-up testing session did not exclude participants from this research.

Participants whom reported as eligible to attend both testing sessions were

subsequently randomised to either the control or training group. All

participants completed an initial testing session, with participants allocated to

the control and training groups returning six-weeks later to complete an

identical follow-up testing session. In this six-week period participants within

the control group continued their regular training practices, while those in the

training group undertook a six-week ACL injury prevention program (see

Figure 3.1, Page 35).

Methodology

35

Figure 3.1 Diagrammatic representation of the experimental flow of

recruited participants. a – four participants (3 could not be

contacted; 1 sustained injury) from the control group did not return for testing session two; b – two participants (could not

be contacted) from the training group did not return for testing

session two.

3.3 Experimental Procedures

A number of experimental procedures were used to meet the aims of this

thesis. Data were collected during one or two laboratory-based testing

sessions, depending on group allocation. Subsequent chapters within this

thesis use data from one or both of these testing sessions. Data from the

baseline testing session was used for the studies where data from a single time-

point was examined (i.e. Chapters Four, Five, Eight and Nine), while data from

both testing sessions was used for the study where comparisons across time-

points were made (i.e. Chapter Six). The tasks completed within these sessions

Methodology

36

were identical across participants and testing sessions. During laboratory-

based sessions, participants completed a range of tasks in a randomised order.

As these tasks required the performance of multiple trials, adequate rest

between trials (i.e. 60-90 seconds) and tasks (i.e. five to ten minutes) was given

in an effort to minimise fatigue. In addition to the laboratory-based testing,

participants allocated to the intervention group completed a six-week injury

prevention program between sessions. The following sections detail how these

experimental procedures were performed, as well as key considerations

surrounding the use of these procedures within laboratory-based testing

sessions.

3.3.1 Netball-Specific Landing Task

The netball-specific landing task chosen for this research was a leap landing.

A leap landing involves a run-up, followed by a single limb take-off and land

on the contralateral limb (see Figure 3.2).[157] This task was examined as part of

this research for three key reasons. First, it represented a sport-specific task

relevant to the cohort being examined. Second, the deceleration and extended

lower limb position associated with the leap landing technique are relevant to

ACL injury mechanisms.[245] Finally, leap landings are a known mechanism for

ACL injuries in netball.[245]

Figure 3.2 Sagittal view of the leap landing task.

Methodology

37

Leap landings are often performed in games when catching a pass,[157] and the

inclusion of a ball in laboratory-based testing can influence lower limb

neuromuscular and biomechanical measures.[54, 154, 155] Therefore, participants

were required to catch a pass distributed with a flat trajectory while

completing the landing task.[248, 253] Each trial was initiated by breaking and

accelerating from a static defender (model skeleton) from a position six-metres

from the landing area, followed by a run-up and leap landing whilst catching

a pass. Passes were received from 2.5 metres in front of the landing area at a

30 degree angle on both the left and right sides (see Figure 3.3, Page 39). To

ensure consistency the pass was completed by the same experienced player

for all trials, with the passer instructed to pass between waist and head height.

The quality of the pass was observed by an investigator during each trial, and

trials were excluded if the pass fell outside of the height range or was not

within half a metre of the participant. Trials were also excluded if the

participant failed to catch the pass.

During testing, participants performed 20 successful leap landing trials on a

designated ‘landing limb.’ The landing limb was determined prior to data

collection by asking participants which limb they would be more comfortable

landing on when catching a pass during a game. Further to this, players were

given an opportunity to trial landing on each limb to identify which they were

more comfortable landing on. Of the 20 successful trials recorded; ten landings

each were performed with the participant receiving passes from their left and

right (see Figure 3.3, Page 39), with the order of these randomised throughout

trials. Preliminary analysis of data revealed no difference between the side the

pass was received from for neuromuscular and biomechanical variables. All

trials, irrespective of the pass side, were subsequently analysed together. A

successful trial was characterised by: (1) performing the leap landing on the

Methodology

38

designated landing limb with the entire foot landing within the borders of the

force plate; (2) successfully catching a pass thrown between chest and head

height; and (3) coming to a complete stop within one step after landing.

Participants were instructed to perform the run-up and landing at ‘game-

pace.’ Approach speeds were monitored using photoelectric timing gates

(SpeedLight, Swift Performance Equipment, Carole Park, Queensland,

Australia) placed at one metre intervals over the three metres prior to the

landing. Approach speeds within a 3.5 – 5.5 m·s-1 range have been observed

for leap landings performed within a game setting,[254] therefore trials

performed outside of this range were excluded from analysis. Further details

on approach speed data collection and analysis are outlined in Section 3.4.4 of

this chapter. Video recordings and force plate data were used to inspect

landings to ensure a leap landing was performed and the criteria for a

successful trial was met. Trials were repeated until the 20 successful trials were

recorded. The average number of trials performed by participants across all

sessions was 27 ± 4. Participants were allocated 60-90 seconds rest between

trials in an effort to minimise fatigue.[134, 255] The laboratory set-up and

instrumentation used to record leap landing performances is illustrated in

Figure 3.3 (see Page 39).

Methodology

39


instrumentation used for the leap landing task.

Methodology

40

3.3.2 Bilateral Drop Vertical Jump

Participants were required to perform a bilateral drop vertical jump task

incorporating both horizontal and vertical displacement. Participants jumped

from a 30 cm high box positioned at a distance 50% of their height from the

landing area.[91] Upon landing, participants were required to immediately

rebound for a maximal vertical jump (see Figure 3.4).[91]

Figure 3.4 Sagittal view of the bilateral drop vertical jump-landing task.

Participants were instructed on how to complete the task with an emphasis

placed on jumping as high as possible during the vertical jump phase, and

were not provided any feedback during or in between trials unless they

performed the task incorrectly.[91] After task instruction, participants were

given as many practice trials as necessary to become proficient at the task.

Participants performed three successful trials of the drop vertical jump-

landing task. A successful trial was characterised by: (1) jumping off both feet

from the box; (2) jumping forward, but not vertically to reach the landing area;

and (3) completing the task in a fluid motion.[91]

Video data from the drop vertical jump task were used to assess landing

technique using the Landing Error Scoring System (LESS).[91] The laboratory

Methodology

41

set-up and instrumentation used to record bilateral drop vertical jump trials is

illustrated in Figure 3.5. The limb of interest during the drop vertical jump task

was the ‘landing limb’ used for the leap landing task. Therefore, the lateral

video camera was moved to the left of the landing area for participants who

identified their left leg as the ‘landing limb.’


instrumentation used for the drop vertical jump-landing task.

Positioning of the lateral video camera is representative of a

participant who identified their right side as their preferred ‘landing limb.’

3.3.3 Tuck Jumps

Participants were required to perform a repeat tuck jump task, attempting to

bring the knees to the chest during each jump, for a period of ten

seconds.[86, 256] Participants were instructed to begin in a starting position with

feet shoulder width apart and initiate the jump with a slight downward crouch

while extending their arms behind them; bring the knees up as high as possible

Methodology

42

during the jump; and on landing immediately begin the next jump (see Figure

3.6).

Figure 3.6 Sagittal view of the tuck jump task.

Participants received further basic instructions and goals for the performance

of tuck jumps, including; attempting to lift the thighs to a parallel height,

aiming to minimise the time in contact with the ground after each jump, and

trying to land on the same spot after each jump.[256] Participants performed one

successful trial of the tuck jump task. Successful trials were characterised by

the participant maintaining the performance of tuck jumps for the ten second

period. Video data from the tuck jump task were used to assess the

participants landing technique using the Tuck Jump Assessment.[86] The

laboratory set-up and instrumentation used to record tuck jump trials is

illustrated in Figure 3.7 (see Page 43).

Methodology

43


instrumentation used for the tuck jump task.

3.3.4 Single-Leg Squat

Participants were required to perform a single-repetition single-leg squat task.

The starting position for the single-leg squat task[257] is illustrated in Figure 3.8.

Figure 3.8 Starting position for the single-leg squat task.

Methodology

44

Participants were instructed to keep their arms parallel to the floor with the

forearms together while performing the squat at a “comfortable pace,”

dropping down to a “comfortable depth,” and then returning to their starting

position (see Figure 3.9). Pace and depth of the squat were not controlled for

across participants as it was hypothesised these factors would be relevant to

the participants self-selected movement strategy, and the strategy employed

may reveal important information regarding their ability to stabilise the lower

limb.

Figure 3.9 Sagittal view of the single-leg squat task.

Participants performed one successful trial of the single-leg squat task.

Successful trials were characterised by the participant maintaining their

balance (i.e. not placing their opposite foot on the ground) and the appropriate

arm position through the entire squatting movement. The laboratory set-up

and instrumentation used to record single-leg squat trials is illustrated in

Figure 3.10 (see Page 45). Participants performed the single-leg squat task on

the limb they used as their ‘landing limb’ for leap landing task, with the lateral

video camera always placed on the squatting limb side of the participant (i.e.

camera was moved to the left for who identified their left leg as the ‘landing

limb’).

Methodology

45


instrumentation used for the single-leg squat task. Positioning of the lateral video camera is representative of a participant

who identified their right side as their preferred ‘landing

limb.’

Methodology

46

3.3.5 Injury Prevention Program

Participants allocated to the training group undertook a six-week injury

prevention program between testing sessions. The injury prevention program

used was the ‘Down 2 Earth’ program.[258] ‘Down 2 Earth’ (D2E) is an

established netball-specific injury prevention program targeting lower limb

injuries by encouraging safe and effective landing technique.[258, 259] The

program involved a total of three sessions per week; where one session was

completed under supervision of a research team member, and the remaining

two conducted as ‘home-based’ sessions completed by the participant

individually. The same researcher completed the supervised training sessions

across all weeks for all participants. The home-based sessions were designed

to replicate the training sessions conducted under supervision, with some

modifications where applicable.[254] For each participant, attendance at the

supervised sessions was recorded. Participants were given a booklet that

guided them through each session, and also contained a diary page to monitor

attendance to training sessions (see Appendix D, Page 379). The exercises used

within the program were based on those used in injury prevention programs

shown to reduce the risk of ACL injury,[37, 212] with necessary adaptations to

make the exercises applicable to netball.[254] The D2E program includes

guidelines for safe and effective landing technique, progressive exercises with

specified sets and repetitions, and instructions for providing/receiving

feedback during sessions.[258, 259]

Methodology

47

The main tenet of the program is that improved landing technique can be

facilitated via a feedback schedule, directing focus to three key guidelines for

safe and effective landing technique,[258] those being; (i) control the hip and the

knee when landing, do not allow the hip and knee to swing inward or

outward on landing; (ii) keep the knee and toe direction the same; and (iii)

ensure a ‘soft’ and slightly bent knee on landing. During supervised testing

sessions, verbal feedback was received from the supervisor surrounding these

three key landing features. Participants were provided with a visual iteration

of these guidelines within the program booklet (see Figure 3.11), which they

used to facilitate feedback during the home-based sessions.



program. Adapted and reproduced with permission from Saunders et al.[258]

The program incorporated a core set of exercises around this feedback

protocol; involving stationary jump-landings, directional jump-landings,

lunging, and netball-specific landing movements.[258] These core exercises were

progressed throughout the six-week program by adding to or varying the

Methodology

48

movement to increase difficulty. Specifically, stationary jump-landing

movements were performed bilaterally and unilaterally, and progressed via

the inclusion of a ball and the addition of rotation (see Figure 3.12, Page 49).

Directional jump-landings were performed bilaterally and unilaterally, and

were progressed via the inclusion of a ball (see Figure 3.13, Page 50). Lunging

movements were performed in a static, dynamic and plyometric manner, in

forward and sideward directions, and were progressed via the inclusion of a

ball (see Figure 3.14, Page 51). The sport-specific landing movement was

progressed via the addition of a defender, and subsequent pivoting and

passing movements following landing (see Figure 3.15, Page 52).

The reasoning behind the use of the D2E program was two-fold. First, the D2E

program has been shown to positively alter neuromuscular and biomechanical

risk factors associated with injuries to the lower limb, particularly those to the

ACL.[254] Second, it was important that the findings surrounding the injury

prevention program were translatable to wider community settings. There is

evidence that community level netball coaches are willing to implement the

D2E program as they believe it can improve athletic attributes and reduce

injury risk.[259] Based on these findings, it was believed the D2E program was

appropriate for use within this thesis.

Methodology

49

Figure 3.12 Variations in stationary jump-landing exercises used within

the ‘Down 2 Earth’ injury prevention program. A) Bilateral

stationary jump-landing; B) Unilateral stationary jump-

landing; C) Bilateral stationary jump-landing with catch; D) Unilateral stationary jump-landing with catch; E) Unilateral

stationary jump-landing with rotation. Adapted and

reproduced with permission from Saunders et al.[258]

Methodology

50

Figure 3.13 Variations in directional jump-landing exercises used within

the ‘Down 2 Earth’ injury prevention program. A) Bilateral directional jump-landing task; B) Unilateral directional jump-

landing task; C) Bilateral directional jump-landing task with catch. ‘NSEW’ is used to describe a sequence that involves

jumping in a forward, backward, left and right pattern.

Adapted and reproduced with permission from Saunders et al.[258]

Methodology

51

Figure 3.14 Variations in lunging exercises used within the ‘Down 2 Earth’

injury prevention program. A) Stationary lunging exercise; B)

Step lunging exercise; C) Stationary lunging exercise with catch; D) Side-step lunging exercise; E) Stationary jump-

landing exercise; F) Stationary jump-lunging exercise with catch. Adapted and reproduced with permission from

Saunders et al.[258]

Methodology

52

Figure 3.15 Variations in sport-specific landing exercises used within the

‘Down 2 Earth’ injury prevention program. A) Netball-specific

leap landing task with catch; B) Netball-specific leap landing task with catch and pass to a known player; C) Netball-specific

leap landing task with catch and pass to an unknown player; D) Netball-specific leap landing task with catch against a

defender and pass to an unknown player. Adapted and

reproduced with permission from Saunders et al.[258]

Methodology

53

3.4 Data Collection and Analysis

Within testing sessions, electromyography, kinematic and kinetic data were

collected from participants. Participants were also monitored using video

footage from frontal and sagittal views. The following sections discuss key

considerations regarding the use of these procedures in biomechanical

research, provide a detailed outline of how these procedures were

implemented, and outline the general processing methods used to analyse

collected data. Throughout chapters, readers will be informed as to which data

were used within each study.

3.4.1 Electromyography (EMG)

The complex movement patterns associated with dynamic sporting and

landing manoeuvres are produced by forces generated by the muscular

system.[260] The electrical signal associated with motorneuron depolarisation

causing contraction of a muscle can be measured via electromyography

(EMG).[261] EMG provides a valuable tool for assessing neural control of human

movement.[262] The eventual signal recorded via EMG originates within the

central nervous system. Motor units, which contain an alpha motor neuron

connected to a number of muscle fibres via motor end plates, form the basic

functional unit for excitation and contraction of skeletal muscle.[263] Motor

units are recruited in an orderly fashion by the central nervous system to

produce skeletal muscle contractions.[261] The activation of a motor neuron

results in the conduction of an excitatory signal along the motor nerve.[263] As

this signal reaches the muscle fibres at the motor end plates, the diffusion

characteristics of the muscle fibre membranes are briefly modified causing

membrane depolarisation.[261, 263] This depolarisation results in a ‘wave’ along

the direction of the muscle fibres.[261] It is this depolarisation and subsequent

Methodology

54

repolarisation wave that is recorded by EMG electrodes.[261] The EMG signal

gives information regarding which muscle(s) are active during certain

movements.[261] Therefore, EMG affords the opportunity to measure the

electrical activity associated with muscle activation and identify the neural

contribution of muscles to human movement.

Although EMG provides valuable information surrounding the neural control

of human movement, it is not without limitations. A major limitation of EMG

is that it does not provide information about muscle force.[262] As the EMG

signal is a measure of electrical and not mechanical events, it does not provide

a direct measure of the force or torque being produced from the muscle

contraction.[261] The physiological characteristics of a muscle can also alter the

interpretation of EMG signals. At the same activation level, a given muscle can

produce more force if it has a larger cross-sectional area, or is operating at a

more optimal length and/or velocity.[262] An additional limitation associated

with EMG recordings is the inherent electromechanical delay between the

electrical signal and subsequent mechanical effect of the muscle

contraction.[262] To assume that when a muscle is activated it instantly provides

a mechanical effect is erroneous. This delay must therefore be considered

when attempting to identify the muscle activation characteristics that support

against external loads during high-risk sporting tasks. In order to be effective,

the muscle must be activated prior to when support against external loading

is required to ensure the muscle is mechanically effective. Despite the

limitations, EMG provides an option for identifying neuromuscular

contributions to ACL injury risk due to the link between neural and muscular

systems. As the muscular system is responsible for driving movement,

abnormalities or variation in muscle activation may be the underlying

mechanism behind undesirable lower limb biomechanics during high-risk

Methodology

55

sporting tasks. The association of muscle activation characteristics to these

undesirable lower limb biomechanics provides a basis for identifying

neuromuscular risk factors. Multiple studies have highlighted our ability to

alter muscle activation patterns through injury prevention programs.[34-45] As

the neuromuscular system is modifiable, identification of neuromuscular risk

factors is pertinent to ACL injury prevention research. Within this thesis EMG

was used to measure the activity of lower limb muscles, and the subsequent

links between activation of these muscles and lower limb biomechanics during

the netball-specific landing task. Preparation of electrode sites, location of

electrodes, signal processing, and reporting of EMG data were in line with

standardised procedures.[264-266]

3.4.1.1 Electrode Properties

Surface electrodes are commonly used within biomechanical research

examining jump-landing or sport-specific tasks,[44, 56, 131, 133, 134, 171, 178, 181, 267-269] as

they are non-invasive and easily applied. Surface electrodes provide a gross

representation of muscle activity underlying the placement area.[270, 271] This

type of electrode is often preferable to record the activity of large superficial

muscles as it ensures the EMG recording is representative of the activity of

many muscle fibres. Due to the large recording area of surface electrodes, there

is the potential for crosstalk in contaminating the recorded signal from

muscles near the muscle of interest.[272] Surface electrodes have shown good

reliability both within- and between-sessions for lower limb EMG recordings

during a range of dynamic tasks.[273-278] However, selecting the appropriate

electrode size and location of electrodes must be carefully planned to avoid

crosstalk.[266]

Methodology

56

The configuration, size, material, and inter-electrode distance all may affect

the quality of EMG recordings when using surface electrodes. Wireless

parallel-bar electrodes (Delsys Trigno, Delsys Incorporation, Boston, MA,

United States) were used for the studies within this thesis. Large contact areas

and inter-electrode distances contribute to a greater risk of crosstalk from

adjacent and underlying muscles, while contamination of the EMG signal can

occur from movement of cables connecting electrodes to an amplifier.[272] The

wireless parallel-bar electrodes used within this research are advantageous in

minimising these effects by providing a small and consistent inter-electrode

distance (10mm) and eliminating the need for cables connecting the electrode

to an amplifier.[279]

3.4.1.2 Surface Electrode Placement and Fixation

EMG data were collected unilaterally from eight lower limb muscles.

Appropriate skin preparation ensured adequate electrode-skin contact,

reducing noise and artificial artefact, thereby improving the quality of EMG

recordings.[264] Prior to electrode attachment, the skin was prepared using a

standardised procedure in accordance with published recommendations.[264]

Placement sites were dry shaven (if required) to remove any hair using a

disposable razor, followed by gentle abrasion and sterilisation of the area

using Nuprep® skin preparation gel (Weaver and Company, Aurora, CO,

United States) and isopropyl alcohol applied via non-sterile gauze swabs

(Sentry Medical, Wetherill Park DC, New South Wales, Australia). Following

skin preparation, electrodes were affixed to the skin at the muscle sites of

interest. Tuf-Skin® adhesive spray (Cramer Sports Medicine, Cramer Products

Inc., Gardner, KS, United States) was applied to improve electrode adherence,

with the electrode affixed to the skin via a specially tailored adhesive interface

(Delsys Trigno, Delsys Incorporation) (see Figure 3.16, Page 58). Electrodes

Methodology

57

were further affixed with stretchable, adhesive, non-woven tape (Fixomull®

Stretch, Smith & Nephew, London, England) and variable-sized elastic net

bandage (Surgifix®, Smith & Nephew) applied over electrodes to improve

adherence and reduce movement on the skin.

Data were collected from the gluteus maximus, gluteus medius, rectus

femoris, vastus medialis, vastus lateralis, medial hamstrings, lateral

hamstrings, and lateral gastrocnemius. These muscles were chosen as they all

have the ability to influence motion at the hip and/or knee joint, and have been

studied in the context of neuromuscular control and ACL injury risk during

athletic tasks.[56, 131, 134, 171, 178, 184-187, 269, 280-282] Electrodes were orientated parallel to

the direction of the muscle fibres and positioned relative to anatomical

landmarks for optimal signal quality and avoidance of innervation zone

locations in accordance with established procedures (see Figure 3.17, Page 58

and Table 3.1, Page 59).[265, 283] Palpation and requesting the participant to

voluntarily contract the desired muscle were used to confirm correct electrode

placement on the muscle belly. Electrode placement and manual muscle

testing were performed by the same investigator on every participant during

testing sessions.

Methodology

58

Figure 3.16 Image of Delsys Trigno parallel-bar surface electrode with

specially tailored adhesive skin interface attached.

Figure 3.17 Locations of EMG surface electrodes for an example

participant (right leg as landing limb). The view of the gluteus maximus (GMAX) and gluteus medius (GMED) locations are

obstructed by the participant’s shorts.

Methodology

59

Table 3.1 Guidelines used for surface electrode placement.[265, 283]

Muscle Electrode Placement Guidelines

Gluteus Maximus (GMAX) Placed on the posterior aspect of the buttocks, 55% of the distance along a line from the second sacral vertebrae to the greater trochanter.

Gluteus Medius (GMED) Placed on the lateral aspect of the buttocks, 50% of the distance along a line from the iliac crest to the greater trochanter.

Medial Hamstrings (MHAM) Placed on the posterior-medial aspect of the thigh, 50% of the distance along a line from the ischial tuberosity to the medial side of the popliteus.

Lateral Hamstrings (LHAM) Placed on the posterior-lateral aspect of the thigh, 50% of the distance along a line from the ischial tuberosity to the lateral side of the popliteus.

Rectus Femoris (RF)

Placed on the anterior aspect of the thigh, 130 mm from the superior aspect of the patella, along a line from the superior aspect of the patella to the anterior superior iliac spine.

Vastus Lateralis (VL)

Placed on the anterior-lateral aspect of the thigh, 120 mm from the superior-lateral aspect of the patella, along a line from the superior-lateral aspect of the patella to the anterior superior iliac spine.

Vastus Medialis (VM)

Placed on the anterior-medial aspect of the thigh, 37 mm from the superior-medial aspect of the patella, along a line from the superior-medial aspect of the patella to the anterior superior iliac spine.

Lateral Gastrocnemius (GAS) Placed on the posterior-lateral aspect of the shank, 25% of the distance along a line from the lateral knee joint line to the lateral malleolus.

Methodology

60

3.4.1.3 EMG Data Acquisition and Signal Processing

EMG data were collected digitally through Vicon Nexus (Vicon, Oxford

Metrics Limited, Oxford, United Kingdom) sampling at 2000 Hz, and online

bandpass filtered between 10 and 500 Hz. It is recommended that the sampling

rate for EMG data be at or higher than the frequency of the online bandpass

filter,[266] therefore the sampling rate of 2000 Hz was deemed appropriate. Prior

to movement trials, a 30 second baseline file with the participant standing

quietly was collected to determine baseline EMG levels. Following this,

maximal voluntary isometric contractions (MVICs) were performed for the

muscles/muscle groups of interest with participants placed in standardised

positions (see Appendix E, Page 390) in accordance with established

procedures and guidelines.[263, 264, 266] Three maximal effort contractions were

performed against manual resistance for each muscle/muscle group. For each

contraction, participants increased muscle tension over three seconds,

maintained tension for three seconds, and released tension over three seconds.

Collected EMG data were exported from Vicon Nexus and processed offline

using custom MATLAB (version R2014a) (The Mathworks Incorporated,

Natick, MA, United States) programs. Prior to processing, all signals were

checked for artefacts or artificial noise. Some trials (less than one per

participant per testing session) were lost due to significant impact artefacts

contributing noise to or saturating the signal. Raw EMG signals were corrected

for bias by subtracting the mean of each participants’ baseline recording for

the respective muscle. Signals were submitted to a fourth-order, zero-lag

Butterworth digital high-pass filter ( = 20 Hz) to minimise movement

artefact.[82, 284, 285] The filtered EMG data were then full-wave rectified and

converted to a linear envelope via a fourth-order, zero-lag Butterworth digital

low-pass filter ( = 6 Hz).[171, 269] The linear envelopes of EMG signals were

Methodology

61

subsequently used in providing an indication of the activation patterns of the

muscles.[261]

Amplitude normalisation of EMG signals was undertaken to account for

variability among recording factors (e.g. changes in skin-electrode impedance,

subcutaneous tissue thickness, and variations in electrode placement) and to

facilitate comparisons between individuals and muscles. Studies examining

athletic tasks have normalised EMG data to the maximum activation level

achieved during MVICs,[59, 62, 63, 120, 131, 178, 180, 286, 287] the maximum activation level

achieved during movement trials,[171] or a combination of whichever value is

higher from either MVICs or movement trials.[282, 288, 289] Within this research,

dynamic EMG data recorded during movement trials were amplitude

normalised (% MVC) to the larger of the peak value recorded during the

MVICs or the peak EMG activity recorded during any movement trials, similar

to existing studies.[282, 288, 289] A 10 millisecond (ms) moving window was used

to identify peak EMG activity across each movement trial. Using this method

results in all normalised EMG data being at or below 100% of maximal

activation.[288] Following signal processing, the area of interest in the EMG

waveform was extracted from the overall trial and used for subsequent

analyses. For example, EMG data was extracted from the leap landing task

over the period of 150 ms prior through to 300 ms post initial landing contact.

This time window was chosen as previous research examining a similar single-

limb catch and land task has shown lower limb muscle onsets occur up to 150

ms prior to landing and remain active for a period of up to 380 ms.[54] A

schematic representation of the EMG processing framework is presented in


Methodology

62

Figure 3.18 Example processing framework for EMG data. A) Uncorrected

raw EMG signal; B) Corrected raw EMG signal filtered with a

20 Hz high-pass fourth order Butterworth filter; C) Full wave

rectified EMG signal; D) EMG signal filtered with a 6 Hz low-pass fourth order Butterworth filter to generate a linear

envelope, dashed red lines indicate area of data to be extracted

for further analysis; E) Extracted linear envelope EMG signal; F) EMG signal normalised to percentage of maximum

activation achieved by the participant for the example muscle.

3.4.1.3.1 Muscle Synergies

Within this thesis, muscle synergies were used as a means to identify and

characterise muscle activation profiles. A recent study by Kipp et al.[290] utilised

muscle synergies to identify the different neural control strategies employed

by prepubescent boys and girls during a drop landing task. Muscle activation

profiles employed during movement tasks are proposed to stem from a set of

relatively few neural commands, known as muscle synergies.[290-293] Muscle

Methodology

63

synergies are the activation patterns of individual or multiple muscles by a

single neural command, and can provide meaningful physiological

information about the control of muscle activation.[290, 292, 293] The hypothesis

that synergistic neural control of muscles is employed within biomechanical

tasks is well supported.[48-50, 290-294] Analysis of muscle synergies provides an

approach to study how these neural commands translate to muscle

activation.[291] Muscle synergies provide a useful strategy for the control of

complex movements as it reduces the number of output patterns the

neuromuscular system must specify for a large number of muscles.[49] For

example, it has been shown that five and three muscle synergies account for

the majority of activation of lower limb muscles during walking[293] and

landing[290] movements, respectively. In addition, muscle synergies are often

associated with the major kinematic and kinetic events of the movements

being analysed.[48, 49, 290, 293] A muscle synergy consists of a time-variant

activation coefficient (hereby referred to as “activation coefficients”), which

represents the relative contribution of the muscle synergy to the overall

activation; and time-invariant weighing coefficients (hereby referred to as

“muscle weighing coefficients”), which represent the relative weighting of

activation for each of the muscles of interest within the synergy (see Figure

3.19, Page 64).[48, 49, 290, 295, 296]

Methodology

64

Figure 3.19 Illustration of muscle synergies. The activation of individual

muscles (M1 – M6) is determined by the sum of the product of

the weighing coefficient (labelled as ‘W’) and activation coefficient (labelled as ‘c’) across all muscle synergies. The

weighing coefficients determine how many muscles are

recruited within each synergy along with how much the muscles are activated. The activation coefficients capture when

these muscles are active. Three muscle synergies (black, dark gray and light gray shading) contributing to overall muscle

activation are illustrated. Reproduced with permission from

Kipp et al.[290]

Factor analysis, independent component analysis, and non-negative matrix

factorisation have all been used in previous research to extract muscle

synergies from EMG data.[49] Although each of these analytical approaches

place different restrictions on the outcomes, they all result in a similar solution

relating to the temporal structure of the muscle activation patterns.[297] Non-

negative matrix factorisation (NMF) is the most common method used for the

extraction of muscle synergies from EMG data.[48-50, 296] Unlike other

decomposition methods, NMF results in positive values being obtained for

both the muscle activation and weighing coefficients, which leads to a simpler

and more physiologically accurate interpretation.[296] For these reasons, NMF

was used as the analytical approach for extracting muscle synergies within this

Methodology

65

thesis. Specifically, the Lee and Seung algorithm[298] was implemented. This

algorithm assumes that the EMG signals from each muscle can be described

as a linear combination of a set of muscle activation and weighing

coefficients.[294] This assumption can be described by the equation:

where is an m x n matrix (m and n being the number of muscles and

synergies, respectively) that specifies the time-invariant relative weighing

coefficients of each muscle within the synergy; is an n x t matrix (t being the

number of sample frames of EMG data) that specifies the time-varying

activation coefficients (i.e. the relative contribution of each synergy to overall

muscle activation); is an m x t matrix representing the reconstructed

EMG data from the multiplication of and ; and is the residual error

between and the experimentally collected EMG data ( ).[294] The

algorithm can be iterated to explore a number of different solutions (i.e.

different number of muscle synergies). At each iteration, and are updated

in order to minimise the residual error between and .[48, 49, 294, 298]

Where implemented within this thesis, the NMF algorithm was used to

decompose a matrix containing all of the processed EMG signals into two

separate matrices containing the muscle activation and weighing coefficient

values. This analysis was iterated by varying the number of synergies between

one and six. The number of muscle synergies extracted was determined by

calculating the variance accounted for ( ) value between and

across the different synergy solutions.[48, 49, 290, 294, 296] It is commonly

accepted that a mean total (i.e. across all muscles) over 90% is an

appropriate threshold to accept reconstruction of the EMG data.[48, 49, 294, 296]

Methodology

66

Therefore, the least number of synergies that accounted for a mean total

> 90% was used as the final solution.[49] Mean total was calculated as:[49, 294]

where m and t represent the number of muscles and frames of the extracted

EMG data, respectively. The for individual muscles ( ) was also

calculated to ensure that each muscles activity patterns were accounted for by

the retained synergies (i.e. > 75%).[48] was calculated as:[296]

Where and represent the reconstructed and experimental

EMG signals at n frames of time (t) of the extracted EMG data, respectively.

Following factorisation, the muscle activation coefficients were normalised to

their maximum value (i.e. each muscle activation coefficient had a maximum

value of 1), with the muscle weighing coefficients within each of the muscle

synergies scaled by the same respective factor.[49, 294] This ensured that

comparisons of muscle weighing coefficients could be made between trials

and participants both within and between muscle synergies. These normalised

muscle weighing coefficients from muscle synergies were used in subsequent

analyses to characterise muscle activation profiles.

3.4.2 Kinematics and Kinetics

Present computer aided optical motion capture systems currently provide the

most accurate method for quantifying kinematics during human

movement,[261] and are commonly used for the examination of athletic

tasks.[126, 131, 138, 149, 166, 167, 299-304] The relative angular rotations of individual body

Methodology

67

segments can be estimated from the positional data of retro-reflective markers

placed on anatomical landmarks and segments using established

musculoskeletal models and techniques.[305-308] Force platform technologies can

accurately quantify the external forces being distributed over an area of the

body (e.g. ground reaction forces under the area of the foot).[261] These forces

are often represented as vectors originating from a central point of pressure,

and are used in conjunction with kinematic data in inverse dynamics

calculations to determine moments of force acting upon segments of the

body.[261] Within this thesis, joint kinematics of the lower limb were measured

to describe the patterns of motion during the movement tasks examined. In

addition, ground reaction force data were collected to determine and calculate

the external forces and joint moments acting on the lower limb.

3.4.2.1 Kinematic Data Acquisition

Three-dimensional (3D) kinematics of the lower limb were measured using an

eight-camera Vicon MX motion analysis system (Vicon, Oxford Metrics

Limited) sampling at 250 Hz. Eight cameras were used to monitor a capture

volume of five by three metres, with the minimisation of marker dropout the

primary factor in determining camera height and placement. Prior to data

collection, a two-part calibration procedure was undertaken in accordance

with standard Vicon procedures. First, a dynamic calibration was performed

by moving a 5-marker L-frame wand throughout the capture volume. The

same wand was then used for a static calibration to define the origin (i.e. X-Y-

Z coordinates of 0-0-0) of the capture volume and orientation of the global

coordinate system (X, Y and Z representing anterior-posterior, medial-lateral

and vertical directions, respectively). The dynamic calibration calculated

camera residuals providing reconstruction accuracy for each individual

camera. Camera residuals of less than 0.3 pixels (< one millimetre) were

Methodology

68

established as a requirement for the calibration to be accepted. Where

residuals were above this threshold, the dynamic calibration process was

repeated until the predetermined values were achieved.

3.4.2.1.1 Marker placement and kinematic model

Thirty-two 14-millimetre retro reflective markers were attached to

participants’ pelvis and lower limbs in accordance with an established

musculoskeletal model.[305] This model has been previously described and

validated[305] and is widely used in biomechanical studies examining athletic

tasks.[56, 126, 138, 149, 176, 300, 309-311] Markers were attached bilaterally on the iliac crest

(ILCR), anterior superior iliac spine (ASIS), posterior superior iliac spine

(PSIS), greater trochanter (GTR), lateral aspect of the thigh (LTHI), anterior

aspect of the thigh (THI), lateral epicondyle of the femur (LEP), medial

epicondyle of the femur (MEP), anterior-proximal aspect of the shank (PSH),

lateral aspect of the shank (LSH), anterior-distal aspect of the shank (DSH),

lateral malleolus (LMAL), medial malleolus (MMAL), base of the third

metatarsal head (TOE), base of the fifth metatarsal head (5TH), and calcaneus

(HEE) (see Figure 3.20, Page 69). All markers were initially secured using

double-sided adhesive tape. Soft tissue artefact is a common source of error

associated with marker-based motion capture.[312] To improve adherence,

markers were further affixed by applying stretchable, adhesive, non-woven

tape (Fixomull® Stretch, Smith & Nephew). The same investigator performed

marker placement on all participants across all testing sessions. After marker

placement and prior to the undertaking of movement trials, a two-second

calibration trial was recorded with the participant standing in a stationary

(neutral) position (see Figure 3.20, Page 69).[56, 305, 310] Markers on the left and

right MEP and MMAL were for calibration purposes only, and removed

following the calibration trial.

Methodology

69

Figure 3.20 Marker locations and the stationary calibration pose used to

define a kinematic model comprised of seven skeletal

segments (pelvis; right and left thigh; right and left shank; and right and left feet). The medial epicondyle (MEP) and medial

malleoli (MMAL) markers (red) were removed for the

recording of movement trials. Adapted and reproduced with permission from McLean and Samorezov.[310]

A kinematic model[305, 310] consisting of seven skeletal segments (pelvis; right

and left thigh; right and left shank; and right and left feet) and 24 degrees of

freedom (DoF) was defined within Visual 3D software (C-Motion, Rockville,

MD, United States) from the participant’s stationary (neutral) trial. The pelvis

was assigned three translation and rotational DoF relative to the global

(laboratory) coordinate system, with the proximal and distal ends of the

segment defined by markers placed on the iliac crest and greater trochanters,

respectively.[56, 310] The hip, knee and ankle joints were defined locally and each

assigned three rotational DoF.[56, 310] Rotations of the hip joint were defined

about three orthogonal axes passing through a fixed joint centre defined

according to Bell et al.[306]. Knee joint motion was described by rotation about

Methodology

70

a fixed flexion-extension axis,[176] passing through a joint centre defined by the

midpoint between the lateral and medial epicondyle markers in accordance

with Vaughan et al.[308]. Ankle joint motion occurred around a joint centre

defined as the midpoint between the lateral and medial malleoli, with the

plantarflexion-dorsiflexion axis extending laterally from this point.[176, 307]

3.4.2.2 Kinetic Data Acquisition

Ground reaction force (GRF) data was collected using a single 600 x 900

millimetre in-ground AMTI force plate (Advanced Mechanical Technology

Incorporated, Watertown, MA, United States) sampling at 1000 Hz. To avoid

deterioration of the signal over the testing session, the force plate was zeroed

between trials for all tasks. Participants were made aware of the position of

the force plate and were instructed to position their ‘landing limb’ within the

boundaries of the force plate during relevant landing movements. Visual

targeting of force plates has been investigated as a potential confounding

factor in biomechanical studies where ground reaction force measures are

taken.[313, 314] While visual targeting strategies (e.g. adjusting stride pattern or

increasing gaze towards force plate) have been identified within laboratory-

based sessions incorporating a force plate, there appears to be no effect on

kinematic or kinetic outcomes when these strategies are adopted.[313, 314]

3.4.2.3 Signal Processing

Three-dimensional marker trajectory data during movement trials were

reconstructed and labelled within Vicon Nexus (Vicon, Oxford Metrics

Limited). Any gaps in marker trajectories were filled using either a cubic spline

or pattern fill method. The method used was determined by visual

examination of the proposed fill trajectory and selection of the best fit. Where

the pattern fill was used, a marker on the same segment was used as the source

Methodology

71

to construct the fill trajectory. Trials where large gaps (i.e. > 25 frames)

occurred were discarded. The appropriate placement of cameras ensured this

rarely occurred, with less than one trial per participant on average being lost

for this reason. Following these steps, marker trajectory and GRF data were

exported to Visual 3D (C-Motion) for further analyses.

There is a need to low-pass filter marker trajectory and GRF data when

examining high-impact athletic or sporting movements to remove excessive or

random noise.[315, 316] While marker trajectory data is generally filtered with a

5-20 Hz cut-off frequency, GRF data has been filtered using higher cut-off

frequencies.[315] However, recent research has shown that using difference cut-

off frequencies for marker trajectory and GRF data within inverse dynamics

calculations results in variations in the magnitude of peak joint moments, as

opposed to matched cut-off frequencies.[315, 317] The variations in joint moment

data are proposed to be a result of impact artefacts that present during rapid,

high impact athletic activities.[315, 318] Applying the same filter and cut-off

frequency to both marker trajectory and GRF data can reduce the possibility

of artefacts in joint moment data.[319] Inaccuracies in the reporting of joint

moment data can influence conclusions of studies examining risk factors for

ACL injury.[320] Therefore, for all analyses within this thesis, an identical low-

pass filter and cut-off frequency was used to process marker trajectory and

GRF data, with the optimal cut-off frequency determined from analysis of

marker trajectory data.[315, 318, 320]

Raw marker trajectory and GRF data from movement trials were low-pass

filtered with a cubic smoothing spline.[321] A number of methods have been

proposed for determining the optimal cut-off frequency for filtering kinematic

and kinetic data.[261, 322] Residual analysis[261] is a common technique used in

biomechanical studies.[83, 166, 167, 255] This technique involves plotting the residual

Methodology

72

between the filtered and unfiltered signal at a range of cut-off frequencies.[261]

In general, the optimal cut-off frequency is determined by finding the point

that balances the magnitude of signal distortion and noise present following

low-pass filtering at various cut-off frequencies.[261] For all analyses within this

thesis, the optimal cut-off frequency for the low-pass filter was determined via

a combination of residual analysis of marker trajectory data and visual

inspection of kinematic and kinetic data filtered at varying levels.[83, 166, 167, 255]

Data from different movement tasks were filtered at variable levels depending

on the optimal cut-off frequency determined for that task. The specific cut-off

frequencies used are outlined in the methodology sections of each study.

Hip, knee and ankle joint rotations from the participants’ landing limb were

calculated from the filtered marker trajectories from each movement trial. Joint

rotations were calculated using an X-Y-Z Cardan angles sequence; with the X,

Y and Z axes representing joint rotations in the sagittal, frontal and transverse

planes, respectively.[138] Kinematic data were expressed relative to each

participant’s stationary (neutral) trial.[56, 310, 311] Positive values represented

positions of hip flexion, adduction, and internal rotation; knee extension,

adduction (i.e. abduction of the tibia relative to the femur), and internal

rotation (i.e. internal rotation of the tibia relative to the femur); and ankle

dorsiflexion, inversion, and internal rotation. Negative values represented

positions of hip extension, abduction, and external rotation; knee flexion,

abduction (i.e. abduction of the tibia relative to the femur), and external

rotation (i.e. external rotation of the tibia relative to the femur); and ankle

plantarflexion, eversion, and external rotation (see Figure 3.21, Page 73).

Methodology

73

Figure 3.21 Illustration of sign conventions used to represent hip, knee

and ankle joint rotations about three rotational degrees of

freedom. Reproduced with permission from McLean and

Samorezov.[310]

Three-dimensional intersegmental forces and moments were obtained by

submitting the filtered marker trajectory and GRF data to a conventional

inverse dynamics analysis.[261] The choice of reference frame in which to

express lower limb joint moments can affect the results obtained, particularly

during high-impact landing movements.[318, 323] Projection to the joint

coordinate system (JCS) is the only method where joint moments correspond

to muscle and ligament loading, and thus is a natural choice for a standard of

joint moment reporting.[318] Joint moments were therefore expressed in the

reference frame of the hip, knee and ankle JCSs. For each JCS, the sagittal axis

was defined as the medio-lateral axis of the proximal segment; the transverse

Methodology

74

axis as the longitudinal axis of the distal segment; and the frontal axis as the

cross product of the sagittal and transverse axes.[318] Normalisation of joint

moment data has been highlighted as a key step when comparisons are being

made across individuals of different masses and heights.[324, 325] To ensure

appropriate comparisons could be made across participants, joint moments

were normalised by dividing the total joint moment (Nm) by the product of

the participant’s weight and height (kgm). Joint moments within this thesis

represent the external loads applied at each joint. For example, knee abduction

moments refer to an external load that moves the knee into an abducted

position. Positive values represented hip extension, abduction, and external

rotation; knee flexion, abduction, and external rotation; and ankle

plantarflexion, eversion, and external rotation moments. Negative values

represented hip flexion, adduction, and internal rotation; knee extension,

adduction, and internal rotation; and ankle dorsiflexion, inversion, and

internal rotation moments.

3.4.2.3.1 Principal Component Analysis (PCA)

Principal component analysis (PCA) is often applied as a means for data

reduction.[326] The main purpose of PCA is to summarise the key information

in the data, and represent the observations/variables simultaneously in a

limited number of principal components.[326] Mathematically, PCA consists of

an orthogonal transformation, converting n variables ( ) into n

principal components ( ).[326] The principal components are

arranged in decreasing order of the variance in the data they explain, where

the initial principal components explain the largest variations in the data.[326]

The number of components retained is often guided by this magnitude of

variance, where the principal components that sum to explain a

predetermined percentage are retained. Cut-off points of 85-100% of variance

Methodology

75

explained have been used within biomechanical research.[327-334] The principal

component model also contains principal component loading vectors (

), which are the eigenvectors of the covariance matrix of .[326]

The use of PCA in biomechanical studies as a means to identify patterns in

continuous kinematic and kinetic data has increased in popularity over recent

years.[250, 326, 328, 329, 331, 332, 335] PCA has been used to identify neuromuscular and

biomechanical differences between sexes during high-risk cutting

manoeuvres.[250, 328, 329, 336] However, outside of these studies there has been

minimal use of this technique for processing biomechanical data within the

context of ACL injury risk and prevention. When applied to continuous data

(i.e. kinematic and kinetic waveforms), the variables ( ) refer to individual

temporal data points of the waveform.[326] PCA is applied to a p x n matrix,

where p (rows) equals the number of trials/subjects and n (columns) the

number of data points. The principal component loading vectors ( ) are an

orthogonal basis set for the waveform data, and each principal component

represents specific features of the waveform data.[326] Each individual

waveform is allocated a score for the extracted principal components ( ), with

this score measuring the degree to which the shape of a given waveform

corresponds to the equivalent principal component.[326] The extracted principal

components, often labelled as ‘principal patterns,’[276, 327] must be interpreted to

identify the feature of the waveform the pattern describes. Interpretation of

principal patterns obtained from waveform data is often accomplished by

examining the shape of the loading vector ( ), and the individual waveforms

that correspond to high and low principal pattern scores.[326, 327, 337] The principal

pattern can also be interpreted by examining the waveforms obtained by

adding and subtracting a suitable multiple of the components loading vectors

( ) to the mean of the relevant waveform.[333]

Methodology

76

Using functional data analysis methods, such as PCA, to analyse

biomechanical data in a continuous fashion has multiple benefits over discrete

point analysis techniques. Overall, using a discrete point style of analysis can

often discard the majority of data in a waveform.[338] PCA views the entire

waveform as a function, preserving the main features and patterns of the

curve.[333] Using discrete point methods also involves the pre-selection of

features thought to make up the important components of a given

waveform.[338] This pre-selection process is strongly dependent on previous

knowledge or research within the area, and has the potential to discard

important information contained within the waveform.[333, 338, 339] In addition,

examining discrete points does not always take into account the timing of the

points of interest.[338] For example; the amplitude of a peak joint angle across

two trials may be identical, however this peak may occur with substantially

different timing. Simply comparing the peaks between these two trials may

result in inaccurate conclusions being drawn from the data. Lastly, using

discrete point analysis does not allow for the examination of features that

occur over a specific duration of time.[338] Specific phases may hold key

information regarding the pattern of the waveform, with these phases unable

to be captured using a discrete time point approach.[338]

3.4.3 Video Footage

3.4.3.1 Video Data Collection

Two commercial digital video cameras (Mini DV MD225, Canon, Macquarie

Park, Sydney, Australia) recording at 30 frames per second were used to

capture frontal and sagittal plane views of the movement tasks performed

during laboratory testing sessions. Both the frontal and sagittal cameras were

initially positioned 3.54 metres from the landing area. For the most part, this

Methodology

77

distance ensured the full body of participants could be viewed throughout all

movement tasks. However, if the entire body of the participant could not be

viewed, cameras were repositioned to a point where this was achieved. The

sagittal plane video camera was always placed on the side that could view the

participant’s predetermined ‘landing limb.’ The main purpose of video data

was to complete the screening protocols for the Landing Error Scoring System

and Tuck Jump Assessment, as these methods required visual observation of

the drop vertical jump and tuck jump tasks, respectively. Video footage was

also used to ensure the criteria for successful trials were met across movement

tasks.

3.4.4 Approach Speed

Approach speed is important to consider when examining sporting

movements that incorporate a run-up approach. It is particularly pertinent to

do so for the netball-specific landing task examined within this thesis, as

previous work has shown approach speed variations can influence GRF

magnitudes during this task.[340] Studies examining sporting manoeuvres with

a run-up approach often limit trials to a specific approach speed range.[131, 134,

138, 250, 309, 341] However, it is important to initially consider approach speed as a

continuous variable in order to examine its potential effect on lower limb

biomechanics. The procedures used to measure approach speed during testing

sessions, and the means with which approach speed was assessed and

accounted for within subsequent chapters of this thesis are outlined in the

following sections.

3.4.4.1 Approach Speed Data Collection

Approach speed was measured during the netball-specific landing task using

photoelectric timing gates (SpeedLight, Swift Performance Equipment, Carole

Methodology

78

Park, Queensland, Australia). Timing gates were placed at one metre intervals

to monitor speed over the three metres prior to the landing (see Figure 3.3,

Page 39). Approach speed was calculated as the average speed over this

distance.

3.4.4.2 Approach Speed Data Analysis

The 20 successful landing trials from all participants who underwent

laboratory testing were used for approach speed analysis. As not all

participants completed both testing sessions, data from the baseline session

was used. Previous work has shown that approach speed can have a

modulating effect on biomechanical measures during netball-specific

landing.[340] Subsequently, the effect of approach speed on hip and knee joint

rotations and moments was examined. One-dimensional statistical parametric

mapping (SPM1D) canonical correlations analysis and linear regression were

used to examine whether changes in approach speed influenced hip and knee

joint rotations and moments (see Section 3.5.2 within this chapter for a

description of SPM1D procedures). Approach speed was found to have a

statistically significant effect along the majority of hip and knee kinematic and

kinetic waveforms (see Appendix F, Page 391 for full results). Based on these

results, efforts were made to counteract the potentially confounding influence

of approach speed. First; only trials that fell within a 3.5 – 5.5 m·s-1 range were

utilised for analysis. This range was chosen as studies examining the netball-

specific landing task used within this thesis have reported approach speed

values that fall within these ranges.[245, 254] Second; where comparisons of

kinematic and kinetic data were made between groups (i.e. Chapters Four,

Five and Six), approach speeds were initially compared across these groups to

determine whether differences in speed may have influenced the conclusions

drawn from these group comparisons.

Methodology

79

3.5 Statistical Analysis

In addition to the analytical methods outlined in the previous sections, a

number of statistical methods were used across the studies within this thesis.

This section details the statistical methods used and considerations in

applying these methods to neuromuscular and biomechanical data. The

specific application of statistical analysis methods are detailed in their relevant

chapters.

3.5.1 Cluster Analysis

Statistical analyses within neuromuscular and biomechanical research are

generally performed using group-based designs (e.g. males vs. females; trials

from one task vs. trials from another task).[342] In single-group designs, it is

often assumed that movement strategies are similar across cohorts.[343]

However, neuromuscular and biomechanical strategies differ across

individuals,[342] and the intra-individual variability observed in

neuromuscular and biomechanical measures during athletic tasks[344-347]

suggests the same individual may employ a number of different strategies.

Drawing conclusions from a single-group design without accounting for

variations in neuromuscular or biomechanical strategies, particularly those at

the individual level may mask important factors relating to the different

strategies used. An alternative to single-group analysis is sub-group analysis,

which classifies similar movement strategies based on patterns (e.g. EMG or

kinematic waveform patterns) into sub-groups or clusters.[348] An optimal

clustering of trials or individuals maximises the ability to identify outcomes

stemming from the differing strategies utilised within a single-cohort

design.[348] A sub-group approach may elicit more relevant outcomes when, for

Methodology

80

example, examining the neuromuscular contributions to lower limb

biomechanics during high-risk sporting tasks.

Cluster analysis is a statistical method for identifying homogenous cases and

grouping these into clusters. Cases within clusters are closely related and share

many characteristics, while also being dissimilar to cases belonging to other

clusters.[343] Cluster analysis has been used to examine movement strategies

employed during human gait,[331, 349-352] as well as in identifying muscle

activation strategies and their subsequent effect on biomechanical

outcomes.[353, 354] A number of different clustering algorithms exist, with these

algorithms varying in their purpose, strengths and weaknesses. Different

clustering approaches were employed for analyses within this thesis. The

determination of the clustering approach was based on the aims of the study,

as well as the strengths and weaknesses of the clustering algorithms. The

following section details the strengths and weaknesses, and considerations of

use for the clustering algorithms used.

3.5.1.1 Density-Based Spatial Clustering (DBSCAN)

Density-based spatial clustering of applications with noise (DBSCAN) is a

clustering algorithm that groups data points that are close together (i.e. points

with many neighbours in high density regions), while marking data points

that lie in low-density regions as outliers (i.e. anomalies). The DBSCAN

algorithm requires three input parameters, those being: the data being

clustered; Epsilon (ε; the distance radius required for two points to be

considered neighbours) and the minimum number of points reachable by a

data point (minPts) using the distance ε for it to be considered part of a

cluster.[355] The latter two of these parameters must be determined based on the

input data and aims of the clustering. The minimum number of points

Methodology

81

reachable by a data point for it to be considered part of a cluster (i.e. minPts)

can be chosen based on the goals of the analysis. For example, isolated data

points with no reachable neighbours could be considered as outliers or

anomalies. In this case, the minPts parameter would be set to one and data

points would not be considered as outliers or anomalies if they had at least

one neighbouring data point. Analytical approaches can be used to calculate

the ε parameter. Daszykowski et al.[355] and Karami and Johansson[356] present

an equation that can be used to calculate ε based on the minPts parameter and

input data used for clustering. Utilising this analytical approach ensures an

objective ε parameter is chosen specific to the data being clustered.

The DBSCAN algorithm assesses each point in a dataset using these

parameters, with data points allocated to clusters when they are within a

distance of ε to the minPts (see Figure 3.22, Page 82).[355] If a point does not

meet these conditions, it is labelled as an outlier.[355, 357] This is the main strength

of clustering using the DBSCAN algorithm. Other clustering approaches are

sensitive to outliers and include these data points within clusters.[343] As the

aim of study one (i.e. Chapter Four) was to identify anomalies in participant

data, the DBSCAN approach was considered appropriate for this study. The

specific application of the DBSCAN algorithm to this data is outlined in the

relevant chapter.

Methodology

82

Figure 3.22 Illustration of the DBSCAN cluster analysis algorithm using a

minimum points value of one. The radius of the large circles

represent Epsilon (i.e. the distance radius required for two

points to be considered neighbours). The green and blue points, respectively, are within this distance to the number of

minimum points required and hence form clusters. The grey

point is not within this distance to any points and is thus considered an outlier.

3.5.1.2 Latent Profile Analysis (LPA)

Latent profile analysis (LPA) is a multivariate, model-based clustering

technique using a general approach of finite mixture modelling.[358] Model-

based clustering techniques assign data points to clusters based on their fit to

a given mathematical model, with the most common used being a Gaussian

mixture model.[343] The Gaussian clustering model is based on the assumption

that the distribution from the examined data is generated by multiple

probability distributions,[343, 359] with the clustering algorithm aiming to detect

these underlying distributions. In a neuromechanical context, the idea of

Gaussian mixture modelling is that different neuromuscular or biomechanical

Methodology

83

strategies can be represented by a specific probability distribution, among the

overall probability distribution of the data.[343, 359] Hence, neuromuscular or

biomechanical strategies can be presented as a cluster with its own probability

distribution.[343] Maximum likelihood estimation, employing the Expectation-

Maximisation algorithm,[359] is the most popular and common method for

identifying these underlying probability distributions. The Expectation-

Maximisation algorithm starts by generating n random probability

distributions, and subsequently calculates the probability of membership for

each data point to one of the n estimated probability distributions.[343] The

calculated probability of membership is used to update (or re-estimate) the

mean and standard deviation of the generated n probability distributions.[343]

This process is repeated until the mean and standard deviation of the

probability distributions converge,[359] with this being the final clustering

solution.

Determining the optimal number of clusters or profiles is an important step

when employing LPA. A common approach is to fit models with a different

number of components (i.e. clusters or profiles) to the data, and evaluate the

fit of each model using evaluation criterion.[360] The Akaike (AIC) and Bayesian

(BIC) information criterion are the most commonly used indices for model

evaluation.[358, 360] The AIC and BIC compare the fit of a range of statistical

models to data based on maximum likelihood estimates.[360] Lower AIC and

BIC values indicate a better fit, therefore the optimal number of model

components can be determined by finding the minimum AIC and BIC

values.[358, 360] The ability to utilise these information criterion as a measure of

model fit and accurately estimate the number of clusters within the data is a

key strength of LPA and model-based clustering.[360] The aim of study two (i.e.

Chapter Five) was to cluster landing trials into sub-groups based on the

Methodology

84

muscle activation profiles employed. No previous studies have examined

muscle activation profiles in this manner, therefore no direction as to the

number of profiles to expect or extract was available. The ability to use

objective metrics to quantify the number of profiles to be extracted from the

data was a key reason for LPA being considered appropriate for this study.

Further, LPA was recently used to identify biomechanical profiles used by

athletes during a side-step cutting manoeuvre.[358] The specific application of

the LPA algorithm is outlined in the relevant chapter.

3.5.2 One-Dimensional Statistical Parametric Mapping (SPM1D)

One-dimensional statistical parametric mapping (SPM1D) is a statistical

technique designed specifically for continuous field analysis.[361] Within

biomechanical research, EMG and kinematic/kinetic data are often manifested

as one-dimensional (1D) scalar trajectories (i.e. continuous waveforms) in the

form of ; where represents a particular muscle’s activation level or a

given joint rotation/moment, and represents 1D time or space.[362] However,

these data are often examined discretely by extracting summary metrics from

specific points or regions of the waveform (e.g. peak muscle activation, peak

joint angles).[361] As these discrete measures only encapsulate a small portion

of the waveform, this regional focus bias can result in important information

relating to other points in the data to be missed. SPM1D is designed

specifically for non-directed hypothesis testing of continuous scalar

trajectories, and can partially offset the limitations of discrete point analyses

by providing a framework for the continuous statistical analysis of waveform

data.[361, 363] Another potential source of bias that can be overcome by SPM1D is

covariance among the components being examined.[362] Separate analysis of

individual vector components (e.g. sagittal, frontal and transverse plane joint

rotations in isolation) assumes independence between the components, which

Methodology

85

may be unjustified in particular instances.[362] For example, joint rotations are

unlikely to be independent as muscle lines of action are generally not parallel

to defined joint axes.[362] This inter-component covariance bias can be

overcome by using SPM1D to examine data as a multi-component vector-

field.[362] SPM1D has been highlighted as a superior technique for non-directed

biomechanical hypothesis testing by overcoming the potential regional focus

and inter-component covariance biases associated with a priori selection of

specific points within a waveform.[362] Recent studies have utilised SPM1D to

identify variations in landing strategies between sub-groups.[364, 365] SPM1D has

also been used to identify movement strategies that predict high-risk knee

joint moments during an unplanned side-step cutting manoeuvre.[366]

SPM1D results in the calculation of an SPM statistic, a continuous scalar

trajectory represented as a function of time.[361, 362, 367] It is worth noting that

‘SPM1D’ refers to the overall methodological approach, while the SPM statistic

refers to the calculated scalar trajectory variable.[367] The scalar trajectory

indicates the magnitude of difference or relationship between the waveforms

of interest.[367] Random field theory[368] is used to assess the field-wide

significance of an SPM statistic.[361] To accept or reject the null hypothesis, a

critical threshold at which only (e.g. 5% for an alpha level of 0.05) of

smooth random curves would be expected to traverse is calculated.[367]

Calculation of this critical threshold is based upon estimates of trajectory

smoothness via temporal gradients, and based on this smoothness, random

field theory expectations regarding the field-wide maximum.[361, 363, 367, 368] As

with traditional hypotheses tests, SPM1D produces type I error at a rate of

alpha,[367] thus the critical threshold level must be corrected when multiple

comparisons are made. While SPM1D does not prevent type I error, it does

tightly control its rate of occurrence.[367] Statistical significance, and thus

Methodology

86

rejection of the null hypothesis, is specified by the SPM statistic exceeding the

calculated critical threshold.[367] Multiple adjacent points of the curve often

exceed the critical threshold, with these areas termed as ‘supra-threshold

clusters.’[367] Cluster specific p-values can then be calculated using random

field theory expectations regarding supra-threshold cluster size (see Figure

3.23).[367, 368]

Figure 3.23 Examples of SPM curves where: A) both the upper and lower

critical thresholds have been exceeded, resulting in two supra-

threshold clusters; B) only the upper critical threshold has been exceeded, resulting in one supra-threshold cluster; and C)

neither the upper or lower critical thresholds have been

exceeded, resulting in no supra-threshold clusters and signifying acceptance of the null hypothesis. Reproduced with

permission from Pataky et al.[362]

SPM1D supports both parametric and nonparametric hypothesis testing,[362, 363]

and can be applied to test both difference (e.g. detecting differences between

two waveforms) and relationship (e.g. examining the relationship between a

predictor variable and a continuous waveform) directed hypotheses.[361, 362]

SPM1D has equivalents to the majority of traditional hypothesis tests used

within biomechanical research; including univariate and multivariate t-tests,

linear regression, analysis of variance (ANOVA), and multivariate analysis of

variance (MANOVA). For univariate (i.e. single-field) SPM1D tests, the 1D

Methodology

87

scalar trajectory (e.g. frontal plane knee moment) is used as the input for the

relevant test. Multivariate (i.e. vector-field) SPM1D tests combine the 1D scalar

trajectories of interest to form the multi-component vector-field.[362] For

example, sagittal (MX), frontal (MY) and transverse (MZ) plane knee moment

vectors are combined to represent the knee moment vector-field MXYZ.[366]

The outputs obtained and interpretation of results is similar across tests. Each

test outputs a continuous scalar trajectory that lie directly in the original 1D

sampled continua.[361] Statistical significance is identified as per the methods

discussed above; where a critical threshold based on a specified alpha level is

calculated, and where the scalar trajectory exceeds this threshold (i.e. supra-

threshold clusters) the null hypothesis is rejected.[361, 362, 367, 369] The choice of

SPM1D statistical test, like with traditional hypotheses testing, is based on the

nature of the hypothesis being examined. Where the aim is to examine

differences between two waveforms, the single- or vector-field SPM1D

equivalent of a t-test (paired or two-sample) can be conducted. Supra-

threshold clusters resulting from this test indicate a statistically significant

difference between the two waveforms being tested at the points along the 1D

continua where these clusters lie.[361, 362, 367] SPM1D linear regression or

canonical correlations analysis (CCA) can be used to determine the strength of

the relationship between a discrete variable at specific points along a

continuous 1D or vector-field trajectory, respectively.[362, 366, 370] In this case; a

statistically significant increase in the waveform is predicted by an increase or

decrease in the discrete variable where the SPM statistic exceeds the upper or

lower critical thresholds, respectively, along the 1D continua. A key advantage

of SPM1D is the ease of which statistical outputs can be interpreted. The SPM

statistic lies directly in the original 1D sampled continua, meaning it is the

same length as the original waveforms and is represented on an identical time-

Methodology

88

scale.[361] This allows investigators to easily identify where differences or

relationships lie along the waveforms by examining the supra-threshold

clusters against the originally sampled time-scale.

It is important to acknowledge the limitations associated with SPM1D post-

hoc analyses. Post-hoc analyses of waveform pairs assumes the post-hoc tests

are independent, which is usually not the case, and also involves separate

smoothness assessments for each test.[371] Limitations also exist for post-hoc

analyses of multivariate SPM1D tests. Where a statistically significant result is

found from a vector-field test, post-hoc analysis of the individual vector

components can be undertaken using single-field SPM1D tests.[362, 371] This

approach may indicate which vector components contribute to the statistically

significant result, however it neglects the inter-component covariance between

vectors.[371] While the resulting errors associated with these limitations are

expected to be small,[371] interpretation of post-hoc results from individual

waveform pairs and vector components should be made with caution.

Although these limitations exist, the advantages and recent work highlighting

the applicability of SPM1D in examining biomechanical research

hypotheses[361, 362, 364, 366, 367, 370] highlights this method as a useful tool for the

analysis of complex neuromuscular and biomechanical datasets.[362]

89

Chapter Four: Intra-Individual Anomalies in Muscle Activation during a Landing Task

4.1 Introduction

A perplexing factor associated with noncontact anterior cruciate ligament

(ACL) injuries is the ‘one-off’ nature in which they occur, where injuries are

often sustained during sporting tasks athletes have replicated safely countless

times.[46] The manner with which these injuries occur suggests a random,

multifaceted intrinsic dysfunction may play a role in injury events. Injuries to

the ACL occur when forces applied to the ligament are greater than the loads

it can withstand.[158] An athlete’s biomechanical technique during high-risk

sporting tasks directly influence the external loads applied to the lower

limb.[149, 164, 304, 372] Subsequently, the musculature surrounding the hip and knee

must act to stabilise the knee joint and provide support against these external

loads,[170-172] minimising the forces applied to soft tissue structures (e.g. the

ACL). ACL injuries likely occur when there is a mismatch between external

loading and muscular support. Therefore, both biomechanical technique and

activation profiles of lower limb muscles are vital in protecting the ACL

during high-risk sporting tasks. Despite both biomechanical and

neuromuscular factors playing a role in the ACL injury events, this study

focuses on the potential neuromuscular aspect of this dysfunction.

Given the high frequency with which high-risk tasks are performed within

competitive sporting environments,[156, 157] the muscles of the lower limb must

repeatedly function in an appropriate manner during these tasks to prevent

ACL injuries from occurring. Research has shown inter-trial variability in

Neuromuscular Anomalies during Landing

90

muscle activation characteristics across a range of tasks.[48-50, 254] This suggests

that a number of muscle activation profiles may be utilised during the

performance of a given movement. Previous studies have identified

neuromuscular risk factors for ACL injury by linking characteristics from an

individual’s mean muscle activation profile to future ACL injuries or lower

limb biomechanics known to increase ACL loads.[131, 174, 177, 178, 186-191] These

approaches fail to acknowledge the inherent variability present in the

neuromuscular system and the potential relationship of this to the ‘one-off’

nature in which noncontact ACL injuries occur. It is conceivable that a random

intrinsic dysfunction within the neuromuscular system is the inciting event for

these ‘one-off’ injury occurrences. Deviations from an individual’s usual or

‘normal’ muscle activation profile (i.e. intra-individual neuromuscular

anomalies) during high-risk sporting tasks may embody this dysfunction, and

could induce changes in lower limb biomechanics that place the ACL at

greater risk of injury. The concept of neuromuscular ‘anomalies’ and their

impact on lower limb biomechanics and ACL injury risk is yet to be examined.

No studies have determined whether intra-individual anomalies in muscle

activation profiles occur during repeat performance of a high-risk sporting

task, and if so, how these anomalies impact on lower limb biomechanics

associated with increased ACL loading and injury risk.

The purpose of this study was to determine whether intra-individual

anomalies in muscle activation profiles occurred during repeat performance

of a high-risk landing task, and if so, whether these neuromuscular anomalies

influenced lower limb biomechanics. It was hypothesised that intra-individual

anomalies in muscle activation profiles would occur, and that anomalous

activation profiles would result in lower limb biomechanics consistent with


91

increased ACL loads or injury risk when compared to the remaining typical

(i.e. ‘normal’) trials.

4.2 Methodology

The majority of the methods employed in this study are comprehensively

outlined in Chapter Three. The following methods are an abridged version of

the sections specific to this chapter.

4.2.1 Participants

Thirty-two healthy sub-elite female netball players (age = 23.1 ± 3.0 years;

height = 171.4 ± 7.8 cm; mass = 67.8 ± 8.2 kg; netball experience = 12.45 ± 4.54

years) volunteered and provided written consent to participate in this study.

All participants met the inclusion/exclusion criteria detailed in Section 3.2 of

Chapter Three.

4.2.2 Experimental Procedures

Data from the netball-specific landing task (i.e. the leap landing) as described

in Section 3.3.1 of Chapter Three, performed during testing session one (i.e.

baseline) were used within this study. Participants performed 20 successful

trials of the leap landing on a predetermined ‘landing limb.’ The landing limb

was determined by asking participants which limb they would be more

comfortable landing on when catching a pass during a game. Further to this,

players were given an opportunity to trial landing on each limb to identify

which they were more comfortable landing on. Each landing trial was initiated

by breaking and accelerating from a static defender (model skeleton) from a

position six-metres from the landing area, followed by a run-up and leap

landing whilst catching a pass (see Figure 4.1, Page 92). Trials were repeated


92

until the 20 successful trials were recorded. Participants were allocated 60-90

seconds rest between trials in an effort to minimise fatigue.[134, 255]

Electromyography, kinematic and kinetic data collected from the participants

‘landing limb’ during all landing trials were used within this study.


4.2.3 Data Collection and Analysis

Electromyography (EMG), kinematic and kinetic data from the leap landing

task were collected and analysed as per the procedures outlined in Section

3.4.1 and 3.4.2, respectively, of Chapter Three. Briefly, surface EMG data were

collected unilaterally from eight lower limb muscles (gluteus maximus

[GMAX], gluteus medius [GMED], medial hamstrings [MHAM], lateral

hamstrings [LHAM], rectus femoris [RF], vastus lateralis [VL], vastus medialis

[VM], and lateral gastrocnemius [GAS]) on the participants’ ‘landing limb’

using wireless parallel-bar electrodes (Delsys Trigno, Delsys Incorporation,

Boston, MA, United States). All EMG data were processed using custom

MATLAB (version R2014a) (The Mathworks Incorporated, Natick, MA,

United States) programs. Raw EMG signals were online processed using a

bandpass filter between 10 and 500 Hz, and later offline processed using a 20

Hz high-pass fourth-order zero-lag Butterworth digital filter to reduce the

effects of movement artefact; full-wave rectified; and converted to a linear

envelope via a 6 Hz low-pass fourth-order zero-lag Butterworth digital filter.

The EMG data was then amplitude normalised (% MVC) to the larger of the


93

peak value recorded during maximal voluntary isometric contractions or the

peak EMG activity during landing trials.[282, 288] EMG data were extracted from

150 ms prior through to 300 ms post initial contact (IC) across all landing trials

for subsequent analyses. This window was chosen as previous research

examining a similar single-limb catch and land task has shown lower limb

muscle onsets occur up to 150 ms prior to landing and remain active for a

period of up to 380 ms.[54]

This study used muscle synergies to represent muscle activation profiles

during the leap landing task. The extracted EMG data were factorised by

applying a non-negative matrix factorisation (NMF) algorithm[298] (as outlined

in Section 3.4.1.3.1 of Chapter Three) to obtain the muscle synergy

components, those being the time-variant muscle activation coefficients and

time-invariant muscle weighing coefficients. The NMF algorithm was iterated

by varying the number of synergies between one and six, with the number of

synergies that resulted in greater than 90% mean total variance accounted for

( ) between experimental and reconstructed EMG data retained. was

also calculated across individual muscles ( ) to ensure patterns were

accounted for by the retained synergies (i.e. > 75%).[48] Following

factorisation, the muscle activation coefficients were normalised to their

maximum value (i.e. each activation coefficient had a maximum value of 1),

with the muscle weighing coefficients within each of the muscle synergies

scaled by the same respective factor.[49, 294]

Three-dimensional (3D) kinematics and kinetics of the lower limb were

computed using an established musculoskeletal model.[305, 310] An eight camera

3D motion capture system (Vicon, Oxford Metrics Limited, Oxford, United

Kingdom) sampling at 250 Hz, synchronised with a 600 x 900 millimetre in

ground force plate (Advanced Medical Technology Incorporated, Watertown,


94

MA, United States) sampling at 1000 Hz collected marker trajectory and

ground reaction force (GRF) data. Marker trajectory and GRF data were low-

pass filtered using a cubic smoothing spline with a cut-off frequency of 12 Hz.

Three-dimensional hip and knee joint rotations were calculated from filtered

marker trajectory data using an X-Y-Z Cardan angles sequence; with the X, Y

and Z axes representing joint rotations in the sagittal, frontal and transverse


participant’s stationary (neutral) trial.[56, 310, 311] Three-dimensional hip and knee

external joint moments were calculated by submitting the filtered marker

trajectory and GRF data to a conventional inverse dynamics analysis.[261] Joint

moments were expressed in the reference frames of the joint coordinate

systems and were normalised by dividing the total joint moment (Nm) by the

product of the participant’s weight and height (kgm). Joint rotation and

moment data were extracted from the instance of IC to 150 ms post IC across

all landing trials. This time window was chosen for biomechanical variables

as ACL injuries are thought to occur within the early deceleration phase of

landing.[161]

4.2.4 Statistical Analysis

The muscle weighing coefficients of all muscle synergies from each

participants’ 20 leap landing trials were extracted into separate matrices for

further analyses. A density-based spatial clustering of applications with noise

(DBSCAN) algorithm (as outlined in Section 3.5.1.1 of Chapter Three)[357] was

applied to each participants’ data to identify trials that employed their typical

or usual (hereby referred to as ‘normal’) muscle activation profile(s), while

also identifying where outlier trials (i.e. intra-individual anomalies in muscle

activation profiles) were present. The DBSCAN algorithm grouped data points

that were close together (i.e. points with many neighbours in high-density


95

regions), while also marking data points that lied in low-density regions as

outliers (i.e. anomalies).[355, 357] Prior to clustering, the muscle weighing

coefficient values of each participant were normalised to their correlation

matrix, as this can result in better classification of waveform data when using

partitional classification techniques.[348] The correlation matrix was created by

calculating the Pearson’s r value for the extracted muscle weighing

coefficients, quantifying the relationship between individual landing trials.

The correlation matrix was subsequently used as the input data for the

DBSCAN algorithm. The additional parameters for the DBSCAN algorithm,

Epsilon (ε) and the minimum number of points reachable by a data point

(minPts) for it to be considered part of a cluster were determined using

established procedures.[355-357] The individual muscle weighing coefficient

correlation matrices from each participant were used as the data inputs in the

equation presented by Daszykowski et al.[355] and Karami and Johansson[356],

resulting in participant-specific ε parameters being calculated. This study

considered isolated data points with no reachable neighbours as anomalies,

therefore the minPts was set at one across all participants. Trials partitioned to

clusters were considered to represent a participant’s selective ‘normal’ muscle

activation profile or profiles, while trials labelled as outliers were deemed as

intra-individual anomalies in muscle activation.

Trials considered as ‘normal’ and anomalous were grouped, exported and

further analysed via open source code[369, 371] in Python 2.7 using Canopy 1.5

(Enthought Incorporated, Austin, TX, United States). One-dimensional

statistical parametric mapping (SPM1D) was used to compare mean hip and

knee joint angles and moments between ‘normal’ and anomalous trials. A

hierarchical process was used to identify multi- and single-planar differences

in lower limb biomechanics between ‘normal’ and anomalous trials. The mean


96

hip and knee joint rotation (RXYZ) and joint moment (MXYZ) vector-fields were

compared between groups using a SPM1D paired Hotelling’s T2 test. Post-hoc

comparisons of the joint rotation (RXY, RXZ, RYZ) and moment (MXY, MXZ, MYZ)

couples using paired Hotelling’s T2 tests and individual joint rotation (RX, RY,

RZ) and moment (MX, MY, MZ) components using paired t-tests between

groups were also conducted. As a paired statistical design was used to

compare ‘normal’ vs. ‘anomalous’ trials, participants who did not exhibit any

anomalous trials were excluded from these analyses. To retain a Type I family-

wise error rate of alpha = 0.05, a corrected threshold based on the number of

vector components (i.e. joints = 2, planes = 3) was calculated,[362] resulting in a

corrected alpha level of 0.0083. Subsequently, for each SPM statistic a critical

threshold was calculated at the corrected alpha level to determine where

statistically significant differences were present. A statistically significant

difference was identified where the SPM statistical curve exceeded this critical

threshold. Based on the corrected alpha level, a supra-threshold cluster

indicated that the same result would be produced by random curves in only

0.83% of repeated tests. The subsequent probabilities (i.e. p-values) with which

supra-threshold regions of the SPM statistical curve could have resulted from

repeated samplings of equally smooth random curves were then calculated.

4.3 Results

Using the criteria previously described in Chapter Three (i.e. overall >

90% and > 75%), three muscle synergies were retained from the

extracted EMG data (see Figure 4.2, Page 97). Inspection of the activation

coefficients associated with each muscle synergy (see Figure 4.3, Page 98)

indicated that the first synergy captured muscle activation around initial

contact (i.e. at landing), the second captured muscle activation after initial


97

contact (i.e. post landing), and the third captured muscle activation prior to

initial contact (i.e. preparatory activity).

A.

B.


number of extracted muscle synergies. B) relative to the number of extracted muscle synergies. GMAX – gluteus

maximus; GMED – gluteus medius; MHAM – medial hamstrings; LHAM – lateral hamstrings; RF – rectus femoris;

VL – vastus medialis; GAS – gastrocnemius.


98

Figure 4.3 Normalised activation coefficients of the three muscle

synergies extracted from the factorisation of EMG data. Dashed vertical line represents initial landing contact (IC).

The DBSCAN algorithm identified trials with anomalous muscle activation

profiles in the majority of participants (27 out of 32). Specifically; six

participants recorded one anomalous trial, 17 participants recorded two

anomalous trials, and four participants recorded three anomalous trials (see

Table 4.1, Page 99). This resulted in 52 anomalous from 640 total trials (8.1%).


99

Table 4.1 Summary of the number of 'normal' and anomalous trials across individual participants. Gray shading indicates

participants exhibiting no anomalous trials.

Participant (P) ‘Normal’ Trials (No. of Profiles Identified) Anomalous Trials

P1 18 (2) 2

P2 18 (2) 2

P3 17 (1) 3

P4 19 (2) 1

P5 18 (1) 2

P6 18 (1) 2

P7 19 (2) 1

P8 18 (1) 2

P9 18 (1) 2

P10 19 (1) 1

P11 19 (2) 1

P12 18 (1) 2

P13 17 (1) 3

P14 18 (1) 2

P15 17 (1) 3

P16 17 (1) 3

P17 18 (1) 2

P18 18 (1) 2

P19 20 (2) 0

P20 18 (1) 2

P21 18 (2) 2

P22 18 (1) 2

P23 18 (1) 2

P24 18 (2) 2

P25 20 (1) 0

P26 18 (1) 2

P27 20 (3) 0

P28 19 (1) 1

P29 20 (2) 0

P30 19 (1) 1

P31 20 (2) 0

P32 18 (2) 2

TOTAL 588 52


100

Due to the influence of approach speed on landing biomechanics (see

Appendix F, Page 391), approach speeds between the ‘normal’ and

‘anomalous’ trial groupings were compared (see Figure 4.4). An independent

sample t-test found no statistically significant difference for approach speed

between groups (p > 0.05), therefore it was concluded that any differences in

landing biomechanics between groups was not due to differences in approach

speed.

Figure 4.4 Mean and standard deviation of approach speeds across

'normal' and anomalous trials.

No statistically significant differences between ‘normal’ and ‘anomalous’ trials

where identified via SPM1D for any of the joint rotation (see Figure 4.5, Page

101) or joint moment (see Figure 4.6, Page 102) vector-field combinations or

individual vector components.


101

Figure 4.5 Mean hip and knee joint rotations of ‘normal’ and anomalous


landing task. FLEX – flexion; EXT – extension; ADD –

adduction; ABD – abduction; IR – internal rotation; ER – external rotation.


102

Figure 4.6 Mean hip and knee joint moments of ‘normal’ and anomalous


landing task. Mom. – moment; EXT – extension; FLEX – flexion; ABD – abduction; ADD – adduction; ER – external

rotation; IR – internal rotation.


103

4.4 Discussion

The ‘one-off’ nature in which ACL injuries occur suggests a random intrinsic

dysfunction may play a role in injury events. Given the neuromuscular

system’s role in protecting the lower limb from hazardous postures and loads,

‘anomalies’ in muscle activation profiles could leave athletes susceptible to

ACL injury. The purpose of this study was to determine whether intra-

individual anomalies in muscle activation profiles occurred during repeat

performance of a high-risk landing task, and if so, whether these

neuromuscular anomalies influenced lower limb biomechanics. The findings

of this study suggest that anomalies in muscle activation can occur when

performing repetitions of a high-risk landing task within laboratory-based

testing, however there is no consistent effect on lower limb biomechanics

when these occur.

Within this study, the majority of participants (27 out of 32) demonstrated at

least one anomalous trial across all of their landings (see Table 4.1, Page 99).

However, the presence of these anomalies in general did not induce changes

in lower limb biomechanics consistent with increased ACL loads or injury risk.

No differences were observed when comparing hip and knee joint rotations

and moments from the anomalous to the remaining ‘normal’ trials. It is well

documented that the neuromuscular system is highly redundant, meaning a

single movement task can be performed in many ways with a similar

result.[49, 200, 274, 373] Variability within the neuromuscular system has been

established across a range of movements, even when the task is kept

consistent.[49, 50, 274] The redundancy of the neuromuscular system,[373] allowing

multiple muscle activation solutions to be utilised for the same task may be a

factor in why no differences were observed between ‘normal’ and ‘anomalous’


104

trials. Contrary to the hypothesis, anomalies in muscle activation profiles may

have the capacity to protect athletes’ from ACL injury by maintaining

consistent task performance. Perturbations during high-risk sporting tasks

performed within competitive environments may play an important role in

ACL injuries.[103] Reacting to a stimulus has been frequently shown to influence

athlete’s biomechanics.[138, 171, 193, 302, 374, 375] Side-step cutting movements have

been associated with ACL injuries within competitive sporting

environments.[5, 102] These injuries may occur due to excessive knee loads

stemming from an unplanned or late decision to initiate the change of

direction movement in response to a sudden external stimulus, such as

another athlete or movement of the ball,[101, 247, 302] or as an instantaneous

manoeuvre to gain a better position within the playing area.[247, 376] Sports that

include a focus on controlling and moving the ball with the feet, such as soccer,

may also present an obstacle in the planning of change of direction

movements.[101] Landing movements are also often cited as the task being

performed when ACL injuries occur.[98, 99, 103] In many team sports where ACL

injuries are observed during landings, players may be required to pre-empt

subsequent movements to the landing prior to completing the movement.[103]

Noncontact ACL injuries have been observed during landings where athletes’

appear to alter the alignment of the trunk in preparation for subsequent

movements.[103] Changes in trunk position during high-risk sporting tasks

have been linked to ACL injuries[159, 377] and can increase joint moments at the

knee that place a load on the ACL.[166, 167, 372] These varying circumstances

experienced within competitive sporting environments may necessitate

alterations to the muscle activation profile to keep task performance consistent

and avoid high knee loads that could cause ACL injury. Neuromuscular

anomalies may therefore not represent a dysfunction of the neuromuscular

system, rather these may be required to adapt to the different external loading


105

environments presented during competitive sporting environments. As the

landing task examined within this study was kept consistent across trials and

did not encompass anticipatory factors found within competitive sporting

environments, these concepts cannot be examined. A focus on unplanned

high-risk sporting tasks with external perturbations may yield valuable

information surrounding the role of neuromuscular anomalies in modulating

ACL injury risk.

While the results of this study support the theory that consistent task

performance can be achieved with anomalies in muscle activation profiles, it

may be unwise to dismiss the potential role of neuromuscular anomalies in

ACL injuries. The ‘one-off’ nature in which noncontact ACL injuries occur,[46]

combined with the neuromuscular system’s role in counteracting potentially

hazardous lower limb postures and loads[170-172] does support the theory that a

random intrinsic neuromuscular dysfunction may play a role in the ACL

injury mechanism. Abnormal knee joint postures and loading places the ACL

at risk of rupture.[138, 378] Anomalous muscle activation characteristics could

potentially increase the risk of injury to the ACL by increasing the likelihood

of an abnormal and possibly hazardous performance occurring. As discussed

earlier, neuromuscular anomalies in certain situations may represent a

mechanism to keep task performance consistent and protect from injury.

However, in other situations where variations in muscle activation profiles are

not required, neuromuscular anomalies may invoke abnormal and potentially

hazardous lower limb postures that increase the risk of injury.

It may also be that specific contrasting anomalies in muscle activation induce

desirable and undesirable changes, respectively, with regard to lower limb

biomechanics associated with ACL loading or injury risk. For example,

increasing activation of the medial quadriceps and hamstrings is considered


106

an appropriate strategy for protecting the ACL by supporting against external

knee abduction moments.[161, 170] Anomalous muscle activation characteristics

representing reduced versus over activation of these medial knee muscles

could therefore have opposing effects, increasing and decreasing the load

placed on the ACL, respectively. Examination of participants’ individual

anomalies compared to their ‘normal’ activation profile(s) revealed a diverse

range of anomalies across different trials, muscles and participants.

Anomalous trials were identified via both increases and decreases in the

activation of muscles within the different muscle synergies. Further, large

variations in the timing of peak muscle activation also appeared to trigger the

identification of anomalous activation profiles. The identification of

anomalous activation profiles can be best explained by inspection of an

example of a participant’s ‘normal’ versus anomalous trials. The muscle

synergy weighing coefficient and EMG waveform data of a sample

participants (P22) ‘normal’ and anomalous trials are presented in Figure 4.7

(see Page 108) and Figure 4.8 (see Page 109), respectively. Figure 4.7 can be

used to determine where the anomalies in muscle activation lie relative to the

identified muscle synergies. For the first identified anomalous trial (indicated

in red), the main anomalies compared to the ‘normal’ profile (indicated in

black) are present in the muscle weighing coefficients of gluteus maximus in

synergy one; as well as the medial and lateral hamstrings, and gastrocnemius

in synergy three. These results would suggest that this trial had a substantial

increase in activation of the gluteus maximus around initial contact; while also

having large increase in medial and lateral hamstrings, and gastrocnemius

preparatory activation relative to ‘normal’ trials. For the second identified

anomalous trial (indicated in blue), the main anomalies can be seen in the

muscle weighing coefficients of the gluteus medius and lateral hamstrings in

synergies one and three, respectively. These results would suggest that this


107

trial had a substantial increase in gluteus medius activation around initial

contact and a large reduction in preparatory activation of the lateral

hamstrings. These interpretations can be confirmed by examination of the

EMG waveform data (see Figure 4.8), comparing the mean data from the

‘normal’ trials to data from the anomalous trials.


108

Fi

gure

4.7

In

divi

dual

par

ticip

ant (

P22)

exa

mpl

e of

mus

cle

syne

rgy

wei

ghin

g co

effi

cien

ts fr

om ‘n

orm

al’ a

nd a

nom

alou

s tr

ials

.


109

Figure 4.8 Individual participant (P22) example of muscle activation

waveform data from ‘normal’ and anomalous trials.

It is possible that the grouping of dissimilar anomalies for comparison to the

participants’ ‘normal’ trials influenced the ability to detect differences in lower

limb biomechanical outcomes. Opposing effects on lower limb biomechanics

compared to ‘normal’ trials may arise from contrasting anomalies. The

investigation of specific anomalies in muscle activation was, however, outside

the scope of this study. The experimental design limited the ability to isolate

and test the effect of specific anomalies in muscle activation on lower limb

biomechanics. Only 52 anomalous trials were observed across 640 total trials.

Splitting these trials based on the type of anomaly that occurred would result


110

in minimal numbers across the various specific anomalies, and it would be

difficult to draw conclusions from such small samples.

Musculoskeletal modelling and computer simulation methods may be a viable

approach for examining the influence of specific muscle activation anomalies

on lower limb biomechanics across various sporting situations where ACL

injuries occur. Studies have utilised modelling and simulation methods to

identify how alterations in muscle activation can impact human

movement.[379-381] The findings of the present study suggest that it may be

important to consider anomalies in muscle activation profiles within

modelling and simulation studies examining the relationship between

neuromuscular control and ACL injury risk during high-risk sporting tasks.

Simulation-based methods could allow for specific neuromuscular anomalies

to be investigated in isolation and replicated across a range of different

muscles, individuals or tasks. Environmental related perturbations that have

been associated with ACL injuries could also be included within these

simulations to understand the potential protective or injurious effects of

variations in muscle activation profiles. Future studies in these areas may

allow for a greater understanding of the differential effects of various

neuromuscular anomalies on lower limb biomechanics associated with ACL

loading during high-risk sporting tasks across a range of situations.

Interestingly, a select number of participants did not exhibit any anomalous

trials (see Table 4.1, Page 99). This could be an indication that certain

participants were more consistent in the activation profiles they used in

completing the landing task. Consistent muscle activation could ensure that

the landing task is repeatedly performed in a safe manner, preventing injuries

from occurring. However, if the individual’s muscle activation profile were

insufficient to produce a landing technique that protected the ACL from high


111

loads, this consistency may result in the individual being frequently placed at-

risk of injury. In addition, the inability to adapt their muscle activation profile

to varying situations may also place the individual at increased risk of injury

when unexpected perturbations to the movement are experienced. The need

to utilise consistent or varying muscle activation profiles may be an important

component to consider for ACL injury prevention programs. ACL injury

prevention programs incorporating repetitive exercises in a controlled

environment may only encourage a single neuromuscular solution for the

high-risk task being trained. Should more variable neuromuscular control be

deemed important for protecting athletes from ACL injury, injury prevention

programs must reflect this need. Further research on the impact of varying

muscle activation profiles on lower limb biomechanics during high-risk

sporting tasks may determine whether injury prevention programs should

encourage consistent or variable neuromuscular strategies.

The lack of anomalies for certain participants could also have been reflective

of the number of trials collected (n = 20). By definition, an anomaly is

something that deviates from what is normal or expected. Therefore, many

trials may be necessary for this to occur within individuals who display more

consistent muscle activation. This need for a large number of trials could be a

limitation for examining this phenomenon using data collected from

experimental methods alone. Collecting a substantial number of trials of a

high-risk sporting task within laboratory settings may be impractical due to

the time required, but may also impact the merit of results. Alterations in

muscle activation characteristics have been observed during high-risk

sporting tasks under fatigued conditions.[382] Performing an extensive number

of trials could induce fatigue, therefore it may be difficult to discern if the


112

anomalies in neuromuscular control were due to an intrinsic dysfunction or

the presence of fatigue.

The grouping of ‘normal’ muscle activation profiles may have been another

limiting factor in this study. While certain participants’ ‘normal’ trials were

grouped in a single muscle activation profile and could be considered

homogenous, other participants exhibited multiple ‘normal’ activation

profiles. This is not surprising, as studies have shown different neuromuscular

solutions can be utilised for performance of the same task.[48-50, 254] Grouping the

various ‘normal’ activation profiles identified and comparing each of these to

the anomalous trials may have resulted in certain differences in lower limb

biomechanics being identified. Nonetheless, the focus of this study was to

examine the impact of anomalies in muscle activation on lower limb

biomechanics during a high-risk landing task, and this was achievable by

grouping trials considered as ‘normal’ for comparison. It is important to note

that ‘normal’ muscle activation within this study was only representative of

the individuals’ typical or usual muscle activation profile(s). It may be

erroneous to assume that these ‘normal’ activation profiles incorporated

adequate muscle activation to prevent lower limb postures that would

increase ACL loading or injury risk. The identification of multiple ‘normal’

profiles does, however, open up further avenues for exploring neuromuscular

contributions to ACL injury risk. It is possible that different muscle activation

profiles result in different lower limb biomechanics, and thus impact on ACL

injury risk. Linking specific muscle activation profiles to lower limb

biomechanics consistent with increased ACL loads may further our

understanding of neuromuscular contributions to ACL injury risk. This area

will be explored within Chapter Five of this thesis.


113

It is important to acknowledge that this study focused solely on muscle

activation during the high-risk sporting task. The physiological characteristics

and strength of muscles are likely to be key factors in protecting the knee and

ACL during high-risk landing tasks, and thus also require consideration in

identifying neuromuscular contributions to ACL injury risk. A recent

prospective study found that reduced isometric hip abduction and external

rotation strength was predictive of future noncontact ACL injury in

athletes,[198] highlighting the importance of adequate muscle strength in

reducing ACL injury risk. A muscle with a larger physiological cross-section

area will produce greater force at the same activation level compared to a

smaller muscle.[262] Similarly, higher muscle forces can be produced at the same

activation level if the muscle is operating at a more optimal length or

velocity.[262] Individuals with differences in muscle strength, size or length may

have variable capacities to cope with anomalies in muscle activation profiles.

Neuromusculoskeletal modelling approaches allow the inclusion of muscle

strength and length parameters, while also calculating muscle forces during

task performances. The use of these approaches in conjunction with the

assessment of neuromuscular anomalies may assist in achieving a greater

overall understanding of neuromuscular contributions to ACL injury risk, and

how this may vary across individuals with different muscle physiology.


114

4.5 Conclusions

This study identified the presence of anomalies in muscle activation profiles

during repeat performance of a high-risk landing task within a laboratory

environment. However, the presence of neuromuscular anomalies did not

result in lower limb biomechanics consistent with increased ACL loads or

injury risk. These findings suggest that consistent performance of a high-risk

landing task can be achieved even with anomalies in neuromuscular control.

Investigating the differential effects of specific neuromuscular anomalies may

elicit further understanding regarding their influence on ACL injury risk.

Multiple ‘normal’ muscle activation profiles were identified within a number

of participants. A greater understanding of neuromuscular contributions to

ACL injury risk may be achieved by examining the influence of these different

normal profiles on lower limb biomechanics during high-risk landing tasks.

115

Chapter Five: The Influence of Muscle Activation Profiles on Lower Limb Biomechanics during a Landing Task

5.1 Introduction

Neuromuscular control is often acknowledged as a key modifiable risk factor

for anterior cruciate ligament (ACL) injuries.[15, 47] Various elements of

neuromuscular control have been implicated in increasing ACL injury risk or

linked to biomechanical factors associated with increased ACL loads.[131, 174, 177,

178, 186-191] Targeting these muscle activation characteristics within injury

prevention programs is likely an effective means for reducing ACL injury risk.

However, previous work linking neuromuscular factors to lower limb

biomechanics associated with ACL loading has tended to focus on discrete

muscle activation characteristics of specific muscles.[54-64] This approach fails to

acknowledge the relationship between multiple muscles and their combined

activation characteristics. As no single muscle crossing the knee is capable of

supporting the joint against injurious loads alone, adequate activation of

multiple muscles is required to protect the ACL from external loading during

high-risk sporting tasks.[161] The combined activation characteristics of lower

limb muscles (i.e. the overall muscle activation profile) may therefore be key

in reducing ACL injury risk. Understanding how the overall muscle activation

profile contributes to lower limb biomechanics associated with ACL loading

during high-risk sporting tasks may improve our understanding of

neuromuscular contributions to ACL injury risk.

Muscle Activation Profiles during Landing

116

Determining how the overall muscle activation profile varies across repeat

task performances within and between individuals is also of importance. The

neuromuscular system is capable of employing different muscle activation

profiles for a given movement.[48-50, 254] The use of different muscle activation

profiles within individuals was evident in the previous study, where a number

of participants employed multiple activation profiles across landing trials.

While the previous chapter focused on intra-individual variation within

muscle activation profiles, the use of muscle activation profiles across

individuals must also be considered. Understanding whether individuals

utilise shared or different activation profiles and how these relate to potential

ACL injury risk is an important factor in ensuring ACL injury prevention

programs are providing a benefit to all involved. Current ACL injury

prevention programs tend to be applied in a universal manner, where the

same neuromuscular deficits are targeted across all individuals.[358] The

identification of different activation profiles within or across individuals

would question whether generic ACL injury prevention programs are

appropriate. Different activation profiles may encompass various

neuromuscular deficits that contribute to ACL injury risk. This may infer to

need for varied injury prevention programs to ensure these neuromuscular

deficits are targeted.

The purpose of this study was to identify the muscle activation profiles

employed during a high-risk landing task and their subsequent impact on

lower limb biomechanics linked to ACL loading or injury risk. It was

hypothesised that individuals would utilise a number of different muscle

activation profiles, and that certain profiles would result in lower limb

biomechanics consistent with increased ACL loading and injury risk.


117

5.2 Methodology




5.2.1 Participants

Thirty-two healthy sub-elite female netball players (age = 23.1 ± 3.0 years;

height = 171.4 ± 7.8 cm; mass = 67.8 ± 8.2 kg; netball experience = 12.45 ± 4.54

years) volunteered and provided written consent to participate in this study.

All participants met the inclusion/exclusion criteria detailed in Section 3.2 of

Chapter Three.


Data from the netball-specific landing task (i.e. the leap landing), as described

in Section 3.3.1 of Chapter Three, performed during testing session one (i.e.

baseline) were used within this study. Participants performed 20 successful

trials of the leap landing on a predetermined ‘landing limb.’ The landing limb

was determined by asking participants which limb they would be more

comfortable landing on when catching a pass during a game. Further to this,

players were given an opportunity to trial landing on each limb to identify

which they were more comfortable landing on. Each landing trial was initiated

by breaking and accelerating from a static defender (model skeleton) from a

position six-metres from the landing area, followed by a run-up and leap

landing whilst catching a pass (see Figure 5.1, Page 118). Trials were repeated

until the 20 successful trials were recorded. Participants were allocated 60-90

seconds rest between trials in an effort to minimise fatigue.[134, 255]


118

Electromyography, kinematic and kinetic data collected from the participants

‘landing limb’ during all landing trials were used within this study.



Electromyography (EMG), kinematic and kinetic data from the leap landing

task were collected and analysed as per the procedures outlined in Section

3.4.1 and 3.4.2, respectively, of Chapter Three. Briefly, surface EMG data were

collected unilaterally from eight lower limb muscles (gluteus maximus

[GMAX], gluteus medius [GMED], medial hamstrings [MHAM], lateral

hamstrings [LHAM], rectus femoris [RF], vastus lateralis [VL], vastus medialis

[VM], and lateral gastrocnemius [GAS]) on the participants’ ‘landing limb’

using wireless parallel-bar electrodes (Delsys Trigno, Delsys Incorporation,

Boston, MA, United States). All EMG data were processed using custom

MATLAB (version R2014a) (The Mathworks Incorporated, Natick, MA,

United States) programs. Raw EMG signals were online processed using a

bandpass filter between 10 and 500 Hz, and later offline processed using a 20

Hz high-pass fourth-order zero-lag Butterworth digital filter to reduce the

effects of movement artefact; full-wave rectified; and converted to a linear

envelope via a 6 Hz low-pass fourth-order zero-lag Butterworth digital filter.

The EMG data was then amplitude normalised (% MVC) to the larger of the

peak value recorded during maximal voluntary isometric contractions or the

peak EMG activity during landing trials.[282, 288] EMG data were extracted from


119

150 ms prior through to 300 ms post initial contact (IC) across all landing trials

for subsequent analyses. This window was chosen as previous research




As in the previous study, muscle synergies were used to represent muscle

activation profiles during the landing task. The extracted EMG data were

factorised by applying a non-negative matrix factorisation (NMF)

algorithm[298] (as outlined in Section 3.4.1.3.1 of Chapter Three) to obtain the

muscle synergy components, those being the time-variant muscle activation

coefficients and time-invariant muscle weighing coefficients. The NMF

algorithm was iterated by varying the number of synergies between one and

six, with the number of synergies that resulted in greater than 90% mean total

variance accounted for ( ) between experimental and reconstructed EMG

data retained. was also calculated across individual muscles ( )

to ensure that each muscle’s activity patterns were accounted for by the

retained synergies (i.e. > 75%).[48] Following factorisation, the muscle

activation coefficients were normalised to their maximum value (i.e. each

activation coefficient had a maximum value of 1), with the muscle weighing

coefficients within each of the muscle synergies scaled by the same respective

factor.[49, 294]




Kingdom) sampling at 250 Hz, synchronised with a 600 x 900 mm in ground

force plate (Advanced Medical Technology Incorporated, Watertown, MA,

United States) sampling at 1000 Hz collected marker trajectory and ground


120

reaction force (GRF) data. Marker trajectory and GRF data were low-pass

filtered using a cubic smoothing spline with a cut-off frequency of 12 Hz.














landing.[161]


Latent profile analysis (LPA) (as outlined in Section 3.5.1.2 of Chapter Three)

was used to classify landing trials with similar muscle activation

characteristics into subgroups (i.e. muscle activation profiles). The normalised

muscle weighing coefficients obtained from the factorisation of EMG data

were used as the input variables for the LPA. These variables represented the

muscle activation characteristics of each landing trial and were therefore

deemed as the most applicable for profile identification. Models with a

different number of profiles (two through to eight) were fit to establish the

number of subgroups that provided the optimal fit for the data.[360] The Akaike

(AIC) and Bayesian (BIC) information criterion indices were used for model


121

selection.[358, 360] Based on these criterion the three component model was

deemed to provide the best fit, therefore three muscle activation profiles

(MAPs) were identified. Although a higher number of profile solutions were

statistically supported by the AIC criterion, solutions above three profiles

resulted in substantial increases in BIC suggesting a poorer fit (see Table 5.1).

Solutions above the three component model also resulted in profiles being

split into subgroups containing a small number of trials (i.e. less than 30 trials).

Table 5.1 Information criterion indices used for model selection in the

latent profile analysis.

Model Components AIC BIC

2 40,398 43,972

3 39,627 43,294

4 38,940 44,735

5 38,232 45,478

6 38,217 46,912

7 38,153 48,298

8 37,795 49,391

AIC – Akaike Information Criteria; BIC – Bayesian Information Criteria Lower values indicate a superior model

Following profile identification, group differences in the muscle weighing

coefficients across all muscle synergies were assessed. Normality of data

distribution within profiles and equality of variance between profiles for the

muscle weighing coefficients was confirmed via non-significant (p > 0.05)

Shapiro-Wilk and Levene statistics, respectively. As the assumptions of

normality and equality of variance were not violated, within- and between-

profile comparisons across the factors of interest were undertaken using

parametric tests. A three-way factorial analysis of variance (ANOVA) with the


122

muscle weighing coefficient values as the dependent variable was used to

examine the main and interaction effects of profile (n = 3), muscle synergy (n =

3) and muscle (n = 8). Partial eta squared (η2) was used as a measure of effect

size for main and interaction effects, with the interpretation of effect size

values guided by Miles and Shevlin[383] (small η2 = 0.01; medium η2 = 0.06; large

η2 = 0.14). Where statistically significant main and interaction effects were

identified (p < 0.05), post-hoc analyses of simple main effects within- and

between-factors were undertaken using Tukey’s honest significant difference

tests to identify where specific differences occurred. A Bonferroni correction

was applied to an initial alpha level of 0.05 for post-hoc tests based on the

number of comparisons being made within a simple main effects test. Cohen’s

d effect sizes were calculated for post-hoc tests, with the interpretation guided

by Cohen[384] (small d = 0.2; medium d = 0.5; large d = 0.8).

To test the effect of different MAPs on lower limb biomechanics, extracted hip

and knee joint rotation and moment waveform data were exported and

analysed via open source code[369, 371] in Python 2.7 using Canopy 1.5

(Enthought Incorporated, Austin, TX, United States). Comparisons of

kinematic and kinetic waveform data across the different MAPs were made

using one-dimensional statistical parametric mapping (SPM1D). A

hierarchical process was used to identify multi- and single-planar differences

in lower limb biomechanics between MAPs. The mean hip and knee joint

rotation (RXYZ) and joint moment (MXYZ) vector-fields were compared across all

profiles using a SPM1D multivariate ANOVA (MANOVA). Where MANOVA

revealed a statistically significant difference, post-hoc comparisons of the joint

rotation (RXYZ) and moment (MXYZ) vector-fields between the individual MAPs

were made using a SPM1D Hotelling’s T2 test. Post-hoc comparisons between

the individual MAPs of the joint rotation (RXY, RXZ, RYZ) and moment (MXY,


123

MXZ, MYZ) couples using SPM1D Hotelling’s T2 tests and individual joint

rotation (RX, RY, RZ) and moment (MX, MY, MZ) components using SPM1D two-

sample t-tests were also conducted. To retain a Type I family-wise error rate

of alpha = 0.05, a corrected threshold based on the number of profiles and

vector components (i.e. profiles = 3; joints = 2; planes = 3) was calculated.[362]

This resulted in a corrected alpha level of 0.0028 being used for subsequent

critical threshold calculations and identification of statistically significant

differences. Subsequently, for each SPM statistic a critical threshold was

calculated at the corrected alpha level to determine where statistically

significant results were present. A statistically significant difference was

identified where the SPM statistical curve exceeded this critical threshold.

Supra-threshold clusters indicated the same result would be produced by

random curves in only 0.28% of repeated tests. The subsequent probabilities

(i.e. p-values) with which supra-threshold regions of the SPM statistical curve

could have resulted from repeated samplings of equally smooth random

curves were then calculated.

5.3 Results

As this study used the same data as Chapter Four, the same three muscle

synergies representing muscle activation around IC (i.e. at landing) (muscle

synergy 1), muscle activation after IC (i.e. post landing) (muscle synergy 2),

and activation prior to IC (i.e. preparatory activity) (muscle synergy 3) were

identified by NMF (see Figure 5.2, Page 124).


124

Figure 5.2 Normalised activation coefficients of the three muscle synergies extracted from the factorisation of EMG data.

Dashed vertical line represents initial landing contact (IC).

5.3.1 Muscle Activation Profiles

As described earlier, the latent profile analysis identified three muscle

activation profiles (MAPs). Of the 640 total trials; 231, 246 and 163 were

clustered to MAP-1, MAP-2 and MAP-3, respectively. A summary of the

number of trials using the different activation profiles by individual

participants is presented in Table 5.2 (see Page 125).


125

Table 5.2 Summary of muscle activation profiles employed by individual participants across trials.

Participant (P) MAP-1 MAP-2 MAP-3

P1 20 0 0

P2 6 1 13

P3 0 19 1

P4 0 1 19

P5 0 5 15

P6 4 1 15

P7 0 20 0

P8 20 0 0

P9 1 1 18

P10 2 18 0

P11 0 20 0

P12 1 19 0

P13 20 0 0

P14 20 0 0

P15 19 1 0

P16 0 0 20

P17 1 0 19

P18 11 0 9

P19 20 0 0

P20 20 0 0

P21 0 0 20

P22 0 19 1

P23 0 20 0

P24 17 0 3

P25 19 0 1

P26 0 20 0

P27 0 20 0

P28 0 20 0

P29 10 1 9

P30 20 0 0

P31 0 20 0

P32 0 20 0

MAP – muscle activation profile


126

Mean EMG data of the three MAPs from 150 ms prior to 300 ms post IC during

the leap landing task is presented in Figure 5.3.

Figure 5.3 Mean electromyography (EMG) waveform data of the three

muscle activation profiles (MAPs) from 150 milliseconds (ms)

prior to initial contact (IC) through to 300ms post IC during the leap landing task. Dashed vertical line indicates IC.

Statistically significant main effects were observed for the factors of profile (p

< 0.001, η2 = 0.008), synergy (p < 0.001, η2 = 0.322) and muscle (p < 0.001, η2 =

0.071). Post-hoc analyses revealed MAP-2 had greater overall activation

compared to MAP-1 (p < 0.001, d = 0.06) and MAP-3 (p < 0.001, d = 0.16), while

MAP-1 had greater overall activation compared to MAP-3 (p < 0.001, d = 0.10).


127

Greater overall activation was observed in muscle synergy one (i.e. at landing)

compared to muscle synergies two (i.e. post landing) (p < 0.001, d = 0.92) and

three (i.e. prior to landing) (p < 0.001, d = 1.10), while muscle synergy two

encompassed greater overall activation compared to muscle synergy three (p

< 0.001, d = 0.18). The highest overall activations were observed for the RF, VL,

VM and GAS compared to all other muscles (p < 0.001 and d = 0.16 – 0.54 across

comparisons). Greater overall activation of GMED and MHAM compared to

LHAM (p < 0.001, d = 0.26 and p < 0.001, d = 0.13, respectively) and GMAX (p <

0.001, d = 0.24 and p < 0.001, d = 0.12, respectively) was also observed.

A statistically significant synergy x muscle interaction was found (p < 0.001, η2

= 0.338). Post-hoc analyses revealed statistically significant differences in

overall activation between muscles within muscle synergies one (see Table 5.3,

Page 128) two (see Table 5.4, Page 129) and three (see Table 5.5, Page 130).


128


activation between muscles within muscle synergy one.

GMAX GMED MHAM LHAM RF VL VM GAS

GMAX ---

GMED n.d. ---

MHAM -0.90 -1.05 ---

LHAM -1.00 -1.17 n.d. ---

RF 0.59 0.47 1.41 1.51 ---

VL 0.71 0.58 1.62 1.74 n.d. ---

VM 0.68 0.56 1.55 1.66 n.d. n.d. ---

GAS 1.09 0.97 1.95 2.07 0.45 0.41 0.40 ---

Table reads along rows. Positive effect size indicates higher activation of muscle in row compared to respective muscle in column; while a negative effect size indicates reduced activation of muscle in row compared to respective muscle in column. p < 0.001 for all statistically significant differences. GMAX – gluteus maximus; GMED – gluteus medius; MHAM – medial hamstrings; LHAM – lateral hamstrings; RF – rectus femoris; VL – vastus lateralis; VM – vastus medialis; GAS – gastrocnemius; n.d. – no difference.


129


activation between muscles within muscle synergy two.


GMAX ---

GMED 0.36 ---

MHAM n.d. -0.43 ---

LHAM n.d. -0.32 n.d. ---

RF 1.52 1.26 1.60 1.51 ---

VL 1.36 1.03 1.48 1.35 -0.40 ---

VM 1.33 1.02 1.44 1.33 -0.38 n.d. ---

GAS n.d. n.d. n.d. n.d. -1.37 -1.18 -1.16 ---



130


activation between muscles within muscle synergy three.


GMAX ---

GMED 0.59 ---

MHAM 1.83 1.36 ---

LHAM 1.42 0.93 -0.41 ---

RF -0.88 -1.60 -2.73 -2.32 ---

VL n.d. -0.85 -2.13 -1.70 0.86 ---

VM n.d. -1.10 -2.33 -1.90 0.57 n.d. ---

GAS n.d. -0.82 -2.00 -1.60 0.57 n.d. n.d. ---


5.3.1.1 Within-Profile Comparisons

Statistically significant profile x synergy (p < 0.001, η2 = 0.048), profile x muscle

(p < 0.001, η2 = 0.018) and profile x synergy x muscle (p < 0.001, η2 = 0.046)

interactions were also observed.

5.3.1.1.1 Muscle Activation Profile 1 (MAP-1)

The muscle synergy weighing coefficient values for MAP-1 are presented in



131

Figure 5.4 Muscle synergy weighing coefficient values for MAP-1.

Within MAP-1, greater overall activation was observed in muscle synergy one

compared to synergies two (p < 0.001, d = 1.31) and three (p < 0.001, d = 1.30).

No overall differences were observed between muscle synergies two and

three. Greater overall activation of RF, VL and VM was recorded compared to

GMAX, GMED, MHAM, LHAM and GAS (p < 0.001 and d = 0.24 – 0.79 across


132

comparisons). No further differences in overall activation between muscles

were observed for MAP-1.

Within synergy one, reduced activation of MHAM and LHAM was observed

compared to all other muscles (p < 0.001 and d = 1.07 – 2.25 across

comparisons); while greater activation of RF, VM and GAS was observed

compared to GMAX (p < 0.001, d = 0.67; p < 0.001, d = 0.49; and p < 0.001, d =

0.41, respectively) and GMED (p < 0.001, d = 0.94; p < 0.001, d = 0.78; and p <

0.001, d = 0.66, respectively).

Within synergy two, greater activation of RF, VL and VM was observed

compared to GMAX, GMED, MHAM, LHAM and GAS (p < 0.001 and d = 1.03

– 1.64 across comparisons). No further differences for MAP-1 within synergy

two were identified.

Within synergy three, the MHAM had the greatest activation (p < 0.001 and d

= 0.47 – 2.12 across comparisons); while the LHAM had greater activation

compared to the remaining muscles (p < 0.001 and d = 1.01 – 1.81 across

comparisons). RF and GAS had reduced activation compared to GMAX (p <

0.001, d = 0.46 and p < 0.001, d = 0.44, respectively), GMED (p < 0.001, d = 0.91

and p < 0.001, d = 0.91, respectively) and VL (p < 0.001, d = 0.91 and p < 0.001, d

= 0.92, respectively); while reduced VM activation was seen compared to

GMED (p < 0.001, d = 0.79) and VL (p < 0.001, d = 0.77).





133



compared to synergies two (p < 0.001, d = 0.43) and three (p < 0.001, d = 0.89);

while synergy two had greater overall activation compared to synergy three

(p < 0.001, d = 0.48). RF, VL and VM had the highest overall activation (p < 0.001

and d = 0.25 – 0.72 across comparisons) within MAP-2. Greater activation of

GAS was found compared to GMAX (p < 0.001, d = 0.47), MHAM (p < 0.001, d


134

= 0.34) and LHAM (p < 0.001, d = 0.29); while GMED had greater overall

activation compared to GMAX (p < 0.001, d = 0.47). No differences in overall

activation between GMAX, MHAM and LHAM were observed for MAP-2.

Within synergy one, GAS had the greatest activation (p < 0.001 and d = 0.57 –

2.39 across comparisons); while VL and VM were found to have greater

activation than the remaining muscles (p < 0.001 and d = 0.51 – 1.80 across

comparisons). Reduced activation of MHAM and LHAM compared to GMAX

(p < 0.001, d = 0.84 and p < 0.001, d = 0.73, respectively), GMED (p < 0.001, d =

1.21 and p < 0.001, d = 1.10, respectively) and RF (p < 0.001, d = 1.11 and p <

0.001, d = 1.01, respectively) was observed. No further differences for MAP-2

within synergy one were identified.

Within synergy two, RF had the highest activation (p < 0.001 and d = 0.94 – 2.56

across comparisons); while VL and VM had greater activation than the

remaining muscles (p < 0.001 and d = 1.04 – 2.20 across comparisons). Reduced

activation of GMAX and MHAM compared to GMED (p < 0.001, d = 0.45 and

p < 0.001, d = 0.65, respectively), LHAM (p < 0.001, d = 0.17 and p < 0.001, d =

0.35, respectively) and GAS (p < 0.001, d = 0.40 and p < 0.001, d = 0.60,

respectively) was observed. No further differences for MAP-2 within synergy

two were identified.

Within synergy three, MHAM and LHAM had the greatest activation (p <

0.001 and d = 1.01 – 3.45 across comparisons); while GMED was found to have

greater activation than the remaining muscles (p < 0.001 and d = 1.11 – 2.47

across comparisons). Greater activation of GMAX was observed compared to

RF (p < 0.001, d = 1.16) and GAS (p < 0.001, d = 0.53). No further differences for

MAP-2 within synergy three were identified.


135



Figure 5.6.



136


compared to synergies two (p < 0.001, d = 1.30) and three (p < 0.001, d = 1.15).

No overall differences were observed between muscle synergies two and

three. GAS had greater overall activation compared to GMAX (p < 0.001, d =

0.60), GMED (p < 0.001, d = 0.31), MHAM (p < 0.001, d = 0.24) and LHAM (p <

0.001, d = 0.49) within MAP-3. Reduced overall activity of GMAX was observed

compared to GMED (p < 0.001, d = 0.36), MHAM (p < 0.001, d = 0.45), RF (p <

0.001, d = 0.51), VL (p < 0.001, d = 0.45) and VM (p < 0.001, d = 0.41); while

reduced activity of LHAM was observed compared to RF (p < 0.001, d = 0.38)

and VL (p < 0.001, d = 0.32). No further differences in overall activation between

muscles were observed for MAP-3.

Within synergy one, GAS recorded the highest activation (p < 0.001 and d =

0.49 – 2.24 across comparisons). Reduced activation of MHAM and LHAM

were observed compared to all other muscles (p < 0.001 and d = 0.48 – 1.65

across comparisons). RF, VL and VM had greater activation than GMAX (p <

0.001, d = 1.14; p < 0.001, d = 0.95 and p < 0.001, d = 0.89, respectively); while RF

and VL recorded greater activation than GMED (p < 0.001, d = 0.58 and p <

0.001, d = 0.45, respectively). No further differences for MAP-3 within synergy

one were identified.

Within synergy two, RF had the greatest activation (p < 0.001 and d = 0.49 –

2.09 across comparisons); while VL and VM were found to have greater

activation compared to the remaining muscles (p < 0.001 and d = 0.66 – 1.74

across comparisons). Greater activation of GMED and MHAM compared to

GMAX (p < 0.001, d = 0.71 and p < 0.001, d = 0.85, respectively) was observed.

No further differences for MAP-3 within synergy two were identified.


137

Within synergy three, MHAM had the greatest activation (p < 0.001 and d =

0.72 – 2.98 across comparisons); while LHAM had greater activation compared

to all remaining muscles (p < 0.001 and d = 0.72 – 2.55 across comparisons).

GMAX, GMED and GAS had greater activation than RF (p < 0.001, d = 1.65; p <

0.001, d = 1.83 and p < 0.001, d = 1.28, respectively), VL (p < 0.001, d = 1.00; p <

0.001, d = 1.29 and p < 0.001, d = 0.92, respectively) and VM (p < 0.001, d = 0.95;

p < 0.001, d = 1.24 and p < 0.001, d = 0.90, respectively). No further differences

for MAP-3 within synergy two were identified.

5.3.1.2 Between-Profile Comparisons

Comparisons of the muscle weighing coefficient values for the three profiles

across the three muscle synergies are presented in Figure 5.7 (see Page 139)

and Table 5.6 (see Page 140).

A number of differences were found between the profiles across the three

muscle synergies. Within synergy one, MAP-1 exhibited greater activation of

RF compared to MAP-2 (p < 0.001, d = 1.14) and MAP-3 (p < 0.001, d = 0.39);

while MAP-2 exhibited reduced activation of RF compared to MAP-3 (p <

0.001, d = 0.74). Within synergy two, MAP-2 exhibited greater activation of VL

and VM compared to MAP-1 (p < 0.001, d = 0.62 and p < 0.001, d = 0.74,

respectively) and MAP-3 (p < 0.001, d = 1.32 and p < 0.001, d = 1.52, respectively);

while MAP-1 exhibited greater activation of VL and VM compared to MAP-3

(p < 0.001, d = 0.55 and p < 0.001, d = 0.60, respectively). Within synergy three,

MAP-1 exhibited greater activation of VL compared to MAP-2 (p < 0.001, d =

0.61) and MAP-3 (p < 0.001, d = 1.58); while MAP-2 exhibited greater activation

of VL compared to MAP-3 (p < 0.001, d = 0.84).

Further, specific differences between the three muscle activation profiles were

observed across the muscle synergies. Compared to MAP-2 and MAP-3, MAP-


138

1 exhibited greater GMAX activity within synergy one (p < 0.001, d = 0.81 and

p < 0.001, d = 0.89, respectively), and reduced MHAM activity within synergy

two (p < 0.001, d = 0.50 and p < 0.001, d = 0.52, respectively). Compared to MAP-

1 and MAP-3, MAP-2 exhibited reduced MHAM activity within synergy one

(p < 0.001, d = 0.38 and p < 0.001, d = 0.35, respectively); greater GMED (p <

0.001, d = 0.53 and p < 0.001, d = 0.72, respectively), LHAM (p < 0.001, d = 0.77

and p < 0.001, d = 0.56, respectively), RF (p < 0.001, d = 1.80 and p < 0.001, d =

1.78, respectively) and GAS (p < 0.001, d = 0.80 and p < 0.001, d = 0.80,

respectively) activity within synergy two; and greater GMED (p < 0.001, d =

0.69 and p < 0.001, d = 0.63, respectively) and LHAM (p < 0.001, d = 0.59 and p <

0.001, d = 0.62, respectively) activity within synergy three. Compared to MAP-

1 and MAP-2, MAP-3 exhibited reduced GMAX activity within synergy two

(p < 0.001, d = 0.61 and p < 0.001, d = 0.88, respectively), reduced activity of VM

within synergy three (p < 0.001, d = 0.51 and p < 0.001, d = 0.81, respectively),

and greater GAS activity within synergy three (p < 0.001, d = 0.64 and p < 0.001,

d = 0.69, respectively).

139


Fi

gure

5.7

M

uscl

e sy

nerg

y w

eigh

ing

coef

fici

ent

valu

es

for

the

thre

e m

uscl

e ac

tivat

ion

prof

iles

(MA

Ps)

iden

tifie

d.

* - s

tatis

tical

ly s

igni

fica

nt d

iffe

renc

e be

twee

n al

l pro

file

s; #

- st

atis

tical

ly s

igni

fica

nt d

iffe

renc

e be

twee

n M

AP-

1 an

d

rem

aini

ng p

rofi

les;

¥ -

sta

tistic

ally

sig

nifi

cant

dif

fere

nce

betw

een

MA

P-2

and

rem

aini

ng p

rofi

les;

¢ -

sta

tistic

ally

si

gnif

ican

t dif

fere

nce

betw

een

MA

P-3

and

rem

aini

ng p

rofi

les.

140


Tabl

e 5.

6 M

ean

(± S

D) m

uscl

e sy

nerg

y w

eigh

ing

coef

fici

ent v

alue

s for

the

thre

e m

uscl

e ac

tivat

ion

prof

iles

(MA

Ps) i

dent

ifie

d.

Mus

cle

Mus

cle

Syne

rgy

1 (A

ctiv

atio

n ar

ound

IC)

M

uscl

e Sy

nerg

y 2

(Act

ivat

ion

afte

r IC

)

Mus

cle

Syne

rgy

3 (A

ctiv

atio

n pr

ior t

o IC

)

MA

P-1

MA

P-2

MA

P-3

M

AP-

1 M

AP-

2 M

AP-

3

MA

P-1

MA

P-2

MA

P-3

GM

AX

3.

67 ±

1.6

6 2.

39 ±

1.4

9 2.

39 ±

1.1

8

0.96

± 0

.93

1.14

± 0

.91

0.53

± 0

.37

1.

10 ±

1.0

2 1.

04 ±

0.7

2 1.

06 ±

0.6

8

GM

ED

3.22

± 1

.61

2.88

± 1

.46

3.13

± 1

.65

1.

10 ±

0.7

6 1.

58 ±

1.0

6 0.

93 ±

0.7

0

1.35

± 0

.73

1.88

± 0

.78

1.36

± 0

.85

MH

AM

1.

70 ±

1.2

0 1.

22 ±

1.2

9 1.

72 ±

1.5

7

0.70

± 0

.49

1.00

± 0

.71

1.02

± 0

.72

2.

74 ±

1.1

8 3.

02 ±

1.0

2 3.

04 ±

1.3

2

LHA

M

1.38

± 1

.13

1.35

± 1

.33

1.54

± 1

.28

0.

70 ±

0.4

8 1.

30 ±

0.9

8 0.

87 ±

0.4

5

2.23

± 0

.99

2.85

± 1

.12

2.18

± 1

.06

RF

4.85

± 1

.86

2.85

± 1

.64

4.14

± 1

.82

2.

03 ±

1.0

4 4.

44 ±

1.5

8 2.

09 ±

0.9

9

0.70

± 0

.70

0.38

± 0

.34

0.22

± 0

.25

VL

4.18

± 1

.53

3.93

± 1

.69

3.94

± 2

.00

2.

23 ±

1.2

7 2.

99 ±

1.1

8 1.

65 ±

0.8

3

1.27

± 0

.55

0.93

± 0

.58

0.50

± 0

.41

VM

4.

49 ±

1.6

6 3.

73 ±

1.8

4 3.

84 ±

1.9

8

2.21

± 1

.30

3.13

± 1

.17

1.55

± 0

.88

0.

80 ±

0.6

6 0.

92 ±

0.5

3 0.

50 ±

0.4

9

GA

S 4.

46 ±

2.1

6 4.

91 ±

1.7

6 5.

02 ±

1.7

8

0.85

± 0

.65

1.52

± 1

.00

0.84

± 0

.66

0.

73 ±

0.6

4 0.

67 ±

0.6

6 1.

34 ±

1.2

2

IC –

initi

al c

onta

ct; M

AP

– m

uscl

e ac

tivat

ion

prof

ile; G

MA

X –

glut

eus m

axim

us; G

MED

– g

lute

us m

ediu

s; M

HA

M –

med

ial h

amst

ring

s; LH

AM

– la

tera

l ham

stri

ngs;

RF –

rect

us fe

mor

is; V

L –

vast

us la

tera

lis; V

M –

vas

tus

med

ialis

; GA

S –

gast

rocn

emiu

s.


141

Due to the influence of approach speed on landing biomechanics (see

Appendix F, Page 391), approach speeds from trials clustered to the different

profiles were compared (see Figure 5.8). A one-way ANOVA revealed no

overall statistically significant difference for approach speed between the

profiles (p > 0.05), therefore it was concluded that any differences in landing

biomechanics between profiles was not due to differences in approach speed.

Figure 5.8 Mean and standard deviation of approach speed for the three

muscle activation profiles (MAPs).

The mean hip and knee joint rotations from the three MAPs from IC through

to 150 ms post IC during the leap landing task are presented in Figure 5.9 (see

Page 142). SPM1D MANOVA revealed a statistically significant (p < 0.0001)

overall difference across the three profiles for both the hip and knee joint

rotation vector-fields (RXYZ). Summary results of the pairwise comparisons

made between profiles for the hip and knee joint rotation vector-fields (RXYZ),

joint rotation couples (RXY, RXZ, RYZ) and individual joint rotation components

(RX, RY, RZ) are presented in Table 5.7 and Table 5.8 (see Pages 143 and 144),

respectively.


142

Figure 5.9 Mean hip and knee joint rotations of the three muscle

activation profiles (MAPs) from initial contact (IC) through to

150 milliseconds (ms) post IC during the leap landing task. FLEX – flexion; EXT – extension; ADD – adduction; ABD –

adduction; IR – internal rotation; ER – external rotation.


143

Table 5.7 Summary of differences identified by SPM1D post-hoc

analyses for hip joint rotations between the three muscle

activation profiles (MAPs).

Vector-Field Time Ranges p-values

Hip RXYZ MAP-1 vs. MAP-2 0 – 150ms p < 0.0001 MAP-1 vs. MAP-3 0 – 75ms; 90 – 129ms p = 0.0003; p = 0.0015 MAP-2 vs. MAP-3 0 – 30ms; 57 – 150ms p = 0.0017; p < 0.0001

Hip RXY MAP-1 vs. MAP-2 0 – 150ms p < 0.0001 MAP-1 vs. MAP-3 0 – 78ms p = 0.0004 MAP-2 vs. MAP-3 0 – 27ms; 57 – 150ms p = 0.0022; p < 0.0001

Hip RXZ MAP-1 vs. MAP-2 0 – 111ms; 144 – 150ms p < 0.0001; p = 0.0027 MAP-1 vs. MAP-3 24 – 72ms; 87 – 132ms p = 0.0010; p = 0.0011 MAP-2 vs. MAP-3 0 – 24ms; 123 – 150ms p = 0.0020; p = 0.0015

Hip RYZ MAP-1 vs. MAP-2 0 – 150ms p < 0.0001 MAP-1 vs. MAP-3 0 – 57ms; 96 – 120ms p = 0.0008; p = 0.0021 MAP-2 vs. MAP-3 0 – 24ms; 54 – 150ms p = 0.0020; p < 0.0001

Hip RX MAP-1 vs. MAP-2 45 – 81ms p = 0.0009 MAP-1 vs. MAP-3 27 – 78ms p = 0.0006 MAP-2 vs. MAP-3 0 – 15ms; 135 – 150ms p = 0.0013; p = 0.0013

Hip RY MAP-1 vs. MAP-2 0 – 150ms p < 0.0001 MAP-1 vs. MAP-3 0 – 48ms p = 0.0008 MAP-2 vs. MAP-3 45 – 150ms p < 0.0001

Hip RZ MAP-1 vs. MAP-2 0 – 45ms; 75 – 114ms p = 0.0004; p = 0.0005 MAP-1 vs. MAP-3 42 – 45ms; 93 – 123ms p = 0.0014; p = 0.0008 MAP-2 vs. MAP-3 0 – 15ms p = 0.0012

MAP – muscle activation profile; ms – milliseconds.


144


analyses for knee joint rotations between the three muscle



Knee RXYZ MAP-1 vs. MAP-2 0 – 150ms p < 0.0001 MAP-1 vs. MAP-3 0 – 150ms p < 0.0001 MAP-2 vs. MAP-3 0 – 150ms p < 0.0001

Knee RXY MAP-1 vs. MAP-2 0 – 150ms p < 0.0001 MAP-1 vs. MAP-3 0 – 150ms p < 0.0001 MAP-2 vs. MAP-3 0 – 111ms p < 0.0001

Knee RXZ MAP-1 vs. MAP-2 0 – 57ms; 93 – 150ms p = 0.0002; p = 0.0002 MAP-1 vs. MAP-3 0 – 150ms p < 0.0001 MAP-2 vs. MAP-3 0 – 150ms p < 0.0001

Knee RYZ MAP-1 vs. MAP-2 0 – 150ms p < 0.0001 MAP-1 vs. MAP-3 30 – 150ms p < 0.0001 MAP-2 vs. MAP-3 0 – 39ms; 54 – 81ms; 141 – 150ms p = 0.0004; p = 0.0010; p =

0.0025

Knee RX MAP-1 vs. MAP-2 0 – 30ms; 90 – 150ms p = 0.0009; p = 0.0002 MAP-1 vs. MAP-3 0 – 150ms p < 0.0001 MAP-2 vs. MAP-3 30 – 111ms p < 0.0001

Knee RY MAP-1 vs. MAP-2 0 – 150ms p < 0.0001 MAP-1 vs. MAP-3 27 – 150ms p < 0.0001 MAP-2 vs. MAP-3 0 – 39ms p = 0.0021

Knee RZ MAP-1 vs. MAP-2 0 – 63ms p < 0.0001 MAP-1 vs. MAP-3 114 – 150ms p = 0.0003 MAP-2 vs. MAP-3 0 – 30ms; 54 – 81ms; 138 – 150ms p = 0.0006; p = 0.0006; p = 0.0011



145

The mean hip and knee joint moments from the three MAPs from IC through

to 150 ms post IC during the leap landing task are presented in Figure 5.10 (see

Page 146). SPM1D MANOVA revealed a statistically significant (p < 0.0001)

overall difference across the three profiles for both the hip and knee joint

moment vector-fields (MXYZ). Summary results of the pairwise comparisons

made between profiles for the hip and knee joint moment vector-fields (MXYZ),

joint moment couples (MXY, MXZ, MYZ) and individual joint moment

components (MX, MY, MZ) are presented in Table 5.9 and Table 5.10 (see Pages

147 and 148), respectively.


146

Figure 5.10 Mean hip and knee joint moments of the three muscle

activation profiles (MAPs) from initial contact (IC) through to 150 milliseconds (ms) post IC during the leap landing task.

Mom. – moment; EXT – extension; FLEX – flexion; ABD –

abduction; ADD – adduction; ER – external rotation; IR – internal rotation.


147


analyses for hip joint moments between the three muscle



Hip MXYZ MAP-1 vs. MAP-2 0 – 150ms p < 0.0001 MAP-1 vs. MAP-3 0 – 150ms p < 0.0001 MAP-2 vs. MAP-3 0 – 120ms p < 0.0001

Hip MXY MAP-1 vs. MAP-2 0 – 150ms p < 0.0001 MAP-1 vs. MAP-3 0 – 75ms; 117 – 150ms p < 0.0001; p = 0.0002 MAP-2 vs. MAP-3 12 – 114ms p < 0.0001

Hip MXZ MAP-1 vs. MAP-2 0 – 150ms p < 0.0001 MAP-1 vs. MAP-3 15 – 48ms; 54 – 150ms p = 0.0002; p < 0.0001 MAP-2 vs. MAP-3 0 – 9ms; 15 – 99ms p = 0.0024; p < 0.0001

Hip MYZ MAP-1 vs. MAP-2 0 – 21ms; 42 – 150ms p = 0.0011; p < 0.0001 MAP-1 vs. MAP-3 0 – 48ms; 60 – 150ms p < 0.0001; p < 0.0001 MAP-2 vs. MAP-3 0 – 9ms; 27 – 87ms p = 0.0023; p < 0.0001

Hip MX MAP-1 vs. MAP-2 0 – 48ms p < 0.0001 MAP-1 vs. MAP-3 51 – 78ms; 129 – 150ms p = 0.0002; p = 0.0004 MAP-2 vs. MAP-3 12 – 96ms p < 0.0001

Hip MY MAP-1 vs. MAP-2 0 – 24ms; 48 – 150ms p = 0.0006; p < 0.0001 MAP-1 vs. MAP-3 0 – 48ms; 117 – 150ms p < 0.0001; p = 0.0002 MAP-2 vs. MAP-3 30 – 78ms p < 0.0001

Hip MZ MAP-1 vs. MAP-2 3 – 21ms; 42 – 150ms p = 0.0007; p < 0.0001 MAP-1 vs. MAP-3 15 – 45ms; 66 – 150ms p = 0.0002; p < 0.0001 MAP-2 vs. MAP-3 0 – 9ms; 24 – 90ms p < 0.0001; p = 0.0012



148


analyses for knee joint moments between the three muscle



Knee MXYZ MAP-1 vs. MAP-2 0 – 150ms p < 0.0001 MAP-1 vs. MAP-3 0 – 33ms; 45 – 150ms p = 0.0002; p < 0.0001 MAP-2 vs. MAP-3 0 – 69ms; 129 – 150ms p < 0.0001; p = 0.0007

Knee MXY MAP-1 vs. MAP-2 0 – 150ms p < 0.0001 MAP-1 vs. MAP-3 0 – 33ms; 51 – 150ms p = 0.0001; p < 0.0001 MAP-2 vs. MAP-3 15 – 69ms; 129 – 150ms p < 0.0001; p = 0.0005

Knee MXZ MAP-1 vs. MAP-2 0 – 90ms; 129 – 150ms p < 0.0001; p = 0.0007 MAP-1 vs. MAP-3 0 – 6ms; 57 – 102ms p = 0.0026; p < 0.0001 MAP-2 vs. MAP-3 0 – 69ms; 129 – 150ms p < 0.0001; p = 0.0006

Knee MYZ MAP-1 vs. MAP-2 0 – 150ms p < 0.0001 MAP-1 vs. MAP-3 0 – 30ms; 48 – 150ms p = 0.0004; p < 0.0001 MAP-2 vs. MAP-3 0 – 54ms p < 0.0001

Knee MX MAP-1 vs. MAP-2 15 – 72ms; 132 – 150ms p < 0.0001; p = 0.0004 MAP-1 vs. MAP-3 0 – 6ms p = 0.0013 MAP-2 vs. MAP-3 15 – 69ms; 126 – 150ms p = 0.0002; p < 0.0001

Knee MY MAP-1 vs. MAP-2 0 – 24ms; 42 – 150ms p = 0.0004; p < 0.0001 MAP-1 vs. MAP-3 0 – 33ms; 51 – 150ms p = 0.0002; p < 0.0001 MAP-2 vs. MAP-3 n.d. ---

Knee MZ MAP-1 vs. MAP-2 0 – 90ms p < 0.0001 MAP-1 vs. MAP-3 54 – 105ms p < 0.0001 MAP-2 vs. MAP-3 0 – 39ms p < 0.0001

MAP – muscle activation profile; ms – milliseconds ; n.d. – no difference.


149

5.4 Discussion

Various elements of neuromuscular control have been implicated in ACL

injury risk or linked to biomechanical factors associated with increased ACL

loads.[131, 174, 177, 178, 186-191] However, these studies have tended to focus on isolated

muscle activation characteristics rather than the combined activation

characteristics of multiple muscles (i.e. the muscle activation profile). The

purpose of this study was to identify the muscle activation profiles employed

during a high-risk landing task and their subsequent impact on lower limb

biomechanics linked to ACL loading or injury risk. A number of muscle

activation profiles were identified, with each profile contributing to the

presence of different biomechanical deficits associated with ACL loading or

injury risk.

The first hypothesis that a number of different muscle activation profiles

would be utilised during the landing task was confirmed. The final model

selected from the latent profile analysis consisted of three muscle activation

profiles (MAPs), with these profiles exhibiting a number of differences in

muscle activation. The magnitude of overall activation, irrespective of muscle

synergy or individual muscles, was found to be different between profiles.

Overall, MAP-2 exhibited the highest levels of activation, followed by MAP-1

and MAP-3. However, these differences had the smallest effect sizes across the

statistically significant comparisons within this study. Larger effect sizes were

observed for the differences within the muscle synergies and individual

muscles between profiles, suggesting profile separation was more so defined

by these specific muscle activation characteristics. MAP-1 was characterised

by elevated preparatory activation of the vastus lateralis, elevated activation

of the gluteus maximus and rectus femoris early after initial contact, and


150

reduced activation of the medial hamstrings during the latter stage of landing.

MAP-2 demonstrated high levels of gluteus medius and lateral hamstrings

activation prior to landing, and low levels of medial hamstrings and rectus

femoris activation early after initial contact. MAP-2 also exhibited higher

activations of multiple muscles in the later stages of landing, including the

gluteus medius, lateral hamstrings, quadriceps muscles, and gastrocnemius.

MAP-3 was characterised by low and high levels of preparatory activation of

the vastus medialis and gastrocnemius, respectively. Low levels of gluteus

maximus activation in the latter stage of landing were also observed within

MAP-3. The fact that these varying activation profiles were identified is not

surprising, as previous studies have found multiple neuromuscular solutions

can be utilised within a range of movement tasks.[48-50, 254]

Examining the MAPs at an individual level revealed that certain participants

utilised multiple activation profiles across their trials (see Table 5.2, Page 125).

While this occurred, it appeared that the majority of participants had a

preferred or ‘dominant’ activation profile. The utilisation of a single activation

profile across all 20 trials was the most common observation (17 out of 32

participants). Similarly, 12 out of 32 participants exhibited a ‘dominant’

activation profile, with at least 15 of the 20 trials being classified to this single

profile. Only three participants (i.e. P2, P18 and P29) demonstrated a more

balanced selection with regard to the activation profiles utilised. However, this

balance tended to be spread across two profiles, with a third rarely or never

being employed. It appears that even though various activation profiles can

be utilised within the landing task examined, individuals tend to persevere

with a preferred neuromuscular strategy. This has implications with regard to

the use of generic or universal neuromuscular training as a method to alter

muscle activation characteristics in an effort to reduce ACL injury risk.


151

Universal neuromuscular training refers to the prescription of an identical

injury prevention program across all individuals, and often attempts to target

specific neuromuscular or biomechanical risk factors associated with ACL

injury.[358] The application of universal neuromuscular training has been

proposed as an effective and efficient method for reducing ACL injury

rates.[226] However, the tendency for preferential neuromuscular strategies

demonstrated by participants within this study may highlight the need for

more targeted injury prevention programs. Varied programs targeting the

specific neuromuscular deficits within the individual’s preferential activation

profile may be more effective in eliciting changes in neuromuscular control

that reduce ACL injury risk.

The selective activation strategies of the musculature around the hip and knee

play an important role in counteracting potentially hazardous joint postures

and external loads.[170-172] The presence of inadequate or poor neuromuscular

control is likely to contribute to lower limb biomechanics that increase ACL

loads or injury risk during high-risk sporting tasks. It was hypothesised that

certain profiles would result in lower limb biomechanics consistent with

increased ACL loading and injury risk, representing a ‘high-risk’ activation

profile; while other profiles would not, representing a ‘safe’ activation profile.

However, this was not the case as various biomechanical deficits were

identified across all profiles. Reduced knee flexion and larger internal rotation

moments at the knee were associated with MAP-1; yet greater hip flexion and

smaller hip adduction and knee abduction angles were also present for this

profile. MAP-2 resulted in the largest hip adduction, hip internal rotation and

knee internal rotation angles, and highest knee abduction moments; but also

demonstrated large knee flexion angles at initial contact. Lastly, limited hip

flexion at initial contact and the largest knee abduction angles were seen with


152

MAP-3; however large knee flexion angles and the lowest knee abduction

moments were associated with this profile. The participants included within

this study were currently injury free and had no history of serious lower limb

injury. As such, the different profiles identified and resultant lower limb

biomechanics may represent normal and safe variations for the landing task

examined. However, considering lower limb biomechanics are the outcome of

a given movement, they provide a means for which to link neuromuscular

control with potential increased injury risk. While ‘high-risk’ and ‘safe’

activation profiles were not identified, investigating the biomechanical

outcomes stemming from the different profiles can assist in improving our

understanding of how the overall activation profile, or aspects of the activation

profile, contribute or prevent lower limb biomechanics consistent with

increased ACL loading or injury risk.

Frontal plane knee postures and moments are acknowledged as the primary

biomechanical risk factors for ACL injury.[20-23, 28, 160] A ‘dynamic valgus’ or

abducted knee position (i.e. abduction of the tibia relative to the femur) and

higher knee abduction moments have been linked to future ACL injury

occurrences[160] and substantially increase the load placed on the ACL during

landing movements.[20-23, 28] During netball leap landings, external knee

abduction moments are observed in the early stages after initial contact, with

these shifting to an external knee adduction moment later in the landing

phase.[248] Athletes must therefore employ adequate activation strategies at

specific periods of the landing phase to counteract these frontal plane

moments. Activation of muscles with appropriate moment arms has been

proposed as a method for counteracting external knee abduction/adduction

moments.[170, 171, 385] As the knee flexes the medial hamstrings have the capacity

to counteract external knee abduction moments, while the lateral hamstrings


153

and lateral gastrocnemius work to counteract external knee adduction

moments.[171] The selective activation of these muscles during the phases of the

landing where knee abduction and adduction moments are present may

therefore be an effective strategy in limiting their magnitude. This notion was

partially supported by the results of this study. MAP-2 exhibited the highest

peak knee abduction moments in the early stages of landing, with this profile

exhibiting lower levels of medial hamstrings activation around the time where

peak knee abduction moments occurred. Increasing the activation of the

medial hamstrings within this profile may have reduced the peak knee

abduction moments to a similar level to that seen with the other profiles. The

largest external knee adduction moments during the latter stages of landing

were seen with MAP-1. Surprisingly, this profile did not exhibit lower levels

of activation for muscles that counteract knee adduction moments, nor did it

exhibit higher levels of activation for muscles known to produce knee

adduction moments. Additional muscles, such as the medial gastrocnemius,

gracilis and sartorius, also have the capacity to counteract external knee

adduction moments.[171] The activation of these muscles may have been

insufficient within MAP-1, leading to the larger knee adduction moments.

However, this cannot be confirmed as these additional muscles were not

assessed in this study.

Preparatory muscle activation has been acknowledged as an important factor

in controlling frontal plane knee motion during landing tasks.[187] Higher peak

knee abduction angles have been associated with increased preparatory

activation of the vastus lateralis and lateral hamstrings.[187] In addition,

elevated preparatory activation of the vastus medialis has been linked to lower

peak knee abduction angles.[187] In the present study, both MAP-2 and MAP-3

recorded higher peak knee abduction angles compared to MAP-1. MAP-2


154

exhibited the highest preparatory activation of the lateral hamstrings. In

addition, low levels of medial hamstrings activation in the early stages of

landing and high levels of lateral hamstrings activation in the later stages of

landing were seen with MAP-2. The combination of these factors may have

impeded the ability of this profile to limit knee abduction during landing.

MAP-3 demonstrated the lowest levels of vastus medialis activity prior to

landing. This low activation of the medial quadriceps may have contributed

to the large degree of knee abduction displacement seen during the first 50-

75ms after initial contact, leading to a large peak knee abduction angle.

Interestingly, MAP-1 recorded the highest levels of preparatory vastus

lateralis activity, yet had the lowest degree of initial contact and peak knee

abduction. While the muscles acting directly at the knee are the most

influential,[187, 386] other muscles may also be relevant in supporting against

excessive frontal plane knee motion.

There are suggestions that the hip abductor and adductor musculature assist

in controlling frontal plane motion at the knee.[170, 387] This notion is supported

by the fact that excessive hip motion, specifically hip adduction and internal

rotation, has the capacity to shift the knee to an abducted position.[149, 162] MAP-

2 exhibited high levels of preparatory gluteus medius activation, however this

did not appear to be protective in limiting the degree of peak knee abduction

experienced. This is in line with previous work, where preparatory activation

of gluteus medius was also found to have no influence on peak knee abduction

angles during landing.[187] Muscles at the hip other than gluteus medius may

also be useful in providing this support. While the gluteus maximus primarily

functions as a hip extensor, it can function like the gluteus medius during

dynamic tasks[178] and may therefore assist in controlling frontal plane motion

at the knee. High activation of the gluteus maximus at initial contact and


155

during the early deceleration phase of landing was observed with MAP-1. This

activation strategy may have been a factor in this profile demonstrating the

lowest hip adduction across the profiles, while also exhibiting minimal hip

internal rotation. Subsequently, the elevated activation of the gluteus

maximus may have served as a compensatory factor for the high levels of

preparatory vastus lateralis activity seen with MAP-1, to avoid an abducted

knee posture by controlling frontal and transverse plane motion at the hip.

Appropriate preparatory activation of muscles at the knee, combined with

appropriate activation of the hip muscles during landing may serve as a useful

neuromuscular strategy for limiting frontal plane knee motion.

Transverse plane rotations and moments at the knee are also often implicated

in ACL injury risk.[20-22, 99] Internal or external rotation at the knee (i.e. internal

or external rotation of the tibia relative to the femur) is frequently observed

during ACL injury occurrences,[99] while large increases in ACL loads are seen

when internal rotation moments are applied at the knee.[20-22] Differences in the

magnitude of transverse plane rotations and moments were found between

the three activation profiles. With regard to transverse plane rotations at the

knee, each profile exhibited certain high-risk characteristics. While MAP-2

displayed minimal transverse plane rotation at initial contact, this profile

demonstrated the largest peak knee internal rotation angles. In contrast, MAP-

2 and MAP-3 landed in a position of external rotation at the knee. These two

profiles also exhibited a degree of internal rotation at the knee throughout the

landing. Based on this data, it appears none of the activation profiles provided

sufficient support against potentially high-risk transverse plane movement

strategies at the knee. With regard to transverse plane moments at the knee,

MAP-1 exhibited larger knee internal rotation moments than the remaining

profiles. While these differences were present, the magnitude of the transverse


156

plane knee moments experienced during the landing task were small and

negligible relative to knee moments in the remaining planes. It is unlikely that

this small difference (~ 0.04 Nm kgm-1) substantially increased the level of ACL

injury risk for this profile. The small magnitude of transverse plane knee

moments seen in this study is not surprising, as comparable values have been

observed in previous work examining a similar netball-specific landing

task.[253]

Limited knee flexion during landing has been linked to higher ACL loads[25]

and ACL injuries often occur with the knee in a relatively extended position.[97-

99] As ACL injuries are proposed to occur within the first 50 ms of landing[98]

and peak ACL loads occur during the deceleration phase of landing,[27] knee

flexion at initial contact and during the early stages of landing is an important

biomechanical strategy for reducing ACL injury risk. Smaller knee flexion

angles at initial contact and through the first 30ms after initial contact were

seen with MAP-1. Greater preparatory activation of the quadriceps relative to

the hamstrings has been linked to a reduction in initial contact knee flexion

angles during landing.[178] While all of the activation profiles exhibited greater

preparatory activation of the hamstrings relative to the quadriceps; MAP-1

tended to exhibit the highest preparatory activation of the quadriceps muscles,

while also exhibiting lower preparatory activation of the hamstrings. In

addition, MAP-1 exhibited the greatest activation of rectus femoris, a primary

knee extensor, during the early stages of landing. These activation

characteristics may have been a contributing factor to the less flexed posture

at the knee at initial contact and through the early stages of landing seen with

this profile.

Despite the differences between profiles, a number of similar neuromuscular

strategies were also observed. Across all profiles, the quadriceps and


157

gastrocnemius muscles tended to be the highest activated muscles within

muscle synergy one, while the quadriceps were also highly activated within

muscle synergy two. The external knee flexion moment was the largest of all

moments measured during the leap landing task. Activation of the quadriceps

would be required to counteract these external loads and decelerate the shank,

with this being a potential reason for the high levels of activation seen in this

muscle group during landing across all profiles. Further, recent work has

suggested the plantarflexors play an important role in unloading the ACL,[184]

and co-contraction of the gastrocnemius and quadriceps may serve to elevate

joint compression and protect the knee from external joint loading during

single-leg landing.[185] Considering these factors, the higher levels of activation

seen in the quadriceps and gastrocnemius at and after initial contact may

represent a requirement rather than a selective activation strategy to ensure

the knee is protected from the external loads experienced during the landing

task. An additional similarity across the profiles was the ratio of quadriceps to

hamstrings activation. The quadriceps and hamstrings tended to be the

highest and lowest activated muscles, respectively, within muscle synergies

one and two (i.e. around and after initial contact). This imbalance may stem

from the female cohort used for this study. Studies comparing the muscle

activation patterns of males and females during landing tasks have found

females tend to exhibit a ‘quadriceps dominant’ activation strategy.[59, 120, 129, 130]

The favoured activation of the quadriceps over the hamstrings may therefore

be an inherent trait of all muscle activation profiles employed by females

during landing tasks. The need to balance hamstrings and quadriceps

activation is important within female athletes, as a high quadriceps-hamstring

activation ratio has been linked to biomechanical factors associated with ACL

loading and injury risk.[177] The presence of an imbalance between quadriceps

and hamstring activation regardless of the activation profile used suggests


158

ACL injury prevention programs targeting female athletes may require a

general goal of increasing hamstrings muscle activation.

As in the previous chapter, it is important to acknowledge that this study

focused solely on muscle activation during the high-risk sporting task. The

physiological characteristics and strength of muscles are likely to be key

factors in protecting the knee and ACL during high-risk landing tasks, and

thus also require consideration in identifying neuromuscular contributions to

ACL injury risk. Individuals with differences in muscle physiology may need

to utilise different activation profiles to protect the ACL from potentially

hazardous joint postures and loads. Neuromusculoskeletal modelling

approaches allow the inclusion of muscle strength and length parameters,

while also calculating muscle forces during task performances. The use of

these approaches in conjunction with the assessment of muscle activation

profiles may assist in achieving a greater overall understanding of

neuromuscular contributions to ACL injury risk, and how this may vary across

individuals with different muscle physiology.

An additional limitation of this study was that data pertaining to trunk

biomechanics, and subsequently the centre of mass position were not

collected. Positioning of the centre of mass plays an important role in

modulating ACL injury risk.[167, 255, 372] Furthermore, the landing task examined

encompassed a flight phase where centre of mass position may have varied

from trial-to-trial. The positioning of centre of mass during the flight phase of

the movement may play an important role in dictating the muscle activation

profile required to perform the landing manoeuvre in a safe manner. Variable

centre of mass positions during the preparatory or flight phase of high-risk

sporting tasks may require different muscle activation profiles in order to

prevent lower limb biomechanical postures linked to ACL loading or injury


159

risk. Future work examining muscle activation profiles during high-risk

sporting tasks should also consider the relationship between preparatory

biomechanics and neuromuscular control.

5.5 Conclusions

This study found multiple muscle activation profiles were utilised during

repeat performances of a high-risk landing task. While multiple activation

profiles were identified, individual’s tended to persevere with a preferred

profile through the majority of landing trials. ‘Safe’ and ‘high-risk’ activation

profiles were not identified. Rather, specific muscle activation characteristics

or combinations of activation characteristics within profiles were linked with

the presence of biomechanical deficits linked to ACL loading or injury risk.

Injury prevention programs directed at improving neuromuscular control

during high-risk landing tasks may need to target these specific muscle

activation characteristics in order to reduce ACL injury risk. As different

activation profiles were seen across individuals, implementing injury

prevention programs in a universal manner may not be appropriate. Tailored

injury prevention programs targeting the specific neuromuscular deficits

relevant to the individual may induce more beneficial neuromuscular


160

Chapter Six: Exploring Individual Adaptations to an Anterior Cruciate Ligament Injury Prevention Program

6.1 Introduction

Numerous studies have highlighted the benefits of anterior cruciate ligament

(ACL) injury prevention programs in altering neuromuscular and

biomechanical factors associated with ACL injury risk.[34-45, 68, 69, 71-75, 79, 166, 201, 213-

217] The success of ACL injury prevention programs is generally defined by a

reduction in the total number or rate of ACL injuries, or by a change in

neuromuscular or biomechanical factors associated with ACL loading or

injury risk. Regardless of the outcome of interest, an overall improvement in

the training group (in isolation or relative to a control group) is deemed

successful, with little attention paid to the individual responses. Despite the

apparent success of injury prevention programs in recent times, ACL injury

rates have remained relatively unchanged.[1, 2] A focus on group results limits

our understanding of how different individuals respond to training, and this

may be a factor that is restricting our understanding of the mechanisms that

underpin failure or success of ACL injury prevention programs. Certain

individuals have been shown to exhibit detrimental responses to training,

even when an overall improvement has been observed at the group level.[71] It

may therefore be erroneous to expect that everyone will gain similar benefits

from the same training stimulus. Despite this, individual responses to ACL

injury prevention programs have received little attention.

Individual Adaptations to ACL Injury Prevention

161

Certain studies have considered the individual responses to ACL injury

prevention programs.[71, 82, 83] Where this has been examined, non-uniform

neuromuscular and biomechanical adaptations have been observed.[71, 82, 83]

Cowling et al.[82] found individuals displayed non-uniform adaptations in

quadriceps and hamstrings muscle activation strategies during a unilateral

netball-specific landing task, despite completing identical training programs.

No differences were found from pre- to post-training when examining the

group, highlighting the potential of individual responses to be masked when

focusing on data at the group level. Variable adaptations to lower limb

biomechanics have also been observed following ACL injury prevention

programs.[71, 83, 84] Pollard et al.[71] concluded an in-season injury prevention

program had a beneficial effect on frontal and transverse plane hip motion

when performing a jump-landing task. Upon examination of individual data,

some athletes were found to have greater improvements in the frontal versus

transverse plane, and vice versa.[71] It was suggested that athletes exhibited

greater changes in the plane they demonstrated a pre-existing high-risk

pattern.[71] However, certain athletes exhibited detrimental adaptations

following training, suggesting that not all responded favourably to the

training program.[71] Further, Dempsey et al.[83] found certain athletes exhibited

large favourable responses in torso rotation during a side-step cutting

manoeuvre following training, while others did not. The authors also

hypothesised that these athletes may have displayed a more high-risk

technique at baseline, and hence obtained a greater benefit from training.

There is evidence to support the hypothesis that the level of ACL injury risk

prior to commencing an injury prevention program is a modulating factor in

the magnitude of response to training.[84] Myer et al.[84] saw a greater decrease

in knee abduction moments during a bilateral drop vertical jump task in

‘high-‘ (knee abduction moment > 22.5 Nm during a drop vertical jump) versus


162

‘low-risk’ athletes. It is, however, yet to be established whether this effect

exists for high-risk sport-specific tasks associated with ACL injuries.

Of the studies which have examined individual changes following an ACL

injury prevention program, there appears to be a consensus that not all athletes

respond in a similar manner to the same training stimulus.

While a large number of studies have assessed biomechanical

and neuromuscular adaptations following an injury prevention

program,[34-36, 38-45, 68, 69, 71-75, 79, 83, 166, 201, 213-217] very few of these have specifically

focused on the variable responses of individuals within the training group.

Failure to examine these individual responses may impact on our

understanding of what does or does not constitute effective ACL injury

prevention practice. For example; where an injury prevention program has

been shown to produce favourable responses at the group level, we may

conclude that this program is effective. However, it may be that not all

individuals who complete the program receive the perceived benefits, and

thus may still be at-risk of ACL injury. A greater understanding of how

different individuals respond to an identical training stimulus can allow for

better development of ACL injury prevention programs that provide a benefit

to all involved. In addition, examining individual responses and the

characteristics of participants exhibiting different responses may maximise

our ability to provide targeted or individualised injury prevention programs.

The purpose of this study was to examine the effects of an ACL injury

prevention program on neuromuscular control and lower limb biomechanics

during a high-risk landing task at the group and individual level. Specifically,

this study aimed to: (i) investigate the effects of an ACL injury prevention

program on neuromuscular control and lower limb biomechanics during a

high-risk landing task at the group level by comparing the overall responses


163

between a control and training group; (ii) investigate the effects of an ACL

injury prevention program at the individual level by examining the individual

neuromuscular and biomechanical responses of athletes within a training

group; and (iii) determine whether individuals’ deemed as ‘high-risk’ at

baseline based on peak knee abduction moments during a high-risk landing

task received a greater prophylactic benefit from the ACL injury prevention

program. It was hypothesised that the training group overall would benefit

from the injury prevention program as indicated by improvements in

neuromuscular and biomechanical risk factors associated with ACL injury;

however, variable responses to the program would be observed across

individuals within the training group. In addition, individuals deemed as

‘high-risk’ for ACL injury would receive a greater prophylactic benefit from

the training.

6.2 Methodology




6.2.1 Participants

Sixteen healthy sub-elite female netball players (age = 22.7 ± 2.7 years; height

= 171.8 ± 7.4 cm; mass = 68.6 ± 8.1) volunteered and provided written consent

to participate in this study. All participants met the inclusion/exclusion criteria

detailed in Section 3.2 of Chapter Three. Upon study enrolment, participants

were randomly allocated to a control or training group. Participant

characteristics of the groups are presented in Table 6.1 (see Page 164). No

statistically significant differences (p > 0.05) were present for age, height,


164

weight, netball experience, and training frequency or duration between

groups.

Table 6.1 Characteristics of participants allocated to control and training groups.

CONTROL (n = 8)

TRAINING (n = 8)

AGE (y) 23.4 ± 2.7 22.0 ± 2.5

HEIGHT (cm) 171.6 ± 7.6 172.0 ± 7.6

WEIGHT (kg) 69.8 ± 11.0 67.9 ± 4.8

NETBALL EXPERIENCE (y) 14.3 ± 3.7 11.3 ± 4.8

TRAINING FREQUENCY 2.0 ± 1.1 2.0 ± 1.1

TRAINING DURATION 146.3 ± 96.8 162.5 ± 109.2

Training frequency reported as number of netball-specific training sessions per week; training duration reported as the average duration (in minutes) of all netball-specific training per week.


Data from the netball-specific landing task (i.e. the leap landing), as described

in Section 3.3.1 of Chapter Three, performed during testing sessions one (i.e.

baseline) and two (i.e. six-weeks) were used within this study. Participants

performed 20 successful trials of the leap landing on a predetermined ‘landing

limb’ at both testing sessions. The landing limb was determined at the baseline

session by asking participants which limb they would be more comfortable

landing on when catching a pass during a game. Further to this, players were

given an opportunity to trial landing on each limb to identify which they were

more comfortable landing on. Each landing trial was initiated by breaking and



a pass (see Figure 6.1, Page 165). Trials were repeated until the 20 successful


165

trials were recorded. Participants were allocated 60-90 seconds rest between

trials in an effort to minimise fatigue.[134, 255] Electromyography, kinematic and

kinetic data collected from the participants ‘landing limb’ during all landing

trials were used within this study.


6.2.2.1 ACL Injury Prevention Program

Participants allocated to the training group were asked to complete a six-week

ACL injury prevention program between testing sessions. The injury

prevention program used was the ‘Down 2 Earth’ (D2E) program.[258] The

specifics of this program, including the exercises and techniques used are

described in Section 3.3.5 of Chapter Three. The main tenet of the program is

that improved landing technique can be facilitated via a feedback schedule,

directing focus to key guidelines for safe and effective landing technique (see

Figure 6.2, Page 166).[258, 259] The program incorporated a core set of exercises

around this feedback protocol; involving stationary jump-landings,

directional jump-landings, lunging, and netball-specific landing

movements.[258] These core exercises were progressed throughout the six-week

program by adding to or varying the movement to increase difficulty.


166



program. Adapted and reproduced with permission from

Saunders et al.[258]

Participants were asked to complete a total of three sessions per week; where

one session was completed under supervision of a research team member, and

the remaining two conducted as ‘home-based’ sessions completed by the

participant individually. The home-based sessions were designed to replicate

the training sessions conducted under supervision, with some modifications

where applicable.[254] Attendance at the supervised sessions were recorded.

Participants were given a training booklet (see Appendix D, Page 379) that

guided them through the home-based sessions and contained a diary page to

monitor attendance to these training sessions. Attendance (%) was calculated

as the total number of sessions completed relative to the total number of

sessions across the program (i.e. three sessions by six-weeks = 18).


167


Electromyography (EMG), and kinematic and kinetic data from the leap

landing task were collected and analysed as per the procedures outlined in

Section 3.4.1 and 3.4.2, respectively, of Chapter Three. Briefly, surface EMG

data were collected unilaterally from eight lower limb muscles (gluteus

maximus [GMAX], gluteus medius [GMED], medial hamstrings [MHAM],

lateral hamstrings [LHAM], rectus femoris [RF], vastus lateralis [VL], vastus

medialis [VM], and lateral gastrocnemius [GAS]) on the participants’ ‘landing

limb’ using wireless parallel-bar electrodes (Delsys Trigno, Delsys

Incorporation, Boston, MA, United States). All EMG data were processed

using custom MATLAB (version R2014a) (The Mathworks Incorporated,

Natick, MA, United States) programs. Raw EMG signals were online

processed using a bandpass filter between 10 and 500 Hz, and later offline

processed using a 20 Hz high-pass fourth-order zero-lag Butterworth digital

filter to reduce the effects of movement artefact; full-wave rectified; and

converted to a linear envelope via a 6 Hz low-pass fourth-order zero-lag

Butterworth digital filter. EMG data recorded during landing trials were

amplitude normalised (% MVC) to the larger of the peak value recorded

during maximal voluntary isometric contractions or the peak EMG activity

during landing trials.[282, 288] EMG data were extracted from 150 ms prior

through to 300 ms post initial contact (IC) across all landing trials for

subsequent analyses. This window was chosen as previous research




As in the previous two studies, muscle synergies were used to represent

muscle activation during the landing task. The extracted EMG data from both


168

groups and time points were collated and factorised by applying a non-

negative matrix factorisation (NMF) algorithm[298] (as outlined in Section

3.4.1.3.1 of Chapter Three) to obtain the muscle synergy components, those

being the time-variant muscle activation coefficients and time-invariant

muscle weighing coefficients. The NMF algorithm was iterated by varying the

number of synergies between one and six, with the number of synergies that

resulted in greater than 90% mean total variance accounted for ( ) between

experimental and reconstructed EMG data retained. was also calculated

across individual muscles ( ) to ensure that each muscle’s activity

patterns were accounted for by the retained synergies (i.e. > 75%).[48]

Following factorisation, the muscle activation coefficients were normalised to

their maximum value (i.e. each activation coefficient had a maximum value of

1), with the muscle weighing coefficients within each of the muscle synergies

scaled by the same respective factor.[49, 294] The muscle weighing coefficients

from individual muscles across the identified synergies were treated as

separate dependent variables within this study. This was done as it was

deemed important that specific changes in muscle activation characteristics

stemming from the ACL injury prevention program could be detected.




Kingdom) sampling at 250 Hz, synchronised with a 600 x 900 mm in ground

force plate (Advanced Medical Technology Incorporated, Watertown, MA,

United States) sampling at 1000 Hz collected marker trajectory and ground

reaction force (GRF) data. Marker trajectory and GRF data were low-pass

filtered using a cubic smoothing spline with a cut-off frequency of 12 Hz.



169













landing.[161]

Previous studies within this thesis have used one-dimensional statistical

parametric mapping (SPM1D) to assess differences in joint rotations and

moments, and have therefore not required any further reduction of the

waveform data. Within this study, a repeated measures approach was used

across statistical tests. Current SPM1D repeated measures procedures only

support balanced designs (i.e. same number of observations for all possible

combinations of factors).[371] As different participants made up the control and

training groups, the criteria for a balanced design was not met and SPM1D

could not be used. When applied to continuous waveforms, such as joint

rotation and moment data, principal component analysis (PCA) has a number

of similar benefits to SPM1D. PCA can be used to view an entire waveform as

function, preserving the main features and patterns of the curve.[333] This

ensures the majority of the data along the waveform is considered in

subsequent analyses. Further, PCA removes the need to pre-select discrete


170

features from the waveform, removing the potential for regional focus bias.[338]

With these factors in mind, PCA was used in place of SPM1D within this study

to reduce the joint rotation and moment waveform data to discrete values that

could be input as dependent variables for repeated measures analyses.

Kinematic and kinetic data were exported to MATLAB and submitted to PCA

to identify the lower limb biomechanical patterns from the leap landing task.

Patterns were extracted for hip flexion/extension (HFLEX/EXT),

adduction/abduction (HADD/ABD), and internal/external rotation (HIR/ER); knee

flexion/extension (KEXT/FLEX), adduction/abduction (KADD/ABD), and

internal/external rotation (KIR/ER); hip extension/flexion moments (HEXT/FLEX

Mom.), abduction/adduction moments (HABD/ADD Mom.), and external/internal

rotation moments (HER/IR Mom.); and knee flexion/extension moments

(KFLEX/EXT Mom.); abduction/adduction moments (KABD/ADD Mom.); and

external/internal rotation moments (HER/IR Mom.). The n number of principal

patterns (PP) that summed to explain at least 90% of the variation in the

waveform were retained and referred to as PP1, PP2, …, PPn.[276, 327] The 90%

criterion was chosen as this value has been previously used when examining

lower limb biomechanical waveforms.[326, 335] Principal pattern scores (PP-

Scores) were then computed for each biomechanical pattern across each joint

rotation and moment variable. Principal pattern interpretation was achieved

by examining the waveforms obtained by adding and subtracting a suitable

multiple of the patterns loading vector to the mean of the relevant

waveform.[333]

Due to its relationship to future ACL injury,[160] the peak knee abduction

moment (pKAM) was used as the dependent variable for categorising

participants as ‘high-’ or ‘low-risk’.[84] Participants who recorded an average

normalised pKAM greater than 0.66 Nm kgm-1 were classified as ‘high-risk,’


171

with the remainder classified as ‘low-risk.’ This value represented the 66th

percentile cut-off for pKAM in the examined cohort, thus the ‘high-risk’ group

consisted of the top third of participants for pKAM. While participants below

this threshold may have exhibited pKAMs that could still place them at-risk

of injury, the percentile cut-off identified the highest risk individuals relative

to the remainder of the cohort.


Normality of data distribution within groups and time points, and equality of

variance between groups and time points was confirmed for all of the

dependent variables via non-significant (p > 0.05) Shapiro-Wilk and Levene

statistics, respectively. As the assumptions of normality and equality of

variance were not violated, parametric tests were utilised.

A series of statistical tests were conducted as follows. First; two separate two-

way factorial analysis of variance (ANOVAs) tests (group x time and

participant x time) were conducted using approach speed as the dependent

variable, to determine whether approach speed was required as a covariate for

subsequent analyses.

Second; separate 2 x 2 (group x time) repeated measures ANOVA were used

to test for the main and interaction effects of group (control vs. training) and

time (baseline vs. six-weeks) on all dependent variables (i.e. muscle synergy

weighing coefficients and biomechanical PP-Scores). This analysis determined

whether there was an effect of the injury prevention program at the group

level. Where statistically significant main or interaction effects were identified,

post-hoc analyses of simple main effects within and between groups and time

points were undertaken using Tukey’s honest significant difference (HSD)

tests to identify where specific differences occurred.


172

Third; data from the training group was isolated and used in separate 8 x 2

(participant x time) repeated measures ANOVA to test for the main and

interaction effects of participant and time (baseline vs. six-weeks) on all

dependent variables. This analysis determined whether there were differences

in the response to the injury prevention program within the training group.[388]

No post-hoc tests were conducted on individual data. Rather, when a

statistically significant participant x time interaction was observed, effect sizes

(Cohen’s d) were used to examine the magnitude of the effect of the injury

prevention program across individuals. Effect sizes (Cohen’s d) were

calculated based on the change in the dependent variable from baseline to six-

weeks across individuals. In addition, individual change scores from baseline

to six-weeks were calculated across all dependent variables for both control

and training group participants. The standard deviation of these change scores

for each group were descriptively compared to examine the magnitude of

individual variation of each group. This analysis helped establish whether any

variability in individual responses observed was a result of the training

program rather than simply test-retest variation.

Fourth; separate 2 x 2 x 2 (group x time x risk level) repeated measures

ANOVA were used to test for the main and interaction effects of group

(control vs. training), time (baseline vs. six-weeks) and risk level (‘high-‘ vs.

‘low-risk’) on all dependent variables. This analysis determined whether

individuals deemed as ‘high-risk’ received a greater benefit from training in

comparison to ‘low-risk’ individuals.[84] As above, statistically significant main

or interaction effects were followed by post-hoc analyses of simple main

effects within and between factors using Tukey’s HSD tests to identify where

specific differences occurred. Non-significant post-hoc tests were expected

due to the limited sample within each group and risk level combination.


173

Cohen’s d was therefore used to estimate the effect sizes for all non-significant

post-hoc comparisons.[269]

An exploratory alpha level of 0.05 was used to determine statistical

significance for the main and interaction effects within all ANOVA tests.[84, 388]

Across all tests, partial eta squared (η2) was used as a measure of effect size for

main and interaction effects, with the interpretation of effect size values

guided by Miles and Shevlin[383] (small η2 = 0.01; medium η2 = 0.06; large η2 =

0.14). A Bonferroni correction was applied to an initial alpha level of 0.05

across all post-hoc analyses based on the number of comparisons being made

within a simple main effects test. Effect sizes (Cohen’s d) for post-hoc

difference tests were calculated based on the change in the dependent variable

from baseline to six-weeks. Interpretation of effect sizes was guided by

Cohen[384] (small d = 0.2; medium d = 0.5; large d = 0.8).

6.3 Results

All participants from the control and training groups completed the baseline

and six-weeks testing sessions. No participants from the training group were

excluded based on their reported attendance to the training program (see

Table 6.2, Page 174). However, as these values varied across participants they

were taken into account when interpreting potential differences in individual

responses.


174

Table 6.2 Reported attendance to the injury prevention program for

participants in the training group (n = 8; T1 – T8).

Participant

SUPERVISED SESSIONS

( / 6)

HOME-BASED SESSIONS

( / 12)

TOTAL SESSIONS

( / 18)

n % n % n %

T1 6 100.0 12 100.0 18 100.0

T2 5 83.3 12 100.0 17 94.4

T3 4 66.7 9 75.0 13 72.2

T4 6 100.0 6 50.0 12 66.7

T5 5 83.3 9 75.0 14 77.8

T6 5 83.3 6 50.0 11 61.1

T7 6 100.0 6 50.0 12 66.7

T8 5 83.3 7 58.3 12 66.7

T – Training Participant


175

Using the criteria previously described in Chapter Three (i.e. overall >

90% and > 75%), three muscle synergies were retained from the

extracted EMG data (see Figure 6.3).

A.

B.


number of extracted muscle synergies. B) relative to the number of extracted muscle synergies. GMAX – gluteus

maximus; GMED – gluteus medius; MHAM – medial

hamstrings; LHAM – lateral hamstrings; RF – rectus femoris; VL – vastus medialis; GAS – gastrocnemius.


176

Despite additional data from testing session two being included in this

analysis, the muscle synergies identified were almost identical to those of the

previous studies (i.e. Chapters Four and Five). Inspection of the activation

coefficients associated with each muscle synergy (see Figure 6.4) indicated that

the first synergy captured muscle activation around initial contact (i.e. at

landing), the second captured muscle activation after initial contact (i.e. post

landing), and the third captured muscle activation prior to initial contact (i.e.

preparatory activity).

Figure 6.4 Normalised activation coefficients of the three muscle

synergies extracted from the factorisation of EMG data. Dashed vertical line represents initial landing contact (IC).


177

Principal patterns were extracted for leap landing joint rotations and moments

via PCA (see Table 6.3, Page 178 and Table 6.4, Page 179, respectively). A

graphical display of the patterns associated with high and low PP-Scores for

each PP is provided in Appendix G (see Page 396).

178


Tabl

e 6.

3 D

escr

iptio

n of

pri

ncip

al p

atte

rns

(PPs

) ext

ract

ed fr

om h

ip a

nd k

nee

join

t rot

atio

n da

ta fr

om th

e le

ap la

ndin

g ta

sk.

Join

t Rot

atio

n PP

D

escr

iptio

n V

aria

nce

Expl

aine

d

HFL

EX/E

XT

1 2 G

ener

al s

hape

and

ove

rall

mag

nitu

de –

hig

her s

core

= g

reat

er o

vera

ll hi

p fle

xion

M

agni

tude

of h

ip fl

exio

n fr

om IC

to ~

50 m

s po

st IC

– h

ighe

r sco

re =

gre

ater

hip

flex

ion

82.4

%

10.6

%

HA

DD

/ABD

1 2

Gen

eral

sha

pe a

nd o

vera

ll m

agni

tude

– h

ighe

r sco

re =

gre

ater

ove

rall

hip

addu

ctio

n M

agni

tude

of f

ront

al p

lane

hip

ang

ular

dis

plac

emen

t – h

ighe

r sco

re =

redu

ced

angu

lar d

ispl

acem

ent

84.6

%

10.0

%

HIR

/ER

1 2 3

Gen

eral

sha

pe a

nd o

vera

ll m

agni

tude

– h

ighe

r sco

re =

gre

ater

ove

rall

hip

inte

rnal

rota

tion

Mag

nitu

de o

f hip

ext

erna

l rot

atio

n fr

om IC

to ~

30 m

s po

st IC

– h

ighe

r sco

re =

redu

ced

hip

exte

rnal

rota

tion

Mag

nitu

de o

f pea

k hi

p in

tern

al ro

tatio

n at

~60

-90

ms p

ost I

C –

hig

her s

core

= g

reat

er h

ip in

tern

al ro

tatio

n pe

ak

73.0

%

16.1

%

7.1%

K

EXT/

FLEX

1 2

Gen

eral

sha

pe a

nd o

vera

ll m

agni

tude

– h

ighe

r sco

re =

redu

ced

over

all k

nee

flexi

on

Mag

nitu

de o

f sag

ittal

pla

ne k

nee

angu

lar d

ispl

acem

ent –

hig

her s

core

= re

duce

d an

gula

r dis

plac

emen

t 83

.3%

8.

9%

KA

DD

/ABD

1 2

Gen

eral

sha

pe a

nd o

vera

ll m

agni

tude

– h

ighe

r sco

re =

redu

ced/

grea

ter o

vera

ll kn

ee a

bduc

tion/

addu

ctio

n Ph

ase

shift

in ti

min

g –

high

er sc

ore

= ea

rlie

r kne

e ad

duct

ion

peak

79

.2%

11

.0%

K

IR/E

R

1 2 3

Gen

eral

sha

pe a

nd o

vera

ll m

agni

tude

– h

ighe

r sco

re =

gre

ater

ove

rall

knee

inte

rnal

rota

tion

Mag

nitu

de o

f kne

e ex

tern

al ro

tatio

n fr

om IC

to ~

30 m

s po

st IC

– h

ighe

r sco

re =

redu

ced

knee

ext

erna

l rot

atio

n M

agni

tude

of p

eak

– hi

gher

sco

re =

gre

ater

pea

k kn

ee in

tern

al ro

tatio

n

58.0

%

22.6

%

13.4

%

HFL

EX/E

XT

– hi

p fle

xion

/ext

ensi

on;

HA

DD

/ABD

– h

ip a

dduc

tion/

abdu

ctio

n; H

IR/E

R –

hip

int

erna

l/ext

erna

l ro

tatio

n; K

EXT/

FLEX

– k

nee

flexi

on/e

xten

sion

; K

AD

D/A

BD –

kne

e ad

duct

ion/

abdu

ctio

n; K

IR/E

R –

kne

e in

tern

al/e

xter

nal r

otat

ion;

PP

– pr

inci

pal p

atte

rn; m

s –

mill

isec

onds

; IC

– in

itial

con

tact

.

179


Tabl

e 6.

4 D

escr

iptio

n of

pri

ncip

al p

atte

rns

(PPs

) ext

ract

ed fr

om h

ip a

nd k

nee

join

t mom

ent d

ata

from

the

leap

land

ing

task

.

Join

t Mom

ent

PP

Des

crip

tion

Var

ianc

e Ex

plai

ned

HEX

T/FL

EX M

om.

1 2 3 4

Gen

eral

sha

pe a

nd o

vera

ll m

agni

tude

– h

ighe

r sco

re =

redu

ced

hip

flexi

on m

omen

t M

agni

tude

of p

eak

– hi

gher

sco

re =

redu

ced

peak

hip

flex

ion

mom

ent

Mag

nitu

de fr

om ~

30-6

0 m

s pos

t IC

– h

ighe

r sco

re =

redu

ced

hip

flexi

on m

omen

t fro

m ~

30-6

0 m

s po

st IC

M

agni

tude

from

~60

-90

ms p

ost I

C –

hig

her s

core

= re

duce

d hi

p fle

xion

mom

ent f

rom

~60

-90

ms

post

IC

61.9

%

19.1

%

7.5%

5.

9%

HA

BD/A

DD

Mom

. 1 2 3

Gen

eral

sha

pe a

nd o

vera

ll m

agni

tude

– h

ighe

r sco

re =

redu

ced

over

all h

ip a

dduc

tion

mom

ent

Mag

nitu

de fr

om IC

to ~

30m

s po

st IC

– h

ighe

r sco

re =

gre

ater

hip

abd

uctio

n m

omen

t fro

m IC

to ~

30 m

s po

st IC

Pr

esen

ce o

f bim

odal

pea

ks in

cur

ve –

hig

her s

core

= p

rese

nce

of h

ip a

bduc

tion

and

addu

ctio

n m

omen

t pea

ks

66.9

%

20.1

%

6.1%

H

ER/I

R M

om.

1 2 G

ener

al s

hape

and

ove

rall

mag

nitu

de –

hig

her s

core

= g

reat

er o

vera

ll hi

p in

tern

al ro

tatio

n m

omen

t M

agni

tude

of p

eak

hip

exte

rnal

rota

tion

mom

ent –

hig

her s

core

= re

duce

d pe

ak h

ip e

xter

nal r

otat

ion

mom

ent

77.1

%

14.2

%

KFL

EX/E

XT M

om.

1 2 3

Gen

eral

sha

pe a

nd o

vera

ll m

agni

tude

– h

ighe

r sco

re =

gre

ater

ove

rall

knee

flex

ion

mom

ent

Phas

e sh

ift in

tim

ing

– hi

gher

scor

e =

dela

yed

peak

kne

e fle

xion

mom

ent

Mag

nitu

de o

f sag

ittal

pla

ne m

omen

t pea

ks –

hig

her s

core

= re

duce

d kn

ee fl

exio

n an

d ex

tens

ion

mom

ent p

eaks

49.3

%

36.4

%

6.2%

K

ABD

/AD

D M

om.

1 2 3

Mag

nitu

de o

f kne

e ad

duct

ion

mom

ent –

hig

her s

core

= re

duce

d kn

ee a

dduc

tion

mom

ent

Mag

nitu

de o

f pea

k kn

ee a

bduc

tion

mom

ent –

hig

her s

core

= g

reat

er p

eak

mom

ent

Tim

ing

of p

eak

knee

abd

uctio

n m

omen

t – h

ighe

r sco

re =

ear

lier p

eak

61.2

%

24.2

%

9.0%

K

ER/IR

Mom

. 1 2

Gen

eral

sha

pe a

nd o

vera

ll m

agni

tude

– h

ighe

r sco

re =

gre

ater

ove

rall

knee

inte

rnal

rota

tion

mom

ent

Mag

nitu

de o

f kne

e in

tern

al ro

tatio

n m

omen

t pea

k –

high

er s

core

= g

reat

er p

eak

mom

ent

81.4

%

12.8

%

HEX

T/FL

EX M

om. –

hip

ext

ensi

on/fl

exio

n m

omen

t; H

ABD

/AD

D M

om. –

hip

abd

uctio

n/ad

duct

ion

mom

ent;

HER

/IR M

om. –

hip

ext

erna

l/int

erna

l rot

atio

n m

omen

t; K

FLEX

/EX

T M

om. –

kn

ee e

xten

sion

/flex

ion

mom

ent;

KA

BD/A

DD M

om.

– kn

ee a

bduc

tion/

addu

ctio

n m

omen

t; K

ER/IR

Mom

. –

knee

ext

erna

l/int

erna

l ro

tatio

n m

omen

t; PP

– p

rinc

ipal

pat

tern

; m

s –

mill

isec

onds

; IC

– in

itial

con

tact

.


180

The following sections detail the results of the statistical tests undertaken.

6.3.1 Approach Speed

The group x time factorial ANOVA revealed no main effect for group (p =

0.421, η2 = 0.004) or time (p = 0.232, η2 = 0.008). Similarly, no group x time

interaction was observed (p = 0.301, η2 = 0.011). This finding required no

further consideration of approach speed for the analyses conducted at the

group level, and indicated that any changes were not due to differences in

approach speed between the two groups or testing sessions. In contrast, the

participant x time factorial ANOVA revealed a statistically significant main

effect for participant (p < 0.001, η2 = 0.618), indicating that approach speed

differed between individuals. Based on these results, approach speed was

included as a covariate for the repeated measures analyses that used

participant as a factor. No main effect was observed for time (p = 0.231, η2 =

0.009) or for the participant x time interaction (p = 0.363, η2 = 0.012), indicating

that any individual changes over time were not due to differences in approach

speed between the two testing sessions.

6.3.2 Group Analysis

The mean muscle synergy weighing coefficient data for the control and

training groups across the two testing sessions are presented in Figure 6.5 (see

Page 181) and Table 6.5 (see Page 182). Mean EMG waveform data from 150

ms prior through to 300 ms post IC during the leap landing task for the control

and training groups across the two testing sessions are presented in Figure 6.6

(see Page 183).

181


Fi

gure

6.5

M

ean

(± S

D)

mus

cle

syne

rgy

wei

ghin

g co

effi

cien

t val

ues

from

the

leap

land

ing

task

for

the

cont

rol a

nd tr

aini

ng

grou

ps fr

om b

asel

ine

to s

ix-w

eeks

.

182


Tabl

e 6.

5 M

ean

(± S

D) m

uscl

e sy

nerg

y w

eigh

ing

coef

fici

ent v

alue

s fr

om t

he le

ap la

ndin

g ta

sk f

or th

e co

ntro

l and

trai

ning

grou

ps fr

om b

asel

ine

to s

ix-w

eeks

.

Mus

cle

Mus

cle

Syne

rgy

1 (A

ctiv

atio

n ar

ound

IC)

M

uscl

e Sy

nerg

y 2

(Act

ivat

ion

afte

r IC

)

Mus

cle

Syne

rgy

3 (A

ctiv

atio

n pr

ior t

o IC

)

CO

NTR

OL

TRA

ININ

G

C

ON

TRO

L TR

AIN

ING

CO

NTR

OL

TRA

ININ

G

BL

6W

BL

6W

BL

6W

BL

6W

BL

6W

BL

6W

GM

AX

3.

20 ±

1.1

5 3.

29 ±

1.2

4 2.

73 ±

0.9

6 3.

00 ±

1.5

6

1.09

± 0

.55

0.86

± 0

.45

1.16

± 0

.63

0.54

± 0

.23

0.

91 ±

0.5

9 1.

11 ±

0.2

9 0.

77 ±

0.5

3 1.

20 ±

0.3

2

GM

ED

2.76

± 0

.89

3.06

± 0

.87

3.67

± 1

.10

3.54

± 0

.78

1.

28 ±

0.6

6 1.

14 ±

0.6

3 1.

46 ±

0.7

7 1.

09 ±

0.6

5

0.97

± 0

.46

1.24

± 0

.43

1.83

± 0

.67

1.85

± 0

.61

MH

AM

2.

29 ±

1.2

9 1.

78 ±

0.7

3 1.

20 ±

0.6

0 1.

50 ±

0.9

0

0.97

± 0

.35

0.91

± 0

.51

0.76

± 0

.34

0.69

± 0

.32

3.

04 ±

0.8

3 2.

74 ±

1.0

8 3.

02 ±

0.8

5 3.

31 ±

1.1

2

LHA

M

2.08

± 1

.82

2.06

± 0

.90

1.88

± 1

.35

1.50

± 1

.03

0.

79 ±

0.4

2 0.

72 ±

0.4

4 1.

18 ±

0.9

5 0.

75 ±

0.4

0

2.16

± 0

.71

2.00

± 0

.96

2.64

± 1

.10

2.76

± 0

.99

RF

4.75

± 1

.16

4.70

± 1

.26

3.49

± 1

.06

3.12

± 1

.22

2.

37 ±

1.2

7 1.

43 ±

0.6

3 4.

29 ±

1.6

2 4.

29 ±

1.4

4

0.47

± 0

.44

0.38

± 0

.49

0.57

± 0

.54

0.58

± 0

.45

VL

4.14

± 0

.65

4.57

± 1

.30

4.10

± 0

.85

3.92

± 1

.29

2.

02 ±

0.8

0 1.

56 ±

0.6

2 2.

49 ±

0.9

0 2.

39 ±

0.9

9

0.90

± 0

.63

0.81

± 0

.72

1.09

± 0

.52

0.98

± 0

.74

VM

3.

83 ±

1.1

5 4.

74 ±

1.2

6 4.

80 ±

1.0

3 3.

78 ±

1.4

4

2.13

± 0

.88

1.91

± 0

.98

2.61

± 1

.06

2.66

± 1

.08

0.

83 ±

0.2

8 0.

86 ±

0.6

9 0.

83 ±

0.6

3 0.

87 ±

0.4

8

GA

S 4.

77 ±

1.6

5 5.

58 ±

1.3

2 5.

08 ±

1.3

8 4.

04 ±

1.3

8

1.18

± 0

.69

0.75

± 0

.93

0.95

± 0

.48

0.65

± 0

.19

1.

04 ±

0.4

1 0.

78 ±

0.4

9 0.

89 ±

0.4

3 1.

46 ±

1.6

1

IC –

initi

al c

onta

ct; B

L –

base

line;

6W

– s

ix w

eeks

; GM

AX

– gl

uteu

s m

axim

us; G

MED

– g

lute

us m

ediu

s; M

HA

M –

med

ial h

amst

ring

s; LH

AM

– la

tera

l ham

stri

ngs;

RF

– re

ctus

fem

oris

; VL

– va

stus

late

ralis

; VM

– v

astu

s m

edia

lis; G

AS

– ga

stro

cnem

ius.


183

Figure 6.6 Mean EMG waveform data from 150 ms prior through to 300

ms post initial contact (IC) during the leap landing task for the

control and training groups at baseline and six-weeks. Dashed vertical line represents IC.

A statistically significant main effect of group was observed for RF activation

within muscle synergy one (i.e. activation around IC) and two (i.e. activation

after IC) (p = 0.015, η2 = 0.352 and p < 0.001, η2 = 0.627, respectively). Inspection

of mean data revealed the training group, irrespective of the time point,

recorded reduced activation of RF at IC but greater activation of RF after initial

contact. A statistically significant main effect for group was also observed for

GMED activation within muscle synergy three (i.e. activation prior to IC) (p =

0.008, η2 = 0.405). Inspection of mean data revealed the training group,


184

irrespective of the time point, exhibited greater GMED preparatory activation.

A statistically significant main effect of time was observed for GMAX

activation within muscle synergy two (p = 0.008, η2 = 0.408). Irrespective of the

group, GMAX activation after IC was higher at baseline compared to six-

weeks. A statistically significant group x time effect was observed for GAS

activation with muscle synergy one (p = 0.035, η2 = 0.282) and MHAM

activation within muscle synergy three (p = 0.019, η2 = 0.333). With regard to

GAS activation at IC, no difference was observed between the training and

control group at baseline (p = 0.143, d = 0.164). However, at six-weeks the

training group exhibited reduced GAS activation at IC compared to the control

group (p < 0.001, d = 0.812). This difference stemmed from a statistically

significant increase in GAS activation at IC within the control group (p < 0.001,

d = 0.426) in conjunction with a statistically significant decrease in GAS

activation at IC within the training group (p < 0.001, d = 0.549). With regard to

preparatory MHAM activation, no difference was observed between the

training and control group at baseline (p = 0.870, d = 0.024). However, at six-

weeks the training group exhibited higher preparatory MHAM activation

compared to the control group (p < 0.001, d = 0.521). This difference stemmed

from a statistically significant increase in preparatory MHAM activation

within the training group only (p < 0.001, d = 0.264). No further statistically

significant main or interaction effects were observed across muscle activation

variables.

The mean hip and knee joint rotations and moments, respectively, from IC

through to 150 ms post IC during the leap landing task for the control and

training groups across the two testing sessions are presented in Figure 6.7 (see

Page 185) and Figure 6.8 (see Page 186).


185



control and training groups at baseline and six-weeks. FLEX –

flexion; EXT – extension; ADD – adduction; ABD – adduction; IR – internal rotation; ER – external rotation.


186

Figure 6.8 Mean hip and knee joint moments from initial contact (IC)


control and training groups at baseline and six-weeks. Mom. –

moment; EXT – extension; FLEX – flexion; ABD – adduction; ADD – adduction; ER – external rotation; IR – internal rotation.

A statistically significant main effect of time was found for hip flexion PP1 (p

= 0.003, η2 = 0.477). Irrespective of the group, greater hip flexion PP1 scores

were observed at baseline compared to six-weeks indicating a reduction in

overall hip flexion. A statistically significant group x time effect was observed

for hip internal/external rotation PP2 (p = 0.013, η2 = 0.369) and knee

flexion/extension moment PP2 (p = 0.016, η2 = 0.347) scores. The control group

exhibited lower hip internal/external rotation PP2 scores at baseline compared

to the training group (p < 0.001, d = 0.394), indicating greater hip external

rotation in the first 30 ms after IC. Decreases in hip internal/external rotation


187

PP2 scores were observed from baseline to six-weeks within the training group

(p < 0.001, d = 0.476), indicating a more externally rotated hip position in the

first 30 ms after IC at six-weeks compared to baseline. This resulted in no

difference between groups for hip external rotation in the first 30 ms after IC

at six-weeks (p = 0.316, d = 0.269). No difference was observed between the

training and control group at baseline for knee flexion/extension moment PP2

scores (p = 0.203, d = 0.668). However, at six-weeks the training group exhibited

higher knee flexion/extension PP2 scores compared to the control group (p =

0.020, d = 1.348), indicating a delayed time to peak knee flexion moment. The

change in knee flexion/extension moment PP2 scores from baseline to six-

weeks failed to reach statistical significance in both the control (p = 0.070, d =

0.380) and training groups (p = 0.129, d = 0.300). No further statistically

significant main or interaction effects were observed across the biomechanical

variables.

6.3.3 Individual Analysis

Statistically significant main effects for time were found for LHAM (p < 0.001;

η2 = 0.088), VM (p < 0.001; η2 = 0.247), and GAS (p < 0.001; η2 = 0.281) within

muscle synergy one; for GMAX (p < 0.001; η2 = 0.337), GMED (p < 0.001; η2 =

0.114), LHAM (p < 0.001; η2 = 0.259) and GAS (p < 0.001; η2 = 0.189) within

muscle synergy two; and for GMAX (p < 0.001; η2 = 0.439) and GAS (p < 0.001;

η2 = 0.270) within muscle synergy three. Irrespective of the participant; muscle

weighing coefficient values were higher at baseline across these dependent

variables, with the exception of GMAX and GAS activation in muscle synergy

three. In addition, statistically significant main effects for time were found for

hip flexion/extension PP1 (p < 0.001; η2 = 0.297), hip internal/external rotation

PP1 (p < 0.001; η2 = 0.256) and PP2 (p < 0.001; η2 = 0.333), knee

adduction/abduction PP1 (p < 0.001; η2 = 0.791) and hip abduction/adduction


188

moment PP1 (p < 0.001; η2 = 0.765). Irrespective of the participant; reduced hip

flexion (hip flexion/extension PP1), reduced hip internal and external rotation

peaks (hip internal/external rotation PP1 and PP2), reduced knee abduction

but greater knee adduction (knee adduction/abduction PP1), and larger

overall hip adduction moments (hip abduction/adduction moment PP1) were

observed at six-weeks compared to baseline.

Statistically significant main effects for participant (p < 0.001; η2 = 0.394 – 0.852)

and participant x time interactions (p < 0.001; η2 = 0.190 – 0.714) were seen

across all muscle activation (i.e. muscle weighing coefficient values) and

biomechanical (i.e. PP-Scores) dependent variables. Effect sizes were

calculated for changes from baseline to six-weeks for participants in the

training group across the muscle activation (see Table 6.6, Page 189; Table 6.7,

Page 190; and Table 6.8, Page 191) and biomechanical (see Table 6.9, Page 192

and Table 6.10, Page 193) dependent variables. Larger standard deviations for

change scores were seen within the training group across the majority of

dependent variables (see Table 6.11, Page 194; Table 6.12, Page 195; and Table

6.13, Page 196). The differing effect sizes between participants across the

majority of dependent variables indicated dissimilar adaptations to the

training program with regard to the direction and magnitude of the response.

Further, the larger standard deviations for changes scores observed in the

training group compared to the control group suggests the variation in

responses were outside the limits of test-retest variation for the majority of

dependent variables.


189

Table 6.6 Effect sizes (Cohen’s d) for individual changes of participants

in the training group (n = 8; T1 – T8) from baseline to six-weeks for muscle weighing coefficient values in muscle synergy one.

Participant

Muscle Synergy 1 (Activation around IC)


T1 2.07 -1.42 0.31 0.30 -0.59 -0.14 -0.47 0.24

T2 0.89 -0.52 -1.05 -2.60 -0.63 0.49 2.17 -1.94

T3 2.16 0.36 1.40 -1.49 -0.13 0.60 -1.49 0.01

T4 -1.21 -0.14 0.74 -1.91 -1.10 -1.61 -1.90 -1.85

T5 -0.63 0.24 0.92 0.73 1.83 -0.05 0.01 -0.56

T6 -0.70 1.07 -0.99 0.63 -0.58 0.80 -1.99 -0.56

T7 -0.97 -0.81 0.32 -0.06 -0.64 -0.07 -1.68 -1.98

T8 -0.98 0.98 0.65 0.54 -0.33 -1.31 -0.79 -1.22

IC – initial contact; GMAX – gluteus maximus; GMED – gluteus medius; MHAM – medial hamstrings; LHAM – lateral hamstrings; RF – rectus femoris; VL – vastus lateralis; VM – vastus medialis; GAS – gastrocnemius. ≤ -0.8 Large Decrease ≥ 0.8 Large Increase > -0.8 ≤ -0.5 Moderate to Large Decrease ≥ 0.5 < 0.8 Moderate to Large Increase > -0.5 ≤ -0.2 Small to Moderate Decrease ≥ 0.2 < 0.5 Small to Moderate Increase


190


in the training group (n = 8; T1 – T8) from baseline to six-weeks

for muscle weighing coefficient values in muscle synergy two.

Participant

Muscle Synergy 2 (Activation after IC)


T1 -1.65 -0.74 -0.75 1.07 -0.83 0.13 -0.09 1.27

T2 -1.59 -2.76 0.09 -3.38 -1.34 -2.03 -2.00 -0.31

T3 -0.43 -0.74 0.60 -1.02 -0.89 -0.26 -0.06 -1.73

T4 -1.15 -0.04 0.06 -0.63 2.56 1.32 3.72 -1.56

T5 -0.51 1.03 -0.22 -0.67 -1.93 -1.71 -1.07 -1.70

T6 -1.65 -0.80 0.35 -0.34 1.08 -1.26 -2.02 -0.54

T7 -0.67 -1.47 -1.04 0.39 -1.61 1.80 1.60 -0.98

T8 -0.94 1.08 -0.60 0.12 2.96 -0.14 1.18 1.04



191


in the training group (n = 8; T1 – T8) from baseline to six-weeks

for muscle weighing coefficient values in muscle synergy three.

Participant

Muscle Synergy 3 (Activation prior to IC)


T1 -0.12 -1.37 0.07 1.05 -0.87 -0.34 -3.14 -1.93

T2 -0.19 0.06 -0.39 -1.11 1.96 -1.39 -1.37 2.51

T3 3.22 1.16 0.07 0.02 -0.39 0.02 -0.67 -2.14

T4 2.99 1.81 -0.12 0.19 1.42 -1.47 2.93 4.56

T5 2.93 0.40 0.76 0.48 -0.35 -3.25 5.41 -1.31

T6 4.27 -1.37 1.11 -0.69 -0.18 1.27 1.93 -0.75

T7 0.60 0.29 1.90 2.02 -0.20 4.34 1.11 0.52

T8 -1.36 -0.32 -0.06 0.35 0.23 -2.38 -1.18 1.01


192


Tabl

e 6.

9 Ef

fect

siz

es (C

ohen

’s d

) for

indi

vidu

al c

hang

es o

f par

ticip

ants

in th

e tr

aini

ng g

roup

(n =

8; T

1 –

T8) f

rom

bas

elin

e to

six-

wee

ks fo

r hip

and

kne

e jo

int r

otat

ion

prin

cipa

l pat

tern

(PP)

sco

res.

Part

icip

ant

HFL

EX/E

XT

H

AD

D/A

BD

H

IR/E

R

K

EXT/

FLEX

KA

DD

/ABD

KIR

/ER

PP1

PP2

PP

1 PP

2

PP1

PP2

PP3

PP

1 PP

2

PP1

PP2

PP

1 PP

2 PP

3

T1

-0.5

9 0.

03

-0

.34

-1.1

1

-1.7

3 -0

.26

-0.0

4

0.99

1.

04

2.

40

0.79

0.53

-1

.76

0.17

T2

-2.2

8 0.

90

-1

.37

0.43

-1.0

3 0.

04

-2.8

7

1.16

3.

34

3.

99

-3.9

1

-0.2

0 0.

05

0.58

T3

-0.2

5 -0

.26

-0

.58

-0.3

6

-0.8

1 -1

.65

-0.4

3

1.94

-1

.64

3.

87

-1.1

7

-0.3

0 1.

40

0.21

T4

-1.3

3 -1

.72

1.

21

-1.7

7

-0.6

7 -2

.78

-1.8

8

0.62

-1

.66

2.

12

0.63

1.12

-1

.53

0.28

T5

-0.5

8 1.

79

-0

.62

0.12

-3.0

0 -0

.28

-1.3

3

0.08

0.

86

-0

.18

-2.0

6

-1.9

4 0.

25

-1.9

2

T6

-2.0

3 0.

22

-0

.67

0.74

0.26

-0

.06

0.01

1.27

-0

.85

5.

75

1.10

-1.1

0 -1

.94

-1.3

4

T7

0.17

0.

42

0.

20

0.14

-0.5

0 0.

47

-0.0

4

0.73

-0

.03

0.

30

0.71

-1.0

0 -0

.55

0.59

T8

0.16

-1

.09

0.

72

-0.6

4

3.02

-2

.04

-1.3

2

0.19

-1

.21

4.

75

-2.2

7

0.04

0.

58

2.64

HFL

EX/E

XT

– hi

p fle

xion

/ext

ensi

on; H

AD

D/A

BD –

hip

add

uctio

n/ab

duct

ion;

HIR

/ER –

hip

int

erna

l/ext

erna

l ro

tatio

n; K

EXT/

FLEX

– k

nee

flexi

on/e

xten

sion

; KA

DD

/ABD

– k

nee

addu

ctio

n/ab

duct

ion;

KIR

/ER –

kne

e in

tern

al/e

xter

nal r

otat

ion;

PP

– pr

inci

pal p

atte

rn.

≤

-0.8

La

rge

Dec

reas

e

≥

0.8

Larg

e In

crea

se

> -0

.8 ≤

-0.5

M

oder

ate

to L

arge

Dec

reas

e

≥

0.5

< 0.

8 M

oder

ate

to L

arge

Incr

ease

>

-0.5

≤ -0

.2

Smal

l to

Mod

erat

e D

ecre

ase

≥ 0.

2 <

0.5

Smal

l to

Mod

erat

e In

crea

se

193


Tabl

e 6.

10

Effe

ct s

izes

(Coh

en’s

d) f

or in

divi

dual

cha

nges

of p

artic

ipan

ts in

the

trai

ning

gro

up (n

= 8

; T1

– T8

) fro

m b

asel

ine

to

six-

wee

ks fo

r hip

and

kne

e jo

int m

omen

t pri

ncip

al p

atte

rn (P

P) s

core

s.

Part

icip

ant

HEX

T/FL

EX M

om.

H

ABD

/AD

D M

om.

H

ER/I

R M

om.

K

FLEX

/EX

T Mom

.

KA

BD/A

DD M

om.

K

ER/I

R M

om.

PP1

P

P2

PP3

P

P4

P

P1

PP2

P

P3

P

P1

PP2

PP1

P

P2

PP3

PP1

P

P2

PP3

PP1

P

P2

T1

-0.7

9 0.

51

-0.9

7 0.

52

-0

.87

1.16

1.

34

1.

06

-0.6

1

-0.6

7 0.

01

0.82

-1.4

2 1.

27

-0.5

7

-0.2

2 -0

.30

T2

2.75

1.

25

-0.5

3 1.

34

-3

.03

1.21

3.

27

2.

23

3.64

2.66

-0

.29

2.48

-1.1

1 -1

.96

-1.6

0

0.47

0.

33

T3

-1.6

6 0.

71

-0.0

9 0.

26

-6

.78

-0.5

7 -0

.13

5.

35

1.18

0.06

-0

.27

-0.9

0

-6.8

5 -0

.16

-2.2

3

-1.9

9 -0

.06

T4

-0.1

7 2.

73

-2.4

2 1.

76

-6

.95

0.96

-4

.27

3.

30

1.16

-1.1

9 -1

.92

3.28

-5.9

8 -2

.16

-3.9

8

-1.4

8 1.

36

T5

-0.3

7 -0

.29

0.33

-0

.42

-0

.63

1.71

-1

.95

1.

93

-2.0

6

-0.2

7 -1

.07

0.37

-1.0

1 2.

14

-1.8

8

0.72

-0

.89

T6

0.69

-1

.19

1.35

1.

55

0.

69

0.24

-0

.30

-3

.02

-0.0

9

-2.6

3 0.

71

0.23

-2.0

2 -0

.06

1.47

-0.1

9 2.

12

T7

-0.0

9 -0

.08

-0.5

4 0.

32

-0

.67

0.23

-0

.40

0.

84

-0.1

0

-0.2

1 -0

.83

0.51

-0.5

1 -0

.38

-0.4

1

0.24

-0

.94

T8

-0.8

3 0.

80

-0.3

9 0.

42

-0

.84

-0.1

0 -0

.12

0.

93

0.57

1.25

-0

.64

0.61

-1.9

6 0.

03

-2.0

8

0.65

-1

.44

HEX

T/FL

EX M

om.

– hi

p ex

tens

ion/

flexi

on m

omen

t; H

ABD

/AD

D M

om.

– hi

p ab

duct

ion/

addu

ctio

n m

omen

t; H

ER/IR

Mom

. –

hip

exte

rnal

/inte

rnal

rot

atio

n m

omen

t; K

FLEX

/EX

T M

om. –

kne

e ex

tens

ion/

flexi

on m

omen

t; K

ABD

/AD

D M

om. –

kne

e ab

duct

ion/

addu

ctio

n m

omen

t; K

ER/IR

Mom

. – k

nee

exte

rnal

/inte

rnal

rot

atio

n m

omen

t; PP

– p

rinc

ipal

pat

tern

.

≤

-0.8

La

rge

Dec

reas

e

≥

0.8

Larg

e In

crea

se

> -0

.8 ≤

-0.5

M

oder

ate

to L

arge

Dec

reas

e

≥

0.5

< 0.

8 M

oder

ate

to L

arge

Incr

ease

>

-0.5

≤ -0

.2

Smal

l to

Mod

erat

e D

ecre

ase

≥ 0.

2 <

0.5

Smal

l to

Mod

erat

e In

crea

se

194


Tabl

e 6.

11

Stan

dard

dev

iatio

ns o

f in

divi

dual

cha

nge

scor

es f

rom

bas

elin

e to

six

-wee

ks f

or t

he c

ontr

ol a

nd t

rain

ing

grou

ps

acro

ss m

uscl

e sy

nerg

ies

one,

two

and

thre

e.

Mus

cle

Mus

cle

Syne

rgy

1 (A

ctiv

atio

n ar

ound

IC)

M

uscl

e Sy

nerg

y 2

(Act

ivat

ion

afte

r IC

)

Mus

cle

Syne

rgy

3 (A

ctiv

atio

n pr

ior t

o IC

)

CO

NTR

OL

TRA

ININ

G

C

ON

TRO

L TR

AIN

ING

CO

NTR

OL

TRA

ININ

G

GM

AX

0.

91

2.00

0.59

0.

50

0.

61

0.66

GM

ED

1.34

0.

97

0.

76

1.00

0.60

0.

49

MH

AM

0.

59

1.04

0.21

0.

47

0.

38

0.51

LHA

M

1.41

1.

92

0.

46

1.05

0.57

0.

48

RF

1.08

1.

19

1.

30

1.99

0.21

0.

25

VL

1.17

1.

04

0.

68

1.14

0.51

0.

62

VM

1.

56

1.96

0.62

1.

45

0.

52

0.77

GA

S 1.

00

1.99

0.76

0.

48

0.

44

1.58

IC –

ini

tial

cont

act;

BL –

bas

elin

e; 6

W –

six

wee

ks;

GM

AX

– gl

uteu

s m

axim

us;

GM

ED –

glu

teus

med

ius;

MH

AM

– m

edia

l ha

mst

ring

s; LH

AM

– la

tera

l ham

stri

ngs;

RF

– re

ctus

fem

oris

; VL

– va

stus

late

ralis

; VM

– v

astu

s m

edia

lis; G

AS

– ga

stro

cnem

ius.


195

Table 6.12 Standard deviations of individual change scores from baseline to six-weeks for the control and training groups for joint

rotation principal pattern scores.

Joint Rotation PP CONTROL TRAINING

HFLEX/EXT PP1 17.61 28.95

PP2 9.06 11.08

HADD/ABD PP1 18.66 19.19

PP2 6.74 14.62

HIR/ER PP1 23.42 36.41

PP2 7.48 15.62

KEXT/FLEX PP1 11.92 8.66

PP2 11.67 19.46

KADD/ABD PP1 8.79 18.19

PP2 21.86 43.68

KIR/ER

PP1 5.65 8.71

PP2 13.51 20.15

PP3 10.48 15.78 HFLEX/EXT – hip flexion/extension; HADD/ABD – hip adduction/abduction; HIR/ER – hip internal/external rotation; KEXT/FLEX – knee flexion/extension; KADD/ABD – knee adduction/abduction; KIR/ER – knee internal/external rotation; PP – principal pattern.


196

Table 6.13 Standard deviations of individual change scores from baseline to six-weeks for the control and training groups for joint

moment principal pattern scores.

Joint Moment PP CONTROL TRAINING

HEXT/FLEX Mom.

PP1 9.24 9.40

PP2 0.91 1.44

PP3 0.33 0.75

PP4 0.41 0.36

HABD/ADD Mom.

PP1 0.23 0.50

PP2 0.58 1.88

PP3 0.44 0.37

HER/IR Mom. PP1 0.25 0.56

PP2 0.41 0.50

KFLEX/EXT Mom.

PP1 0.16 0.22

PP2 0.97 1.22

PP3 0.58 0.57

KABD/ADD Mom.

PP1 0.37 0.51

PP2 0.93 1.60

PP3 0.47 0.80

KER/IR Mom.

PP1 0.36 0.58

PP2 0.21 0.20

PP3 0.07 0.07 HEXT/FLEX – hip extension/flexion; HABD/ADD – hip abduction/adduction; HER/IR – hip external/internal rotation; KFLEX/EXT – knee extension/flexion; KABD/ADD – knee abduction/adduction; KER/IR – knee external/internal rotation; PP – principal pattern; Mom. – moment.


197

6.3.4 Risk Level Analysis

As the same data was used as the group analysis, identical main effects for

group and time as presented in Section 6.3.2 of this chapter were found as part

of this analysis. A statistically significant group x time x risk level interaction

was found for PP1, PP2 and PP3 for knee abduction/adduction moments (p =

0.019, η2 = 0.630; p = 0.042, η2 = 0.525 and p = 0.037, η2 = 0.543). Post-hoc analyses

revealed no statistically significant differences (p > 0.05) from baseline to six-

weeks within either the ‘high-’ or ‘low-risk’ participants from both the control

and training groups. As post-hoc tests did not reach statistical significance,

effect sizes for the changes from baseline to six-weeks within the risk level and

group combinations were calculated across the knee abduction/adduction

moment principal patterns (see Table 6.14, Page 198). The largest effects were

observed in the ‘high-risk’ training group (d > 0.91), with medium effects for

the ‘low-risk’ training group (d = 0.50 – 0.80). In contrast, small to medium

effects were observed in the ‘high-risk’ control group (d < 0.50), while only

very small effects were observed in the ‘low-risk’ control group (d < 0.20). No

additional statistically significant main or interaction effects were found

within this analysis.


198

Table 6.14 Effect sizes (Cohen’s d) for changes of ‘high-’ and ‘low-risk’

sub-groups within the control and training groups from baseline to six-weeks for knee abduction/adduction moment

principal pattern (PP) scores.

CONTROL TRAINING

‘HIGH-RISK’ (n = 3)

‘LOW-RISK’ (n = 5)

‘HIGH-RISK’ (n = 3)

‘LOW-RISK’ (n = 5)

KABD/ADD Mom. PP1 0.33 0.04 1.81 0.77

KABD/ADD Mom. PP2 0.21 0.01 0.91 0.50

KABD/ADD Mom. PP3 0.15 0.10 2.11 0.70

KABD/ADD Mom. – knee abduction/adduction moment; PP – principal pattern.


199

6.4 Discussion

Numerous studies have highlighted the ability of ACL injury prevention

programs in altering neuromuscular and biomechanical factors associated

with ACL injury risk.[34-45, 68, 69, 71-75, 79, 166, 201, 213-217] Throughout these studies, there

has been limited examination[71, 82, 83] of individual responses to ACL injury

prevention programs. The purpose of this study was to examine the effects of

an ACL injury prevention program on neuromuscular control and lower limb

biomechanics during a high-risk landing task at the group and individual

level. Overall effects of the injury prevention program were found at the group

level for neuromuscular and biomechanical variables. However, further

examination revealed dissimilar responses to the training program across

individuals.

An overall effect of the injury prevention program at the group level was

observed for a number of neuromuscular and biomechanical variables. The

training group was found to have decreased their gastrocnemius activation

around initial contact, while the control group increased activation of the

gastrocnemius around initial contact following the six-week period. Fleming

et al.[389] found co-contraction of the gastrocnemius with the quadriceps

elevated ACL strain at low knee flexion angles. However, this was identified

under static loading conditions at fixed knee angles. Recent work focusing on

a dynamic landing task suggests that the gastrocnemius may in fact work with

the quadriceps to elevate joint compression and protect the ACL from external

loading.[185] The reduction in gastrocnemius activation at landing seen within

the training group may therefore represent a shift to a more high-risk

neuromuscular strategy, representing a potential negative effect of the injury

prevention program. It is unclear why an increase in gastrocnemius activation


200

was observed in the control group following the six-week period. This study

was undertaken during participant’s competitive netball season, hence players

in the control group continued their regular in-season training practices.

Donnelly et al.[269] found in-season Australian Rules football training produced

changes in muscle activation during a side-step cutting task. These findings

suggest standard in-season training alone is capable of altering muscle

activation characteristics during high-risk sporting tasks, and could be why

these changes in muscle activation were seen within the control group in this

study.

The remaining changes seen from baseline to six-weeks within the training

group appeared favourable in reducing potential ACL injury risk. An increase

in preparatory activation of the medial hamstrings was observed in the

training group compared to controls following the six-week period.

Appropriate preparatory muscle activation is thought to prevent joint collapse

immediately after initial landing contact, thereby protecting the joint during

deceleration.[187] This appears especially pertinent to the hamstrings, as

inadequate preparatory activation may result in lower hamstrings contraction

forces during the initial portion of ground contact and subsequently reduce

knee stability.[120] Reduced preparatory activity of the medial hamstrings has

also been identified as an important factor in elevating ACL injury risk.[177] The

increase in preparatory activation of the medial hamstrings stemming from

the injury prevention program may therefore be useful in reducing ACL injury

risk during landing. The training group was found to exhibit greater external

rotation at the hip during the early stages of landing following the six-week

period. Excessive internal rotation at the hip when the foot is fixed to the floor

can contribute to a position of ‘dynamic valgus’,[162] and greater hip internal

rotation has been linked to larger peak knee abduction moments during


201

landing.[149] A more externally rotated hip position may therefore serve to

protect the knee and ACL by avoiding lower limb postures consistent with

increased ACL loading and injury risk.

Examination of the participants within the training group revealed large

variation in the individual responses to the injury prevention program.

Statistically significant participant by time interactions within the training

group were found across all of the neuromuscular and biomechanical

dependent variables examined. These findings and the effect sizes calculated

for the changes in individual participant data from baseline to six-weeks

highlight the largely variable neuromuscular and biomechanical responses

observed. There did not appear to be a consistent trend with regard to the

individual responses across the neuromuscular and biomechanical variables

examined. Differential effects were observed across the majority of the

dependent variables examined. Where these differential effects were

observed, there did not appear to be a consistent trend across participants. It

could be expected that participants reporting higher attendance to the training

program would consistently receive the largest responses to the training

program, while those reporting a lesser attendance may exhibit opposing

responses. However, this did not appear to be the case. Although certain

participants reported a higher attendance, there did not appear to be a

consistent effect stemming from this. Using a simple measure of attendance to

gauge the degree of participation in the training program may have limited

this interpretation. Recent work has shown that while athletes may attend

injury prevention program sessions, they do not always perform exercises

appropriately to gain a protective benefit.[225] Although this study involved one

session per week led by a researcher, there was minimal control as to how

participants performed the exercises at home. Examining individual responses


202

to a training program, in conjunction with more detailed measures of

adherence, may lead to a better understanding of how the performance of

exercises within a training program can influence outcomes.

The results of this study highlight the potential for variable individual

responses to the same training program. Further work is required to

understand why these variable responses occur, which may lead to better

design of ACL injury prevention programs that provide a benefit to all

involved. Despite this study providing minimal understanding with regard to

why and how these individual responses occur, further insights from the data

can still be made. Examining the responses of participants within the training

group for variables where an overall group change was observed could

provide insight into how overlooking the individual might influence our

understanding of the effectiveness of ACL injury prevention programs.

Previous work has found that a select few participants who experience a large

response to the training can drive group changes.[83] This did not appear to be

the case with regard to gastrocnemius activation around initial contact (i.e.

muscle synergy one), where the majority of participants experiencing a

reduction in activation following the six-week period drove the group change

(see Figure 6.9, Page 203). As noted earlier, this reduction in gastrocnemius

activation may represent a shift towards a high-risk neuromuscular

strategy,[185] potentially rendering these individuals at greater risk of ACL

injury. Despite moderate to large effects being observed across nearly all of the

participants in the training group, this effect was absent in two participants

(i.e. T1 and T3), highlighting how consistent effects of a training program may

not be apparent across all individuals.


203

T1 T2 T3 T4 T5 T6 T7 T8

d 0.24 -1.94 0.01 -1.85 -0.56 -0.56 -0.67 -0.94


baseline to six-weeks for gastrocnemius weighing coefficient values from muscle synergy one. Cohen’s d effect sizes are

presented for the individual changes from baseline to six-

weeks for individual participants (T1 – T8). Positive and negative effect sizes are indicative of increases and decreases,

respectively, in the weighing coefficient value.


204

In contrast, inspection of the individual responses for medial hamstrings

preparatory activation (see Figure 6.10, Page 205) suggested that the group

improvement was mainly driven by three of the eight participants (i.e. T5, T6

and T7). These three participants demonstrated large increases in medial

hamstrings preparatory activation relative to the remainder of the group.

Interestingly, two of the three participants that received a large effect for

improvements in preparatory medial hamstrings activation (i.e. T5 and T6)

had the highest values for this variable at baseline. In contrast, the two

participants who recorded the lowest medial hamstrings preparatory

activation at baseline (i.e. T1 and T4) received minimal to no effect in this area

from the training. This is potentially problematic, as low preparatory

activation of the medial hamstrings may represent a high-risk neuromuscular

strategy.[120, 177] Although the group results suggested the program had a

beneficial effect on preparatory medial hamstrings activation, it appears this

effect was limited or diminished in those participants that may have required

it the most.


205

T1 T2 T3 T4 T5 T6 T7 T8

d 0.07 -0.33 0.06 -0.12 0.95 1.41 1.97 -0.06

Figure 6.10 Individual changes of participants in the training group from baseline to six-weeks for medial hamstrings weighing

coefficient values from muscle synergy three. Cohen’s d effect

sizes are presented for the individual changes from baseline to six-weeks for individual participants (T1 – T8). Positive and

negative effect sizes are indicative of increases and decreases,

respectively, in the weighing coefficient value.


206

A similar pattern was observed when examining individual responses for hip

internal/external rotation PP2, whereby the group reduction also appeared to

be driven by three participants (i.e. T3, T4 and T8) (see Figure 6.11, Page 207).

In this scenario, however, two of the three participants (T4 and T8) who

experienced large effects exhibited the highest PP-Scores at baseline. Higher

scores for this biomechanical pattern represented larger hip internal rotation,

which could be considered a high-risk biomechanical strategy in the early

deceleration phase of landing.[162] Therefore, experiencing a large reduction

could be considered beneficial, as it represents a shift to a more externally

rotated hip position during this phase of landing. In contrast, the third

participant to receive a large effect for this variable (T3) exhibited the lowest

scores across all participants at baseline, suggesting they already

demonstrated a large degree of hip external rotation prior to the training. The

remaining participants only exhibited small to no response for this variable,

with one participant (T7) demonstrating a moderate increase following the six-

week period. Despite demonstrating a similar, potentially high-risk score to

the participants who received the largest beneficial effects from the training,

the opposite effect was observed for this individual. The increase in score

indicates an increase in hip internal rotation during the early deceleration

phase of landing following training, which could potentially increase the risk

of ACL injury by contributing to a ‘dynamic valgus’ position or large knee

abduction moments.[149, 162]


207

T1 T2 T3 T4 T5 T6 T7 T8

d -0.26 0.04 -1.65 -2.78 -0.28 -0.06 0.47 -2.04


baseline to six-weeks for hip internal (IR) / external (ER)

rotation principal pattern (PP) two PP-Scores. Cohen’s d effect sizes are presented for the individual changes from baseline to

six-weeks for individual participants (T1 – T8). Positive and

negative effect sizes are indicative of increases and decreases, respectively, in the PP-Score.


208

It has been proposed that individuals who demonstrate a pre-existing high-

risk pattern at baseline receive the greatest improvements in these areas

following training.[71, 83, 84] This appeared to be the case in this study, as ‘high-

risk’ individuals (as deemed by higher peak knee abduction moments during

the landing task) were found to respond more favourably to the injury

prevention program as demonstrated by reductions in frontal plane knee

moments. Knee abduction/adduction moment patterns were the only

variables where a statistically significant group by time by risk level

interaction was identified. While this was the case, post-hoc analyses revealed

no statistically significant differences from baseline to six-weeks across any of

the groupings. These non-significant results were to be expected due to the

limited sample sizes within each of the group and risk level combinations.

Examination of effect sizes from baseline to six-weeks across the three knee

abduction/adduction moment patterns revealed the largest effects were seen

for participants deemed as ‘high-risk’ within the training group, followed by

the ‘low-risk’ participants from the training group, with small effects seen for

either risk levels within the control group. The changes in knee

abduction/adduction moment patterns seen within the training group were

largely favourable, particularly for ‘high-risk’ individuals (see Figure 6.12,

Page 209). At six-weeks, ‘high-risk’ individuals within the training group were

able to reduce the knee abduction moments experienced during landing to the

same levels ‘low-risk’ individuals exhibited at baseline. ‘Low-risk’ individuals

within the training group also saw a reduction in knee abduction moments

during landing, albeit smaller.


209

Figure 6.12 Mean knee abduction / adduction moment from initial contact

(IC) through to 150 ms post IC during the leap landing task for

the ‘high-’ and ‘low-risk’ sub-groups from the training group at baseline and six-weeks.

The reductions in knee abduction moments did appear to cause larger knee

adduction moments in the latter stages of landing. The presence of knee

abduction and adduction moments early on and late in the landing,

respectively, is characteristic of netball leap landing manoeuvres.[248]

Considering the link between larger knee abduction moments and increased

ACL loads and injury risk, particularly during the early deceleration phase of

landing,[20-23, 28, 160, 390] reducing the knee abduction moment at the cost of an

increased knee adduction moment may be advantageous. The finding that

individuals receive a more favourable response to training in the area they


210

exhibit a pre-existing high-risk biomechanical strategy (i.e. peak knee

abduction moments) fits with the findings of previous work.[71, 83, 84] However,

this may only be applicable to certain biomechanical or neuromuscular factors.

Individuals exhibiting low levels of preparatory medial hamstrings activation

exhibited minimal to no increases in activation, while one participant

demonstrating potentially high-risk transverse plane hip motion did not

improve in this area following training. The differing effects could be due to

the high-risk movement strategies targeted by the injury prevention program

used in this study. A heavy focus of the D2E programs feedback schedule is

on landing ‘softly’ while keeping the hip and knee aligned in the frontal

plane.[258] Successful application of these movement strategies should result in

minimised landing forces and ‘dynamic valgus’ postures. These

biomechanical factors have been linked to knee abduction moments during

high-risk sporting tasks.[149, 164, 304] Addressing these biomechanical deficits

would subsequently have a beneficial effect in reducing the knee abduction

moments experienced during landing. The combination of a pre-existing risk

pattern at baseline in conjunction with an injury prevention program targeting

frontal plane alignment may have been why larger responses in frontal plane

knee moments were observed within the ‘high-risk’ individuals of the training

group.

The broad implications of this study are that a generic approach to ACL injury

prevention programs may not be effective across all individuals. A potential

shortcoming of current injury prevention programs is that they focus on a

select number of biomechanical deficits, or simultaneously attempt to address

a variety of possible biomechanical deficits across all individuals.[358] The

injury prevention program used within this study focused on movement

strategies that would reduce frontal plane knee moments, as these have been


211

associated with increased ACL loading and injury risk.[20-23, 28, 160, 390]

Subsequently, individuals who demonstrated a pre-existing risk pattern in this

area (i.e. high peak knee abduction moments) received a greater benefit in

reducing frontal plane knee moments during the leap landing task. Additional

variations in individual responses were observed, however there did not

appear to be a consistent trend with regard to these responses. Tailored injury

prevention programs that target the high-risk neuromuscular or

biomechanical strategies of the individual may be the most effective in

reducing an individual’s level of ACL injury risk.

Despite the results of this study, it is important not to discount generic ACL

injury prevention programs applied in a universal manner. While possibly

having a smaller prophylactic effect, universal neuromuscular training may

mitigate the potential increase in injury risk associate with season-long

participation in high-risk sports.[269] Further, while targeted programs may be

more effective in reducing ACL injury risk the cost-effectiveness of this

approach could be considered questionable. Implementing specifically

tailored injury prevention programs across individuals is unlikely to be a cost-

effective strategy, particularly within wider community settings where time

and staff resources may be limited. Striking a balance between the generic and

individualised approaches may be optimal. A recent study by Pappas et al[358]

found that individuals could be categorised into specific sub-groups, with

each group demonstrating a combination of biomechanical deficits during a

side-step cutting task. Tests that could categorise individuals to these deficit

profiles may be useful for prescribing targeted ACL injury prevention

programs.[358] This sub-group approach to tailoring injury prevention

programs may be more effective and efficient than current generic injury

prevention program practices,[358] and may minimise some of the practicality


212

issues stemming from highly individualised program design. Screening

methods that are practical and cost-effective to implement, while also having

the capacity to identify individuals who demonstrate high-risk movement

strategies therefore become an important component in the design of targeted

ACL injury prevention programs. The applicability of current ACL injury risk

screening methods to wider community settings, as well as their capacity to

identify high-risk movement strategies during a high-risk sport-specific task

will subsequently be explored in the following chapters of this thesis.

This study is not without limitations. First, a small sample size was used for

both the control and training groups. Previous studies investigating

neuromuscular and biomechanical responses to an ACL injury prevention

program have used a larger cohort.[40, 68, 76, 213, 216] However, smaller sample sizes

have been used in studies that have investigated group[41, 44, 74, 77-80, 391] and

individual[82, 83] responses to training. Despite the small sample, statistically

significant differences were observed suggesting there was sufficient power to

observe changes in certain neuromuscular and biomechanical variables. The

limitation of sample size was particularly evident within the risk level

analysis, with only three participants able to be allocated to the ‘high-risk’

groups. This limited the capacity of the study to identify statistically

significant differences for this particular analysis. Future investigations

employing this type of analysis would likely benefit from a larger sample size

to ensure adequate participant numbers across sub-groups. A larger sample

size may also allow for stratification into additional sub-groups, which may

provide a better understanding of how individuals with certain characteristics

respond differently to training. Second, the duration of the injury prevention

program used may have limited the generalised training effects observed.

‘Down 2 Earth’ is designed as a six-week ACL injury prevention program.[258]


213

This program was used in this study as it was specifically developed for ACL

injury prevention within a netball-specific cohort.[254, 258] However, it is

important to acknowledge that the duration and subsequent volume of this

program may have impacted on the generalised training effects observed.

While a training effect was seen with certain neuromuscular and

biomechanical variables, these may have been more extensive with an injury

prevention program of longer duration. A review by Bien[392] suggested injury

prevention programs should be eight or more weeks to ensure sufficient

neuromuscular training effects. Despite this, it was acknowledged that limited

evidence exists regarding optimal training duration for obtaining the most

effective results.[392] Shorter duration programs (i.e. six-weeks or less) have

been shown to have a protective effect against ACL injuries or injury risk

factors.[44, 65, 166, 393] Nonetheless, future studies investigating individual

adaptations to ACL injury prevention programs may benefit from utilising

training programs of longer duration (e.g. eight-weeks or longer) where more

generalised training effects may be observed. Lastly, strength and

performance measures were not collected prior to and after the training

program. Understanding how lower limb strength measures changed

subsequent to the training may have provided further assistance in

interpreting changes in muscle activation data.


214

6.5 Conclusions

This study found an ACL injury prevention program induced neuromuscular

and biomechanical adaptations during the performance of a high-risk landing

task. However, examination of individual data revealed dissimilar responses

to the training program. The findings of this study in conjunction with

previous work[71, 82, 83] suggest that variation across individual responses to

training may be inevitable. Future studies investigating the effects of ACL

injury prevention programs must therefore include a focus on the individual

to fully understand the effectiveness, or lack thereof, of the training program.

Individuals displaying a pre-existing high-risk strategy may receive a greater

benefit from ACL injury prevention programs. However, this may only be

applicable if the program targets the relevant high-risk strategy. Further

efforts must be made to ensure that ACL injury prevention programs are

providing a benefit to all involved, particularly in the areas where they

demonstrate neuromuscular or biomechanical strategies that may increase

their risk of ACL injury. Injury prevention programs that target

neuromuscular or biomechanical deficits relevant to the individual may be

useful in ensuring beneficial responses across all involved.

215

Chapter Seven: A Systematic Evaluation of Field-Based Screening Methods for the Assessment of Anterior Cruciate Ligament Injury Risk

Adapted from: Fox AS, Bonacci J, McLean SG, Spittle M, Saunders N. A systematic

evaluation of field-based screening methods for the assessment of anterior cruciate

ligament (ACL) injury risk. Sports Med. 2016; 46(5): 715-35.

7.1 Introduction

There is substantial support for the use of neuromuscular training as part of

anterior cruciate ligament (ACL) injury prevention programs in reducing ACL

injury rates[5, 65-67] and inducing biomechanical and neuromuscular adaptations

proposed to reduce ACL injury risk.[34-45, 68, 69, 71-75, 79, 166, 201, 213-217] A recent cost-

effective modelling study predicted universal neuromuscular training of all

athletes could reduce the incidence of ACL injury from 3% to 1.1% per season;

and would be more cost-effective than no training or screening combined with

training for high-risk athletes.[226] However, potential barriers to implementing

universal neuromuscular training exist. While ACL injury prevention

programs have proved successful, the level of adoption and compliance is

low[227, 228] and their deployment may not be effective in “real-world”

situations.[223, 229-231] Current neuromuscular training programs often involve

the whole team, require considerable time investment and may detract from

sport-specific skill training.[84] These factors may deter coaches from

implementing such programs within their pre-season or in-season

Field-Based Screening Methods for ACL Injury Risk

216

training.[84, 229] Screening methods that identify athletes at-risk of ACL injury

may improve coach intent to implement ACL injury prevention programs for

these athletes.[84] In addition, the ability to identify individuals ‘at-risk’ of ACL

injury appears to be an important step in maximising the effectiveness of ACL

injury prevention programs. Myer et al.[84] found female athletes deemed as

‘high-risk’ (knee abduction moment > 25.25 Nm during a drop vertical jump)

received a greater prophylactic effect from neuromuscular training compared

to ‘low-risk’ athletes. Similar to prevention programs, injury risk screening

also requires time investment from coaches and athletes. It has been suggested

that the equipment and time costs associated with screening for high-risk

athletes is not cost-effective.[226] However, there may be a role for coupling

screening with more targeted and intensive neuromuscular training programs

if the screening methods used have high accuracy with minimal cost.[226]

Identification of screening methods that are both predictive of injury and cost-

effective may provide further support for the integration of screening within

ACL injury prevention efforts.

A screening method must be appropriate and relevant to its selected purpose

and it is imperative that the constructs or criteria examined align with the

constructs of interest.[394] Screening methods for ACL injury risk should

therefore be linked with the mechanical aetiology of ACL injuries in order to

be effective in identifying ‘at-risk’ athletes. Therefore, screening methods that

identify athletes who employ movement strategies linked to increased ACL

loading are likely to be the most effective in classifying ACL injury risk.

Laboratory-based measures (e.g. three-dimensional motion analysis, force

plates) can accurately estimate segment motions and joint kinetics during

athletic tasks, providing an option to evaluate movement strategies or

techniques that may predispose an athlete to ACL injury. This approach,


217

however, requires extensive laboratory-based equipment and training to

complete. The cost of three-dimensional motion analysis is in the range of

$1000 per athlete per test.[87] The cost-effectiveness of this approach is therefore

questionable, and is likely prohibitive to the majority of the wider community.

Screening methods that can be completed in a field or clinical setting may be

more cost-effective and applicable for wider community use in identifying

dysfunction that may lead to an ACL injury. Field-based screening methods

for identifying ACL injury risk are emerging,[85-88, 91] however assessment of

their applicability for use in wider community settings is yet to be undertaken.

This is required to identify the most appropriate method(s) for field-based

evaluation of ACL injury risk.

The purpose of this systematic review was to evaluate and compare field-

based screening methods for ACL injury risk to determine their efficacy of use

in wider community settings. Specifically, this review aimed to: (i) examine

whether the identified screening methods are appropriate for their intended

use based on whether the method aligns with the constructs of interest, has

theoretical and empirical support, and is designed for the population of

interest; (ii) examine the technical adequacy of the identified screening

methods based on their reliability and predictive validity; and (iii) determine

the usability of the identified screening methods based on the balance of the

costs and benefits, feasibility of administration, and utility of outcomes

associated with the method. It was hypothesised that current field-based

screening methods would include scoring criteria appropriate for identifying

high-risk movement strategies associated with ACL injury risk, however,

would have a limited ability to predict future ACL injuries. In addition, some

of the screening methods would be more applicable for use in wider


218

community settings than others due to having minimal equipment and time

requirements to complete.

7.2 Methodology

7.2.1 Literature Search

A literature search was performed to retrieve articles pertaining to field-based

screening or assessment methods used to evaluate ACL injury risk. An

electronic search was conducted on SportDISCUSTM, Medline, AMED and

CINAHL (January 1990 – July 2015) databases. Searches were limited to

English language results only to minimise time and costs related to translation.

The following search term groups were used: group (i) “ACL”, “anterior

cruciate ligament”, “knee injur*” and “ACL injur*”; and group (ii)

“screening”, “injury risk”, “screening tool”, “screening test”, “clinical

screening tool”, “field* screening”, “risk screening”, “preseason screening”,

“mass screening”, “team screening”, “risk assessment” and “evaluation tool”.

Within groups; search terms were connected with OR, and between groups

with AND. All terms were searched for in the titles and abstracts of articles.

The results of all searches were taken together and duplicates removed.

Once the initial literature selection process was complete and screening

methods identified, a secondary search was undertaken to identify any further

studies relevant to each of the screening methods. A search of the same

databases over the same time period using key words included in the name or

description of the screening methods identified (see Table 7.2, Page 224). All

terms were searched for in the titles and abstracts of articles. All studies

identified by the secondary literature search were screened against the same

inclusion/exclusion criteria applied to initially identified studies. Once all


219

relevant literature from database searching was identified, the reference lists

of all included articles were examined to identify any further studies for

inclusion in the review.

7.2.2 Literature Selection

The titles and abstracts of all studies identified by the search strategy were

reviewed for potential relevance. If sufficient information was not contained

in the title and abstract to determine relevance, the full text was evaluated. The

full text of potentially relevant articles were independently analysed by two

reviewers for final inclusion, based on the following inclusion criteria: (i) the

screening method was developed as a means to determine ACL injury risk;

and (ii) the screening method could be undertaken outside of a laboratory

setting (i.e. field and clinical settings). Articles were excluded from the review

if: (i) the article was only in abstract form; or (ii) the screening method was not

applied in an effort to identify injury risk. Any discrepancies between the two

reviewers were resolved with a consensus meeting. If consensus could not be

reached, a third reviewer was consulted. Following the literature selection

process, included studies were grouped according to the specific screening

method they examined.

7.2.3 Methodological Quality Assessment

Included studies were examined for methodological quality using a modified

version of the Cochrane Group on Screening and Diagnostic Test

Methodology quality assessment tool.[395] Eleven criteria (subject

inclusion/exclusion criteria described; information to identify setting

described; study design; age and sex reported; screening method sufficiently

described and reported; groups comparable at baseline; blinding of

investigators to confounding/prognostic factors; mean and variance of


220

variables reported; reliability reported; missing data reported or explained;

and outcome clearly defined with methods explained) which logically applied

to the majority of studies included in this review were used, resulting in a

maximum score of eleven. No study was excluded and no additional analysis

was undertaken on the basis of the assessment of study quality. A full

summary of the methodological quality assessment undertaken on included

studies is presented in Appendix H (see Page 415).

Using relevant articles, an overall description of each screening method,

including the criteria used for scoring, was acquired. Screening methods were

evaluated against the relevant criteria surrounding ‘appropriateness for the

intended use,’ ‘technical adequacy’ and ‘usability’ from Glover and Albers[394]

(see Table 7.1, Page 222). In order to examine each screening method against

these criteria, the following data were extracted pertaining to each method:

1. Has the screening method been prospectively evaluated against ACL

injury occurrences, and if so, was the method successful in identifying

future ACL injuries?

2. Has the screening method been assessed for validity, and if so, how has

this been undertaken?

3. Has intra- and inter-rater reliability for the screening method been

assessed and established?

4. What cut-off value, if any, has been identified for labelling an

individual as ‘at-risk’?

5. What additional, if any, equipment is required to undertake the

screening method?


221

6. What additional, if any, training is required for an individual or group

to capably perform the screening method?

7. Approximately how long does it take to screen one athlete?

8. Is the screening method applicable across multiple sports and athletes?

These criteria were chosen to ensure the identified ACL field-based screening

methods could be evaluated based on their accuracy in identifying future ACL

injury risk, validity in identifying high-risk movements and injury risk factors,

reliability of the method within and between users, and how well the method

could be applied across various settings. Extracted data were used to evaluate

and compare the identified ACL field-based screening methods for their use

in wider community settings.


222

Table 7.1 Considerations and questions for evaluating identified screening methods. Reproduced with permission from Glover

and Albers.[394]

Consideration Questions

Appropriateness for intended use Alignment with constructs of interest

Are the criteria measured as a part of the screening method relevant for determining an individual’s risk status?

Theoretical and empirical support

Have the format and content of the screening method been validated in previous research?

Population fit Is the screening method designed for the population of interest?

Technical adequacy Test-retest (intra-rater) reliability Is the screening method consistent over

time? Inter-rater reliability Is the screening method consistent across

raters? Predictive validity

Sensitivity Of those actually at risk, what proportion is correctly identified?

Specificity Of those actually not at risk, what proportion is correctly identified?

Positive predictive value Of those identified as at risk, what proportion is correctly identified?

Negative predictive value Of those identified as not at risk, what proportion is correctly identified?

Usability Balance of costs and benefits Are the costs associated with the screening

method reasonable? Feasibility of administration Are personnel able to administer the

screening method? Utility of outcomes Can stakeholders understand the

implications associated with the screening methods outcomes? Are the outcomes useful for guiding instruction/intervention?


223

7.3 Results

7.3.1 Search Findings and Study Selection

The electronic search yielded 1,077 citations (SportDISCUSTM – 286, Medline –

499, AMED – 68, CINAHL – 224). Following the removal of duplicates,

elimination of studies on the basis of title and abstract, 25 studies were

identified as potentially relevant and screened against the inclusion/exclusion

criteria. A total of 13 studies[396-408] were excluded due to not meeting the

inclusion/exclusion criteria, resulting in 12 included studies[85, 86, 88-94, 104, 105, 409]

meeting all relevant criteria. Included studies were then grouped according to

the screening method they examined. Five ACL screening methods were

identified via the literature search, those being the Landing Error Scoring

System (LESS),[91, 94, 104, 105, 409] Clinic-Based Algorithm,[88-90, 93] Observational

Screening of Dynamic Knee Valgus (OSDKV),[85] 2D-Cam Method[92] and Tuck

Jump Assessment.[86]

Following the initial literature search and selection process, the secondary

search using the key words and terms deemed relevant to the identified

screening methods (see Table 7.2, Page 224) was undertaken.


224

Table 7.2 Search terms used and search results of the secondary literature search.

Screening Method Search Terms Search Results

Landing Error Scoring System

“landing error scoring system” 27 total citations

SportDISCUSTM – 19

MedLine – 20

AMED – 4

CINAHL – 9

Clinic-Based Algorithm “clinic-based nomogram” OR “clinic-based prediction tool” OR “clinic-based algorithm” OR “clinic-based measurements” OR “clinic based nomogram” OR “clinic based prediction tool” OR “clinic based algorithm” OR “clinic based measurements”

13 total citations

SportDISCUSTM – 3

MedLine – 10

AMED – 0

CINAHL – 2

Observational Screening of Dynamic Knee Valgus

“observ* risk screening” AND “dynamic knee valgus” OR “knee valgus”

1 total citations

SportDISCUSTM – 1

MedLine – 1

AMED – 0

CINAHL – 1

2D-Cam Method “two-dimensional cam* method” OR “2D cam method” OR “two dimensional cam* method” OR “2D-cam method”

1 total citations

SportDISCUSTM – 1

MedLine – 1

AMED – 0

CINAHL – 1

Tuck Jump Assessment “tuck jump assessment” 6 total citations

SportDISCUSTM – 7

MedLine – 3

AMED – 1

CINAHL – 5


225

From the secondary literature search, eight further studies that met all criteria

were included in the review. Specifically, one study relating to the LESS,[410]

three studies relating to the Clinic-Based Algorithm,[87, 411, 412] and three studies

relating to the Tuck Jump Assessment.[256, 413, 414] No relevant studies pertaining

to the OSDKV or 2D-Cam Method were identified via the secondary literature

search. A search of all included studies’ reference lists resulted in one study

relating to the Tuck Jump Assessment[415] being added to the review. Therefore,

a final total of 20 studies[85-94, 104, 105, 256, 409-415] were included in the review. A

diagrammatic representation of the search strategy and study selection

process is shown in Figure 7.1 (see Page 226).


226

Figure 7.1 Diagrammatic representation of search strategy results and study selection process. 1 = [396, 398, 399, 402, 408] 2 = [401, 403, 404, 406] 3 = [397,

405] 4 = [407] 5 = [91, 94, 104, 105, 409] 6 = [88-90, 93] 7 = [85] 8 = [92] 9 = [86] 10 = [410] 11

= [87, 411, 412] 12 = [256, 413, 415, 416]. LESS – Landing Error Scoring System; ACL – anterior cruciate ligament.


227

7.3.2 Screening Methods Identified

7.3.2.1 Landing Error Scoring System (LESS)

The Landing Error Scoring System (LESS) involves the assessment of an

individual’s jump-landing technique based on ‘errors’ identified during the

performance of a bilateral drop vertical jump task.[91] The drop vertical jump

task is performed from a height of 30 cm, with athletes required to land at a

distance of 50% of their height from the box and immediately perform a

maximal vertical jump (see Figure 7.2).[91]

Figure 7.2 Demonstration of the jump-landing task used for the Landing Error Scoring System. Reproduced with permission from

Padua et al.[91]

The LESS involves 17 scored items or ‘errors’, with items dedicated to the

assessment of: lower extremity and trunk position at initial ground contact

(trunk, hip, knee, and ankle flexion angles at initial contact; knee valgus angle

at initial contact; lateral trunk flexion angle at initial contact); foot position at

initial contact and the time between initial contact and maximum knee flexion

(stance width – wide or narrow; foot position – toe in or out; symmetry of

initial foot contact); lower extremity and trunk movements between initial

contact and maximum knee flexion (knee flexion and valgus displacement;


228

trunk and hip flexion at maximum knee flexion); overall sagittal plane

movement; and the general perception of landing quality.[91] Lower LESS

scores are indicative of better jump-landing technique and is taken as an

indication of reduced ACL injury risk.[91] Three trials are performed with

landings recorded using digital video cameras placed 3.54m from the landing

area in frontal and sagittal views.[91] These trials are then assessed using the

scoring items to determine the individuals LESS score out of a total of 19.[91]

An ‘error’ is given when the criteria is met on two of the three trials, with the

individual’s total LESS score out of 19 attained by summing these errors.[91]

A real-time version of the LESS (LESS-RT) was also identified.[410] The jump-

landing task used for the LESS-RT is identical to the originally developed

LESS, however, scoring is undertaken without the assistance of video cameras

and with an additional fourth landing trial completed.[410] Trials one and two

are assessed from a frontal view, while trials three and four are assessed from

a sagittal view.[410] The four trials are used to assess ten scoring items relating

to stance width, foot rotation, symmetry of feet contact, knee and trunk frontal

plane motion, foot contact pattern, knee and trunk sagittal plane motion, and

overall impression of sagittal and frontal plane motion.[410] As with the LESS,

a lower LESS-RT score is indicative of better jump-landing technique and

taken as an indication of reduced ACL injury risk.[410]

A further, more simplified version of the LESS was also identified.[409] The i-

Landing Error Scoring System (iLESS) involves the performance of a single

drop vertical jump task identical to that used in the LESS / LESS-RT, with this

landing allocated to one of two movement classifications.[409] Athletes are

deemed ‘low-risk’ if they exhibit a “good movement pattern,” characterised

by no knee valgus at initial foot contact, no knee valgus displacement from

initial contact to maximum knee flexion, landing with greater than 30 degrees


229

of knee flexion, undergoing greater than 30 degrees of knee flexion

displacement during landing, and exhibiting minimal to no sound upon

landing.[409] In contrast, athletes are deemed as ‘high-risk’ if they exhibit a

“poor movement pattern,” characterised by moderate to large knee valgus

position at initial foot contact, moderate to large knee valgus displacement

from initial contact to maximum knee flexion, landing with less than 30

degrees of knee flexion, undergoing less than 30 degrees of knee flexion from

initial contact to maximum knee flexion, and exhibiting loud sound upon

landing.[409] Scoring is completed via observation of frontal plane video

footage.[409]

7.3.2.2 Clinic-Based Algorithm

The Clinic-Based Algorithm uses a combination of anthropometric measures,

landing biomechanics, and strength measures as a means to predict the

probability of a female athlete exhibiting a high (> 21.74 Nm) peak knee

abduction moment (pKAM) during a jump-landing task. Anthropometric

measures assessed include tibial length and body mass, while quantities of

knee valgus motion and knee flexion range of motion (ROM) during a bilateral

drop vertical jump task from a 31 cm jump-box are used to assess jump-

landing technique.[87, 88, 90] Isokinetic knee extension/flexion strength

(QuadHam ratio) is also used, however, if an isokinetic testing device is

unavailable, a surrogate value for QuadHam ratio can be used based on the

individual’s body mass (QuadHam ratio = [0.01 x body mass] + 1.10) or by

using a previously determined mean value of 1.53.[87, 89] Once collected, these

values are applied to a specifically developed algorithm (see Figure 7.3A, Page

230) which allocates a point value to each measure.[87, 88, 90] The points for each

measure are then summed, with this total value used to estimate the


230

probability of the individual exhibiting a high peak knee abduction moment

during a jump-landing task (see Figure 7.3B).[87, 88, 90]

A.

B.

Figure 7.3 A) Specifically developed nomogram sheet for use with the

Clinic-Based Algorithm. B) Completed example of nomogram

to predict probability of high knee loads. Reproduced with permission from Myer et al.[89] and Myer et al.[87]

7.3.2.3 Observational Screening of Dynamic Knee Valgus (OSDKV)

The Observational Screening of Dynamic Knee Valgus (OSDKV) method

involves the labelling of athletes as either ‘high-’ or ‘low-risk’ based on video

footage of three bilateral drop vertical jump task performances from a 31 cm

jump-box.[85] A set of observational risk screening guidelines are used to

allocate athletes into these groups. Athletes are to be evaluated as ‘high-risk’

if their “patella moves inwards and ends up medial to the first toe,” and ‘low-


231

risk’ if their “patella lands in line with the first toe” (see Figure 7.4).[85] Athletes

are deemed ‘high-risk’ if they meet the high-risk guidelines in any of the three

trials performed.[85]

Figure 7.4 Example of trials deemed as “high-risk” (left) and “low-risk”

as per the Observational Screening of Dynamic Knee Valgus

(OSDKV) method guidelines. Reproduced with permission from Ekegren et al.[85]

7.3.2.4 2D-Cam Method

The 2D-Cam Method involves the assessment of frontal plane knee angle from

a single digital video camera during various athletic task (side-step, side-jump

and shuttle run) performances.[92] Video footage of task performance is taken

from a camera placed in the frontal plane at a point two metres from where

the athlete performs the athletic manoeuvre.[92] This footage is then imported

into video analysis software where estimates of the hip, knee and ankle joint

centres are identified and digitised for each frame of video to determine 2D

frontal plane knee angle during the stance phase of the athletic manoeuvre.[92]

Due to the relationship between peak knee valgus and ACL injuries, a greater

magnitude of stance phase frontal plane knee motion during athletic task

performances is considered as increased injury risk.[92]


232

7.3.2.5 Tuck Jump Assessment

The Tuck Jump Assessment involves athletes performing repeated tuck jumps

bringing their knees to their chest on-the-spot for ten seconds (see Figure 7.5),

with their jump-landing technique scored using ten set criteria.[86]

Figure 7.5 Demonstration of the jump-landing task used for the Tuck

Jump Assessment. Adapted and reproduced with permission

from Myer et al.[86]

Deficits in the athlete’s jump-landing technique are identified via these

criteria, with each deficit given a score of one, resulting in a total score out of

ten.[86, 256] Scoring can be undertaken visually in real-time, or by using photos

or video taken via digital cameras.[86, 256] The ten scoring items are split into

categories, with three criteria relating to knee and thigh motion (lower

extremity valgus at landing; thighs reaching parallel; unequal thighs side-to-

side), four criteria relating to foot position during landing (foot placement

shoulder width apart; foot placement not parallel; unequal foot contact timing;

excessive landing contact noise), and three criteria relating to plyometric

technique (pauses between jumps; technique declines prior to ten seconds;

does not land in same footprint).[86] Lower scores are indicative of fewer

deficits in the athlete’s technique and taken as an indication of reduced ACL

injury risk.


233

7.3.3 Extracted Data

Table 7.3 (see Page 234) presents a summary of the data extracted for each of

the screening methods.

234


Tabl

e 7.

3 Ex

trac

ted

data

for t

he id

entif

ied

fiel

d-ba

sed

scre

enin

g m

etho

ds.

Cri

teri

a LE

SS /

LESS

-RT

/ iL

ESS

Clin

ic-B

ased

A

lgor

ithm

O

bser

vatio

nal S

cree

ning

of

Dyn

amic

Kne

e V

algu

s 2D

-Cam

M

etho

d Tu

ck Ju

mp

Ass

essm

ent

Pros

pect

ive

eval

uatio

n of

ab

ility

to

pred

ict A

CL

inju

ries

LESS

sco

res

did

not

dist

ingu

ish

betw

een

mat

ched

AC

L- a

nd n

on-

inju

red

mal

e an

d fe

mal

e hi

gh sc

hool

and

col

legi

ate

leve

l ath

lete

s.[1

05] L

ESS

scor

es

of fi

ve a

nd a

bove

pre

dict

ed

AC

L in

juri

es w

ith sp

ecifi

city

an

d se

nsiti

vity

val

ues

of 8

6%

and

64%

, res

pect

ivel

y; a

nd

resu

lted

in P

PVs a

nd N

PVs

of 1

.4%

and

99.

8%,

resp

ectiv

ely

in a

you

th

fem

ale

popu

latio

n.[1

04] N

o pr

ospe

ctiv

e ev

alua

tion

of th

e ca

paci

ty o

f the

LES

S-RT

or

iLES

S to

pre

dict

AC

L in

juri

es w

as id

entif

ied.

Ass

ocia

tion

betw

een

pKA

M

dete

rmin

ed b

y th

e C

linic

-Ba

sed

Pred

ictio

n To

ol a

nd

AC

L in

jury

ass

esse

d us

ing

mat

ched

AC

L- a

nd n

on-

inju

red

fem

ale

high

sch

ool

and

colle

ge a

thle

tes.

No

asso

ciat

ion

betw

een

pred

icte

d pK

AM

and

AC

L in

jury

was

iden

tifie

d.[4

12]

No

pros

pect

ive

eval

uatio

n of

th

e ca

paci

ty o

f the

OSD

KV

m

etho

d to

pre

dict

AC

L in

juri

es w

as id

entif

ied.

No

pros

pect

ive

eval

uatio

n of

th

e ca

paci

ty o

f the

2D

-Cam

M

etho

d to

pre

dict

AC

L in

juri

es w

as id

entif

ied.

No

pros

pect

ive

eval

uatio

n of

th

e ca

paci

ty o

f the

Tuc

k Ju

mp

Ass

essm

ent m

etho

d to

pr

edic

t AC

L in

juri

es w

as

iden

tifie

d.

235


Tabl

e 7.

3 Ex

trac

ted

data

for t

he id

entif

ied

fiel

d-ba

sed

scre

enin

g m

etho

ds (c

ontin

ued)

.

Cri

teri

a LE

SS /

LESS

-RT

/ iL

ESS

Clin

ic-B

ased

A

lgor

ithm

O

bser

vatio

nal S

cree

ning

of

Dyn

amic

Kne

e V

algu

s 2D

-Cam

M

etho

d Tu

ck Ju

mp

Ass

essm

ent

Val

idity

as

sess

men

t un

derta

ken

LESS

sco

res

wer

e as

soci

ated

w

ith 3

D m

easu

res o

f hip

fle

xion

and

add

uctio

n; k

nee

flexi

on a

nd v

algu

s; kn

ee

valg

us a

nd in

tern

al ro

tatio

n m

omen

ts; h

ip e

xten

sion

, ad

duct

ion

and

inte

rnal

ro

tatio

n m

omen

ts, a

nd A

TS

forc

e du

ring

a d

rop

vert

ical

ju

mp

task

in m

ales

and

fe

mal

es.[9

1] A

gree

men

t be

twee

n LE

SS s

cori

ng a

nd

3D m

easu

res

was

cla

ssifi

ed

as ‘e

xcel

lent

’ (84

-100

%) f

or

six

scor

ing

item

s, ‘m

oder

ate’

(6

8-74

%) f

or fo

ur s

cori

ng

item

s, a

nd ‘p

oor’

(10-

42%

) fo

r thr

ee s

cori

ng it

ems

in

fem

ale

colle

giat

e at

hlet

es.[9

4]

No

form

of v

alid

ity

asse

ssm

ent h

as b

een

com

plet

ed o

n th

e LE

SS-R

T or

iLES

S.

Val

idat

ed a

gain

st 3

D m

otio

n an

alys

is a

nd la

bora

tory

m

easu

res

of p

KA

M.[8

8]

Repo

rted

to p

redi

ct

adol

esce

nt fe

mal

e at

hlet

es

with

‘hig

h’ p

KA

Ms

(> 2

1.74

N

m) w

ith h

igh

sens

itivi

ty

(77%

) and

spe

cific

ity

(71%

).[90,

93]

Sens

itivi

ty o

f thr

ee ra

ters

us

ing

the

OSD

KV

met

hod

in

iden

tifyi

ng ‘t

ruly

hig

h-ri

sk’

fem

ale

indi

vidu

als b

ased

on

a 3D

cut

-off

of 1

0.83

o of k

nee

valg

us w

as d

eem

ed

‘inad

equa

te’ (

67-8

7%);

whi

le

spec

ifici

ty v

alue

s fo

r id

entif

ying

thes

e in

divi

dual

s w

ere

cons

ider

ed ‘a

dequ

ate’

(6

0-72

%).[8

5]

Fron

tal p

lane

kne

e m

otio

n re

cord

ed v

ia th

e 2D

-Cam

m

etho

d w

as c

onsi

sten

t with

3D

dat

a fo

r the

sid

e-st

ep

(RM

SE =

1.7

o ) a

nd s

ide-

jum

p (R

MSE

= 1

.5o )

task

s, bu

t not

th

e sh

uttle

run

task

(RM

SE =

16

.0o )

in m

ale

and

fem

ale

colle

giat

e ba

sket

ball

athl

etes

.[92]

No

valid

ity a

sses

smen

t was

id

entif

ied

for t

he s

cree

ning

m

etho

d.

236


Tabl

e 7.

3 Ex

trac

ted

data

for t

he id

entif

ied

fiel

d-ba

sed

scre

enin

g m

etho

ds (c

ontin

ued)

.

Cri

teri

a LE

SS /

LESS

-RT

/ iL

ESS

Clin

ic-B

ased

A

lgor

ithm

O

bser

vatio

nal S

cree

ning

of

Dyn

amic

Kne

e V

algu

s 2D

-Cam

M

etho

d Tu

ck Ju

mp

Ass

essm

ent

Relia

bilit

y as

sess

men

t un

dert

aken

Repo

rted

as

havi

ng

‘exc

elle

nt’ (

ICC

= 0

.91;

SEM

=

0.42

) int

ra- a

nd ‘g

ood’

(IC

C =

0.

84; S

EM =

0.7

1) in

ter-

rate

r re

liabi

lity;

[91]

and

‘exc

elle

nt’

intr

a- (I

CC

= 0

.97)

and

inte

r-

(ICC

= 0

.92)

rate

r rel

iabi

lity

for m

ale

and

fem

ale

athl

etes

.[105

] Int

er-r

ater

av

erag

e pe

rcen

tage

ag

reem

ent b

etw

een

‘nov

ice’

(o

ne h

our o

f tra

inin

g) a

nd

‘exp

ert’

(five

yea

rs’

expe

rien

ce u

sing

the

LESS

) ra

ters

rang

ed fr

om 6

5-10

0%

acro

ss s

cori

ng c

rite

ria

in

fem

ale

colle

giat

e at

hlet

es.[9

4]

‘Goo

d’ in

ter-

rate

r rel

iabi

lity

(ICC

= 0

.72-

0.81

; SEM

= 0

.69-

0.76

) was

repo

rted

for t

he

LESS

-RT

for h

ealth

y m

ales

an

d fe

mal

es.[4

10]

‘Exc

elle

nt’ i

ntra

-rat

er

relia

bilit

y re

port

ed fo

r rat

ing

of k

nee

valg

us (I

CC

= 0

.997

) an

d kn

ee fl

exio

n ra

nge

of

mot

ion

ICC

= 0

.952

) dur

ing

the

jum

p-la

ndin

g ta

sk

perf

orm

ed b

y ad

oles

cent

fe

mal

es.[4

12] N

o da

ta

pert

aini

ng to

intr

a-ra

ter

relia

bilit

y ac

ross

oth

er

vari

able

s or

inte

r-ra

ter

relia

bilit

y w

as id

entif

ied.

Intr

a- (κ

= 0

.75-

0.85

) and

in

ter-

rate

r (κ

= 0.

77-0

.80)

ag

reem

ent b

etw

een

thre

e ra

ters

for s

cori

ng fe

mal

e in

divi

dual

s w

as re

port

ed a

s ‘s

ubst

antia

l.’[8

5]

No

asse

ssm

ent o

f int

ra- o

r in

ter-

rate

r rel

iabi

lity

of th

e sc

reen

ing

met

hod

was

id

entif

ied.

Intr

a-ra

ter r

elia

bilit

y w

as

initi

ally

repo

rted

as ‘

high

’.[86]

A

vera

ge p

erce

ntag

e in

tra-

an

d in

ter-

rate

r agr

eem

ent

was

foun

d to

rang

e fr

om 7

5-10

0% a

cros

s sc

orin

g cr

iteri

a fo

r mal

e an

d fe

mal

e at

hlet

es.[2

56] I

n co

ntra

st, i

ntra

-ra

ter r

elia

bilit

y be

twee

n th

ree

rate

rs fo

r ass

essi

ng

mal

es a

nd fe

mal

es h

as b

een

show

n to

rang

e fr

om ‘p

oor’

to

‘mod

erat

e’ (I

CC

= 0

.44-

0.72

); an

d in

ter-

rate

r re

liabi

lity

has b

een

repo

rted

as

‘poo

r’ (I

CC

= 0

.47)

but

did

im

prov

e in

a s

econ

d sc

orin

g se

ssio

n (I

CC

= 0

.69)

.[414

]

237


Tabl

e 7.

3 Ex

trac

ted

data

for t

he id

entif

ied

fiel

d-ba

sed

scre

enin

g m

etho

ds (c

ontin

ued)

.

Cri

teri

a LE

SS /

LESS

-RT

/ iL

ESS

Clin

ic-B

ased

A

lgor

ithm

O

bser

vatio

nal S

cree

ning

of

Dyn

amic

Kne

e V

algu

s 2D

-Cam

M

etho

d Tu

ck Ju

mp

Ass

essm

ent

‘At-r

isk’

cut

-of

f val

ue

A L

ESS

scor

e of

five

or

abov

e ha

s be

en

reco

mm

ende

d as

a c

ut-o

ff va

lue

for i

dent

ifyin

g in

divi

dual

’s a

t-ris

k of

AC

L in

jury

.[104

] No

spec

ific

cut-o

ff va

lues

wer

e id

entif

ied

for

the

LESS

-RT.

No

spec

ific

cut-

off v

alue

s fo

r at

-ris

k at

hlet

es’ w

as

iden

tifie

d fo

r the

Clin

ic-

Base

d A

lgor

ithm

.

Ath

lete

s’ w

ere

deem

ed

‘hig

h-ri

sk’ i

f the

y di

spla

yed

a kn

ee p

ositi

on m

edia

l-to-

the-

toe

duri

ng la

ndin

g.[8

5]

No

spec

ific

cut-

off v

alue

s fo

r at

-ris

k at

hlet

es’ w

as

iden

tifie

d fo

r the

2D

-Cam

M

etho

d.

Ath

lete

s’ e

xhib

iting

six

or

mor

e fla

wed

tech

niqu

es (i

.e.

a sc

ore

of s

ix o

r abo

ve)

duri

ng th

e Tu

ck Ju

mp

Ass

essm

ent s

houl

d be

ta

rget

ed fo

r fur

ther

tr

aini

ng.[8

6]

Add

ition

al

equi

pmen

t re

quir

ed

Tw

o di

gita

l vid

eo

cam

eras

a 3

0 cm

jum

p-bo

x A

sses

smen

t she

et[9

1, 9

4]

31

cm ju

mp-

box

Tap

e m

easu

re

Sca

les

Isok

inet

ic te

stin

g de

vice

b T

wo

digi

tal v

ideo

cam

eras

V

ideo

ana

lysi

s so

ftwar

e N

omog

ram

she

et[9

0]

31

cm ju

mp-

box

Dig

ital v

ideo

cam

era[8

5]

One

dig

ital v

ideo

cam

era

Vid

eo a

naly

sis

softw

are[9

2]

Ass

essm

ent s

heet

D

igita

l cam

erab

Tw

o di

gita

l vid

eo

cam

eras

b[86

, 415

]

Add

ition

al

trai

ning

re

quir

ed

Acc

urat

e sc

orin

g ha

s be

en

repo

rted

by

‘nov

ice’

rate

rs

(one

hou

r of t

rain

ing)

co

mpa

red

to ‘e

xper

t’ ra

ters

(fi

ve y

ears

’ exp

erie

nce)

.[94]

N

o fu

rthe

r ind

icat

ions

on

trai

ning

requ

irem

ents

wer

e id

entif

ied.

No

indi

catio

ns o

n tr

aini

ng

requ

irem

ents

wer

e id

entif

ied.

20-m

inut

e tr

aini

ng v

ideo

s w

ere

supp

lied

to ra

ters

; and

te

n m

inut

es w

as a

lloca

ted

to

revi

ew in

stru

ctio

ns a

nd

exam

ine

prac

tice

exam

ples

pr

ior t

o in

itial

ratin

g se

ssio

ns.[8

5]

No

indi

catio

ns o

n tr

aini

ng

requ

irem

ents

wer

e id

entif

ied.

The

requ

irem

ent

of p

revi

ous

expe

rien

ce in

di

gitiz

ing

land

mar

ks is

ex

pres

sed

by th

e de

velo

pers

.[92]

No

indi

catio

ns o

n tr

aini

ng

requ

irem

ents

wer

e id

entif

ied.

238


Tabl

e 7.

3 Ex

trac

ted

data

for t

he id

entif

ied

fiel

d-ba

sed

scre

enin

g m

etho

ds (c

ontin

ued)

.

Cri

teri

a LE

SS /

LESS

-RT

/ iL

ESS

Clin

ic-B

ased

A

lgor

ithm

O

bser

vatio

nal S

cree

ning

of

Dyn

amic

Kne

e V

algu

s 2D

-Cam

M

etho

d Tu

ck Ju

mp

Ass

essm

ent

Tim

e ta

ken

to s

cree

n in

divi

dual

at

hlet

e

Tota

l tes

ting

time

incl

udin

g se

t-up

per i

ndiv

idua

l was

re

port

ed a

s ty

pica

lly fi

ve

min

utes

or l

ess.

[91]

Tra

ined

ra

ters

can

sco

re th

ree

tria

ls

in th

ree

to fo

ur m

inut

es.[9

1]

The

LESS

-RT

take

s ap

prox

imat

ely

two

min

utes

pe

r ind

ivid

ual.[4

10] N

o in

dica

tion

of th

e tim

e ta

ken

to s

cree

n us

ing

the

iLES

S w

as id

entif

ied.

No

indi

catio

ns o

n th

e tim

e ta

ken

to s

cree

n w

ere

iden

tifie

d.

No

indi

catio

ns o

n th

e tim

e ta

ken

to c

ompl

ete

the

jum

p -la

ndin

g ta

sk w

ere

iden

tifie

d.

Rate

rs w

ere

give

n on

e vi

ewin

g an

d al

loca

ted

fifte

en

seco

nds

to s

core

eac

h in

divi

dual

.[85]

No

indi

catio

ns o

n th

e tim

e ta

ken

to s

cree

n w

ere

iden

tifie

d.

Test

dur

atio

n w

as re

port

ed

as te

n se

cond

s.[8

6] N

o fu

rthe

r in

form

atio

n on

scr

eeni

ng

time

was

iden

tifie

d.

App

licab

ility

ac

ross

spo

rts

and

athl

etes

No

indi

catio

ns o

r res

tric

tions

ar

e gi

ven

as to

spo

rts

and

athl

etes

the

met

hod

can

be

appl

ied

to.

The

met

hod

is d

escr

ibed

as

bein

g ap

plic

able

to ‘f

emal

e at

hlet

es.’[8

8, 9

0, 9

3]

No

indi

catio

ns o

r res

tric

tions

ar

e gi

ven

as to

spo

rts

and

athl

etes

the

met

hod

can

be

appl

ied

to.

No

indi

catio

ns o

r res

tric

tions

ar

e gi

ven

as to

spo

rts

and

athl

etes

the

met

hod

can

be

appl

ied

to.

No

indi

catio

ns o

r res

tric

tions

ar

e gi

ven

as to

spo

rts

and

athl

etes

the

met

hod

can

be

appl

ied

to.

LESS

– L

andi

ng E

rror

Sco

ring

Sys

tem

; LES

S-RT

– L

andi

ng E

rror

Sco

ring

Sys

tem

Rea

l-Tim

e; i

LESS

– i

-Lan

ding

Err

or S

cori

ng S

yste

m; A

CL

– an

teri

or c

ruci

ate

ligam

ent;

PPV

– p

ositi

ve p

redi

ctiv

e va

lue;

NPV

– n

egat

ive

pred

ictiv

e va

lue;

OSD

KV

– O

bser

vatio

nal S

cree

ning

of D

ynam

ic K

nee

Val

gus;

3D

– th

ree-

dim

ensi

onal

; ATS

– a

nter

ior t

ibia

l sh

ear;

pKA

M –

pea

k kn

ee a

bduc

tion

mom

ent;

RMSE

– ro

ot m

ean

squa

re e

rror

; IC

C –

intr

a-cl

ass

corr

elat

ion

coef

ficie

nt; S

EM –

sta

ndar

d er

ror o

f the

mea

n a

– no

t req

uire

d fo

r LES

S-RT

b

– sc

reen

ing

can

be c

ompl

eted

with

out


239

7.4 Discussion

The ability to identify athletes at-risk for future ACL injury is of great

importance across all levels of sport. Screening methods that can be completed

in a field setting provide a valuable option for those in the wider community

without access to laboratory-based equipment. The purpose of this review was

to evaluate and compare current field-based screening methods for ACL

injury risk to determine their applicability of use in wider community settings.

Field-based screening methods had limited predictive validity in identifying

athletes at-risk of ACL injury. Differences between screening methods

surrounding the equipment required, time taken to screen athletes, and

applicability across varying sports and athletes were identified. The data

presented within this review can be used to determine which screening

method may be the best option or more applicable for use in field or clinical

settings, or with specific athlete populations

7.4.1 Appropriateness for Intended Use

7.4.1.1 Alignment with Constructs of Interest

An important consideration when evaluating a screening method is that the

constructs or criteria assessed align with the constructs of interest.[394] In the

context of screening for ACL injury risk, effective screening methods will

target movement strategies or techniques that contribute to increases in ACL

load or injury risk. All identified versions of the LESS included a number of

screening criteria related to movement strategies or techniques that have been

proposed to increase ACL injury risk. These scoring criteria include reduced

sagittal plane joint motion (e.g. reduced trunk, hip and knee flexion); and

increases in knee abduction, lateral trunk motion and internal/external


240

rotation of the shank; all of which have been linked to increases in ACL

loading or injury risk.[20-23, 25, 160, 372] Additional criteria, such as altered stance

width and symmetry of foot contact, while less clearly linked to increases in

ACL loading or injury risk may still provide relevant information pertaining

to an individuals’ injury risk. A wider stance has been shown to increase knee

valgus moments during side-step cutting manoeuvres,[167, 304] and may be

applicable in identifying ACL injury risk. However, this effect has only been

identified in unilateral cutting manoeuvres and may not be present in the

bilateral drop vertical jump task used within the LESS. An asymmetrical foot

contact is proposed to place a greater load on the initial landing limb and may

be indicative of an out-of control or off-balance landing,[410] a factor associated

with ACL injury occurrences.[99] Importantly, the LESS screening criteria

attempts to acknowledge the multiplanar mechanism of ACL injury[33] by

including criteria pertaining to sagittal, frontal and transverse plane

biomechanics.[91]

The Clinic-Based Algorithm predicts the probability for high peak knee

abduction moments during landing as an indication of injury risk.[87, 90] The

individual measures used to calculate this appear well supported through the

literature.[89, 90, 149, 160, 304, 417-419] High peak knee abduction moments during

landing have been prospectively identified as a predictor of ACL injuries,[160]

therefore field-based measures that can estimate this are relevant to

identifying ACL injury risk. Tibial length, knee valgus and knee flexion range

of motion during landing, body mass, and QuadHam ratio make up the values

used to calculate the probability of high knee loading.[87, 88, 90] In isolation, these

measures have been proposed as mechanisms for ACL injury, or linked to

increases in knee abduction moments or ACL loads during landing or cutting

movements. [90, 149, 160, 304, 417-419] Importantly, the combination of these measures


241

within the algorithm has been shown to predict knee abduction moments

during landing with high sensitivity and specificity.[90, 93] It must be noted that

the Clinic-Based Algorithm was established to identify the increase in peak

knee abduction moments associated with pubertal female development.[90]

Particular maladaptations that occur during puberty, such as deficits in

strength and power relative to growth and development are more evident in

females than males.[418, 420] The application of the Clinic-Based Algorithm to a

male cohort is, therefore, not recommended.[88]

Both the OSDKV and 2D-Cam methods focus on a single factor in measuring

injury risk, namely ‘dynamic valgus’ or frontal plane knee motion.[85, 92] This is

likely due to the notion that ‘dynamic valgus’ positions have been shown to

significantly increase ACL loads,[20-23] are often associated with ACL injury

occurrences,[98, 99, 159] and have been identified as a significant predictor of ACL

injuries.[160] While ‘dynamic valgus’ is an important contributor to ACL injury

risk,[160] this is only one factor in a complex problem. Evidence suggests ACL

injuries occur via a multiplanar mechanism,[33] and motion from more

proximal segments (i.e. the trunk and hip) are important when considering

ACL injury risk.[166, 167, 372, 421] The use of two-dimensional methods in

quantifying ‘dynamic valgus’ should also be interpreted with caution.

Dynamic valgus can stem from a combination of three-dimensional

movements at the trunk, hip, knee and ankle.[160, 162, 372] Increased frontal plane

knee motion or a knee position medial-to-the-toe can result from hip

adduction and internal rotation, and do not necessarily indicate greater valgus

positions.[422, 423] The use of two-dimensional measures focused solely on the

knee may be inadequate in evaluating ‘dynamic valgus.’

The scoring criteria used within the Tuck Jump Assessment are proposed to

identify landing technique deficiencies relating to ligament dominance,


242

quadriceps dominance, leg dominance or residual injury deficits, trunk

dominance, and technique perfection.[413] Deficiencies in these areas are

indicative of neuromuscular imbalances that can lead to increases in knee joint

loads and subsequently increase ACL injury risk.[424] Ligament dominance

manifests in an inability to control dynamic knee valgus when landing and

cutting, and is therefore a factor in ACL injury risk.[424] The high effort required

in the tuck jump movement may reveal technique deficiencies and the

magnitude of frontal plane knee motion may be useful in identifying ligament

dominant athletes.[424] There appears to be a disconnect between the remaining

scoring criteria and the technique deficiencies they target. No empirical

support was found for the use of the proposed scoring criteria in identifying

quadriceps dominance (i.e. excessive landing contact noise), leg dominance or

residual injury deficits (i.e. unequal thighs side-to-side; foot placement not

parallel; and unequal foot contact timing), or trunk dominance (i.e. thighs

reaching parallel; pauses between jumps; and does not land in the same

footprint) in athletes. Further, no suggested link between maintaining perfect

tuck jump technique across the ten seconds and ACL injury risk was

identified. Validation linking tuck jump technique deficiencies to ACL injury

risk factors is required before this method can be recommended for ACL injury

risk screening purposes.

7.4.1.2 Theoretical and Empirical Support

Another important consideration when determining the efficacy of screening

methods is whether the format and content of the screening method have been

validated.[394] A number of the identified screening methods had been

evaluated against 3D laboratory-based measures to validate their use. Poor

LESS scores (i.e. indication of increased injury risk) were associated with

reduced hip and knee flexion; greater knee abduction and hip adduction


243

angles; greater hip and knee internal rotation, hip extension and adduction,

and knee abduction moments; and greater anterior tibial shear forces during

the performance of a bilateral drop vertical jump task.[91] These results support

the use of the LESS as an ACL injury risk screening method, as all of these

biomechanical variables have been proposed as risk factors for ACL

injury.[20-22, 26, 29, 121, 149, 160, 162, 167, 191, 425]

Both the OSDKV and 2D-Cam Method were assessed against 3D measures of

knee valgus. The validity of the OSDKV method was determined by

comparing the allocation of individuals to high- and low-risk groups between

3D motion analysis and three raters using the screening method.[85] Compared

to 3D analysis, raters were able to allocate individuals to these groups with

sensitivity and specificity values of 67-87% and 60-72%, respectively.[85] While

specificity values were considered adequate for screening purposes, the

sensitivity of raters was considered inadequate, with up to a third of truly

high-risk individuals being missed.[85] A knee position medial-to-the-toe is not

always indicative of 3D knee abduction[422] and this may be why some truly

high-risk individuals were missed. The 2D-Cam Method showed good

agreement with laboratory-based measures of frontal plane knee motion

during the side-step and side-jump, but not the shuttle run task.[92] The

developers suggested this may be due to the positioning of the athlete relative

to the camera used to digitise joint centres.[92] During both the side-step and

side-jump tasks, the frontal plane of the athlete’s body remains relatively

parallel to the camera during the stance phase of the movement; with this not

being the case during the shuttle run task.[92] This highlights a limitation of the

2D-Cam Method, with this method likely only applicable to athletic tasks

where the frontal plane remains in full view of the camera through the analysis

phase of the motion.


244

While certain screening methods showed good agreement with laboratory-

based measures, these studies tended to utilise generic athletic tasks

performed bilaterally (e.g. drop vertical jump landings) to validate the

method.[88, 90, 91, 93, 94] This potentially stems from the majority of the screening

methods using generic bilateral jump-landing tasks (e.g. drop vertical jump-

landings, repeat tuck jumps) in assessing ACL injury risk.[85-87, 91, 94, 409] The

specificity of these movements to ACL injury risk could be considered

questionable, with previous research showing these generic tasks do not

adequately represent those seen in a competitive sporting environment,[95, 96]

and noncontact ACL injuries are more common during unilateral landing or

plant-and-cut manoeuvres.[97-99] In addition, injuries occur more frequently in

competitive match environments compared to in training.[1] Considering these

factors, it is important that screening methods for ACL injury risk can identify

athletes who perform unilateral sporting manoeuvres with high-risk lower

limb biomechanics. The incorporation of more ecologically valid or unilateral

movement tasks within injury risk screening protocols may achieve this. For

example, unplanned side-stepping more closely emulates a competitive

environment and the potential ACL injury mechanism by limiting decision

time and increasing knee loading.[247] Screening of this movement task could

yield more valuable information pertaining to an individuals’ level of ACL

injury risk. However, a number of factors pertaining to technique can

influence important biomechanical measures of injury risk during more sport-

specific movements.[166, 167, 304] The need to examine these additional factors may

increase the complexity of screening, increasing the analysis and time burdens

involved. Exposing athletes’ to a potentially high-risk task for the purpose of

injury risk screening may also be undesirable for coaches. Simplistic

movement tasks that better replicate the biomechanics of sport-specific

movements may be the most applicable for injury risk screening. It is yet to be


245

determined whether any of the screening methods examined within this

review are valid in identifying athletes who perform more sport-specific

movements in a way that may influence their level of injury risk. Future

examination against laboratory-based measures acquired from sport-specific

movements known to cause ACL injury may aid in identifying whether certain

screening methods are more appropriate for specific sports or sporting

movements.

7.4.1.3 Population Fit

It is helpful to examine whether a screening method is designed for the

population of interest.[394] Screening methods for ACL injury risk should be

most applicable to athletes participating in sports which involve a high

frequency of cutting and landing tasks, as these sports tend to incur the highest

rates of injury.[1] For the majority of screening methods examined, very few

restrictions on the sports or athletes to which they could be applied to were

identified. Both the LESS and Tuck Jump Assessment were accompanied by

general descriptions, such as; an “assessment tool for detecting poor jump-

landing mechanics,”[91] and an assessment for the “identification of lower

extremity landing technique flaws,”[86] respectively. This suggests these

methods may be applicable for screening athletes in sports where any form of

high risk jump-landing movements are present. There is evidence to support

the use of the LESS in a wide range of athlete populations, with the method

being utilised in both male and female military cadets,[91, 426] and collegiate[94]

and youth athletes[427] of varying sports. The LESS may be more effective in a

youth athlete population, where it appears to have some predictive value in

identifying future ACL injuries.[104] The same was not found for an older group

of high school and collegiate athletes,[105] suggesting this screening method

may not be as effective across different populations. While the Tuck Jump


246

Assessment has been used to examine both male and female athletes,[256] the

method has not been as widely applied to a range of populations. It is therefore

difficult to assess whether this method is more or less effective across a range

of populations. The 2D-Cam method had only been applied to a cohort of

collegiate male and female basketball athletes, and was only successful in

measuring the side-step and side-jump tasks.[92] While these are general

athletic tasks, caution may be required if applying this method to athlete

cohorts or sports where athletic manoeuvres differ from those examined.

Two of the screening methods identified, the OSDKV and Clinic-Based

Algorithm, had only been used to examine female athletes.[85, 87, 88, 90] This is

likely due to the sex bias associated with ACL injuries, and thus the

subsequent focus on female athletes in research examining ACL injury risk.

While the Clinic-Based Algorithm had only been applied to female athletes, it

is important to highlight that the method was specifically designed for and is

only applicable to this sex.[87, 88, 90] The method has been utilised to examine

athletes across multiple sports,[88-90, 412] however, the focus has tended to be on

a youth athlete population.[88-90] Considering this method was developed and

validated using a cohort of female adolescent athletes (aged 14.0 ± 2.5 years),[88]

it appears application of the Clinic-Based Algorithm outside of this population

may be unwarranted. The evidence suggests different screening methods may

be more or less applicable to different populations. Consideration must

therefore be given to the specific population being examined to determine

which screening method may be the most applicable in a given situation.


247

7.4.2 Technical Adequacy

7.4.2.1 Test-Retest (Intra-Rater) and Inter-Rater Reliability

The test-retest (intra-rater) and inter-rater reliability of screening methods are

important to ensure accurate screening can be undertaken over time and by

different assessors.[394] The majority of screening methods examined reported

acceptable levels of intra- and inter-rater reliability.[85, 86, 91, 94, 105, 256, 410, 412] An

exception to this were the reliability values reported by Dudley et al.[414] for the

Tuck Jump Assessment method, with intra- and inter-rater reliability ranging

from ‘poor’ to ‘moderate.’ Reliability values did, however, improve upon a

second scoring session suggesting there may be a learning effect with the Tuck

Jump Assessment screening method. Additional information or training may

also be required for accurate Tuck Jump Assessment scores to be assessed. This

could be a reason why the reliability values reported by Herrington et al.[256]

were much higher, as the testers within this study may have had greater

experience or knowledge in using the screening method. Both intra- and inter-

rater reliability are important features for injury risk screening methods.

Screening methods can be used to monitor changes in injury risk over time,

therefore consistency in interpretation and scoring is required to ensure

variations represent true differences and not errors in application. This

becomes especially pertinent if various raters are performing subsequent

analyses on an athlete, where variation between raters may result in erroneous

interpretation of injury risk. The distinct and specific scoring criteria used

within the majority of the screening methods examined[85, 86, 91, 409, 410] is likely to

play a role in the acceptable levels of reliability reported. Giving users specific

instructions and/or scoring criteria leaves less room for subjective

interpretation, thus creating less error between subsequent analyses or

different raters. Screening methods which incorporate clear and easy to


248

understand scoring/rating criteria will be more applicable for use in wider

community settings.

7.4.2.2 Predictive Validity

The most important indicator of a screening methods technical adequacy is the

ability to accurately distinguish between those who will and will not sustain

an injury.[394] Of the screening methods included in this review, only the Clinic-

Based Algorithm and LESS had been prospectively evaluated against future

ACL injury occurrences.[104, 105, 412] Goetschius et al.[412] applied the Clinic-Based

Algorithm to a population of female high school and college athletes, and

determined there was no relationship between the risk of suffering an ACL

injury and the predicted peak knee abduction moment obtained from the

screening method. However, the developers of the screening method

expressed concerns regarding this study, suggesting factors such as the jump-

landing protocol utilised, the lack of use of shoes during testing, and the

absence of control over foot/ankle separation during the jump-landing task

may invalidate these results.[411] Another factor to consider may be the

population examined by Goetschius et al.[412]. As the Clinic-Based Algorithm

was developed and validated against a youth female athlete population (aged

14.0 ± 2.5 years),[88] the inclusion of older female athletes (high school and

collegiate athletes; age range 15-22) may have influenced its predictive ability.

While the Clinic-Based Algorithm was unsuccessful in predicting future ACL

injuries, it has been shown to predict high peak knee abduction moments (>

21.74 Nm) with high sensitivity (77%) and specificity (71%) in female athletes

during a bilateral drop vertical jump task.[87, 93] As previously mentioned, a

higher peak knee abduction moment is a significant predictor of future ACL

injury.[160] The ability to distinguish between female athletes demonstrating

‘high’ and ‘low’ peak knee abduction moment during landing via a field-based


249

screening method may be useful in identifying athletes at-risk of sustaining an

ACL injury. It is, however, important to note that the Clinic-Based Algorithm

has only been shown to predict high peak knee abduction moments in the

same cohort from which the algorithm was developed. Validation against

multiple relevant cohorts is an important step in establishing the efficacy of

screening methods.[428] Further validation of this method may be required

before more widespread use can be recommended.

Results from the LESS have also been shown to have no relation to future ACL

injury status in a population of female high school and collegiate athletes.[105]

In contrast, the LESS may have some capacity to identify athletes at-risk for

future ACL injury in a youth soccer population.[104] Padua et al.[104] found a

LESS score of five or above identified athletes who went on to incur an ACL

injury with sensitivity and specificity values of 86% and 64%, respectively. The

positive and negative predictive values for the LESS in identifying ACL

injuries were low (1.4%) and high (99.8%), respectively.[104] The positive

predictive value (PPV) of any ACL screening method is likely to be low due to

rate of ACL injuries being quite low, even among high-risk populations.[104]

This low PPV would result in a number of low-risk individuals labelled as

high-risk. Considering the consequences associated with ACL injuries and that

low-risk individuals may also benefit from neuromuscular training, the

benefits of identifying all truly high-risk individuals likely outweighs the costs

associated with misclassification of those who are at a low-risk of ACL injury.

The conflicting results with regard to the predictive validity of the LESS when

applied to different age groups and sporting populations is further evidence

that different screening methods may be more or less applicable to different

populations or sports. Examination of screening methods across a range of

populations of different ages and sports appears necessary.


250

7.4.3 Usability

7.4.3.1 Balance of Costs and Benefits

The main costs associated with employing screening methods in wider

community settings pertains to the equipment and time requirements. An

important aspect in ensuring screening methods are applicable for wider

community use is the requirement for minimal equipment. The need for

minimal equipment is an important element for coach and athlete compliance

to ACL injury prevention programs,[392] and is likely a key factor in promotion

and compliance of injury risk screening as well. Nearly all of the screening

methods identified required minimal additional equipment to complete. The

need for a single or two digital video cameras in combination with a jump-

landing box was a requirement of the LESS/iLESS[91, 409] and OSDKV[85]

methods; while the 2D-Cam Method required a single digital video camera

and video analysis software to analyse results.[92] While not required, the use

of a digital camera or digital video cameras to capture the athlete’s jump-

landing technique is suggested and have been used for the Tuck Jump

Assessment.[86, 256, 413] The use of cameras to capture an athlete’s jump-landing

technique during screening may improve the accuracy and consistency of the

screening method, as it allows the assessor to view the technique multiple

times and pause or slow down the movement. However, while digital cameras

have become more affordable in recent years, the need for this piece of

equipment may still present access issues, particularly within youth or

community level sport. Considering the growth in mobile technologies (i.e.

smart phones, tablets) and digital applications as a means to promote injury

prevention strategies,[429, 430] the development or incorporation of screening

methods with these mobile technologies may provide an opportunity to

extend the scope of injury risk screening. Two of the screening methods


251

identified, the real-time version of the LESS (LESS-RT)[410] and Tuck Jump

Assessment[86] combat the need for digital cameras by the assessor performing

the screening in real-time. While this may increase difficulty; these real-time

screening methods may also be a valid option in settings where access to

equipment, in particular to digital video cameras, is limited.

In comparison to the other methods identified, the Clinic-Based Algorithm has

a much greater reliance on additional equipment. This is not surprising, as the

Clinic-Based Algorithm employs a greater number of measures to determine

an athlete’s level of injury risk. While this may enhance the ability of the

method to identify at-risk athletes, it may be impractical for users in certain

settings. One aspect that may limit the use of this screening method is the need

for an isokinetic device to assess quadriceps-to-hamstring strength

(QuadHam) ratio.[87, 88, 90] It was noted that where an isokinetic device is

unavailable, an equation is provided incorporating the athlete’s mass to attain

a surrogate value for the QuadHam ratio.[87, 89] However, it is suggested that

the prediction algorithm is best utilised with QuadHam ratios achieved via

direct measurement.[87] Whether the surrogate value is used or not, the extra

equipment and number of measures recorded may make the Clinic-Based

Algorithm a less valid option in settings where minimal equipment or time is

available for athlete screening.

Another important feature of field-based screening methods is the time

required to screen athletes. In many cases, athlete screening is undertaken

during group sessions with a number of athletes present. Screening methods

that can be completed quickly and implemented into large scale team

screening sessions may be more beneficial. The LESS was the only screening

method examined which reported a specific time frame in which athlete

screening may be completed. Testing time (including set-up) for the originally


252

developed LESS, utilising video footage for scoring, was reported as taking

five minutes or less per subject.[91] A trained rater was reported as being able

to score an athlete’s trials in approximately three to four minutes.[91] With this

added to the testing time, the total time to screen an athlete using the LESS

appears to be approximately ten minutes. The LESS-RT requires less time to

complete, with the real-time method reportedly taking approximately two

minutes per athlete.[410] No indication was given for the time taken using the

iLESS format. Considering this variation was developed to simplify and

shorten the screening process and only requires the evaluation of one jump-

landing trial,[409] it could be assumed that this method would take less time to

complete than the other LESS variations.

The developers of the 2D-Cam method acknowledge there are relatively

lengthy processing requirements inherent in this approach,[92] suggesting this

method may be impractical where time constraints are present. With regard to

the OSDKV method, raters were given a single viewing of the athlete’s jump-

landing technique, followed by 15-seconds to give their rating.[85] While this is

relatively fast, it does not take into account the more time consuming process

of filming athletes performing the jump-landing task. Considering the ratings

can be given at a later stage, the two processes could be blocked and performed

separately (i.e. viewing footage and rating a group of athletes together at a

later time). Thus, this method could be practical for use when screening a large

group of athletes. The Clinic-Based Algorithm is suggested to simplify the

time associated with laboratory analysis for determining peak knee abduction

moment.[87, 89] However, considering the additional number of measures

required, it is likely this method would take longer to complete than the other

screening methods examined in this review. The Clinic-Based Algorithm may

be more practical in a setting where little to no time constraints are placed on


253

injury risk screening, or where individual screening is being undertaken.

Lastly, the simplistic manner in which the Tuck Jump Assessment can be

undertaken and scored in real-time suggests it could be completed in a short

time frame. Considering athletes can be observed and scored while

performing repeat tuck jumps for ten seconds,[86] it is quite possible this

method could be completed in one or two minutes per athlete. An increase in

the time taken to complete the Tuck Jump Assessment would be seen if digital

cameras were used to screen the athlete’s jump-landing technique. As there is

considerable variation in the time required for each of the screening methods,

this may be an important factor in determining the most appropriate method

to employ in a given setting.

7.4.3.2 Feasibility of Administration

In a wider community context, screening for ACL injury risk is likely to be

undertaken by individuals with a broad range of qualifications. Instructions,

examples, or the training accompanying the screening method should

therefore be clear for the user to ensure accurate application.[394] Many of the

screening methods identified presented no specific guidelines for the type or

amount of training required, with some simply claiming the method was easy

to employ. No specific training guidelines were identified for the 2D-Cam

Method, however it was suggested that previous experience in digitising

landmarks may be necessary,[92] potentially rendering the method impractical

in certain settings. No guidelines for training were also identified for the Tuck

Jump Assessment. This method, however, did claim to be ‘clinician/coach

friendly’[86, 415] implying limited previous knowledge or training is required to

complete the screening protocol. While for the most part, the scoring criteria

relating to the Tuck Jump Assessment is simplistic, certain terminology (such

as ‘lower extremity valgus’) may be difficult to interpret by those


254

inexperienced in screening jump-landing technique. The visual aid provided

with this screening tool, containing images of mistakes to look for in jump-

landing technique[86, 413] could assist in rectifying any confusion surrounding

terminology. The Clinic-Based Algorithm also provided no guidelines for

additional training, however, many of the measures used within this method

are simplistic and would be familiar to the majority of clinicians (e.g. height,

body mass, tibia length). Others, such as QuadHam ratio, and knee valgus and

flexion motion, may again be difficult for coaches inexperienced in injury risk

screening of athletes to easily understand and complete. It is important to note

that an attempt to employ the Clinic-Based Algorithm by a group outside of

the developers[412] was highlighted as having methodological errors that

would likely have influenced results.[411] If certain aspects of the method are

difficult to perform, and the accuracy of the algorithm relies on all aspects

being undertaken correctly, further instructions or training guidelines may be

beneficial to assist those wishing to use this method for athlete screening.

Indications as to the level of training required to be competent in the LESS and

OSDKV methods were identified. Onate et al.[94] compared the scoring

responses between an ‘expert’ (co-developer of the screening method) and

‘novice’ rater using the LESS. The ‘novice’ was provided with a one-hour

training session, which consisted of reviewing each of the LESS scoring items

and conducting a practice analysis of an athlete with the help of an ‘expert’

rater.[94] This one-hour session was sufficient for producing excellent reliability

between ‘novice’ and ‘expert’ LESS scores.[94] While it is likely not feasible for

an ‘expert’ rater to provide training to everyone who wishes to use the LESS,

the fact that a short period of training can produce reliable scores is

encouraging. If this information could be successfully transferred to a more

widespread communicable medium (such as a web-based training video),


255

there is the potential for uptake and accurate use of the LESS in wider

community settings. The use of video for training purposes is not a foreign

concept for ACL injury risk screening, with this technique employed to train

raters on how to undertake the OSDKV method.[85] Raters were provided with

a 20-minute training video that included background information on ACL

injury risk factors, detailed rating instructions and footage of athletes for

practice analyses.[85] In addition, prior to the rating session, ten-minutes was

spent reviewing the rating instructions and performing additional practice

ratings.[85] While sensitivity targets (> 80%) for identification of ‘high-risk’

athletes were rarely met by the three raters; consistent agreement was seen

between raters, and the specificity of ratings was considered adequate for field

screening purposes (60-72%).[85] It is important to note that the three raters

used in this study were physiotherapists with at least five years of experience,

and a high level of sports and orthopaedic expertise.[85] It may be the case that

additional training or instruction is required when inexperienced users are

performing the injury risk screening. Nonetheless, there is promising evidence

that competency in injury risk screening can be achieved via a short (e.g. one-

hour) training period. Providing this training in an easily accessible and

readily available form may increase the uptake and use of screening methods

via improved understanding and knowledge.

7.4.3.3 Utility of Outcomes

Perhaps the most important factor when determining the usability of a

screening method is whether the outcomes and implications from the method

can be clearly understood and used to guide further intervention.[394]

Dichotomous classification systems (i.e. ‘high-risk’ or ‘low-risk’), such as those

seen within the OSDKV and iLESS methods,[85, 409] could be considered the

simplest outcome with regard to ACL injury risk screening. The simplicity of


256

this classification presents certain benefits, particularly in wider community

settings. First, a simple decision of ‘high-’ or ‘low-risk’ reduces the analysis

burden on those undertaking screening. Second, the implications and

subsequent intervention associated with these outcomes are easily understood

(e.g. all individuals classified as ‘high-risk’ require subsequent intervention).

However, this simplicity may create issues in accurately classifying difficult

cases. With only two categories available, those individuals deemed as

borderline ‘high-risk’ may be incorrectly classified resulting in reduced

sensitivity. This was identified as a potential limitation by the developers of

the OSDKV method, and it may be beneficial to err on the side of caution by

allocating borderline cases to the ‘high-risk’ group.[85] As mentioned earlier,

the benefits of identifying all truly high-risk individuals likely outweigh the

costs associated with misclassification of certain low-risk individuals.

In contrast, the majority of screening methods (i.e. LESS, LESS-RT, Clinic-

Based Algorithm, 2D-Cam Method and Tuck Jump Assessment) used

continuous measures in classifying an individuals’ level of ACL injury risk.[86,

87, 90-92, 410] These included the use of scores (LESS, LESS-RT and Tuck Jump

Assessment),[86, 91, 410] risk probabilities (Clinic-Based Algorithm)[87, 90] and joint

angles (2D-Cam Method).[92] These continuous measures allow individuals to

be classified on a risk continuum, counteracting one of the limitations

associated with a dichotomous classification system. Of those that used these

continuous measures, higher values/scores were suggested to indicate greater

ACL injury risk.[86, 87, 90-92, 410] This is useful in a comparative context, where

athletes’ receiving higher scores can be identified as at greater risk of injury. It

is important that these measures are coupled with recommendations as to

what the results signify or at what stage intervention is required. The LESS

and Tuck Jump Assessment methods were accompanied with such


257

recommendations, where scores of five or above[104] and six or above[86] on the

respective methods are given as points where intervention is required. The

Clinic-Based Algorithm provides a probability percentage of whether the

individual will exhibit a high knee abduction moment during landing,[87, 88, 90]

however there is no indication as to what point on this probability scale

intervention is required. While a percentage is an easy value to understand,

the interpretation of what is classified as ‘high-risk’ and where to intervene

appears to remain up to the user. Similarly, the 2D-Cam Method indicates that

greater frontal plane knee angles are indicative of increased risk,[92] but also

provides no indication as to when this is reaching high-risk levels. The need

for subjective interpretation of results may make these methods less applicable

in certain wider community contexts where limited knowledge about what

constitutes high-risk performance is present.

It is important to acknowledge that only screening methods specific to ACL

injury risk were included within this review. It is possible that other field-

based screening methods developed for lower limb injury or injuries to the

knee in general may also be useful in screening for ACL injury risk. Two

studies[397, 405] that examined screening methods for general lower limb injury

were excluded from this review. Chorba et al.[397] examined whether the

Functional Movement Screen (FMSTM) could be used to predict injuries in a

population of female collegiate athletes. A FMSTM score of 14 or less resulted

in a 4-fold increase in lower limb injury risk.[397] The specific types of lower

limb injuries sustained were not reported, therefore it is difficult to determine

the application of this method to ACL injuries specifically. Brumitt et al.[405]

examined the utility of standing long jump, single-leg hop for distance, and

lower extremity functional tests in identifying collegiate athletes at higher risk

of injury. Female athletes who completed the lower extremity functional test


258

with a higher score were six times more likely to sustain a thigh or knee injury,

and were four times more likely to suffer a foot or ankle injury if they exhibited

greater than a 10% side-to-side asymmetry between single-leg hop

distances.[405] The specifics of knee injuries sustained were also not reported in

this study, again making it difficult to determine the application to ACL injury.

Nonetheless, it appears there is some validity in utilising screening methods

to identify general lower limb or area specific injury risk. Screening for lower

limb injury or injuries to the knee in general may also be a viable and time

efficient option for injury risk screening in wider community settings.

However, given the traumatic nature and consequences associated with ACL

injuries, it remains important to be able to identify the individuals at greater

risk for this specific injury.

7.5 Conclusions

This study found limited evidence supporting the predictive ability of field-

based screening methods in identifying future ACL injuries. The LESS was

predictive of ACL injuries in a youth athlete population,[104] however this result

did not transfer to a high school and collegiate athlete cohort.[105] The Clinic-

Based Algorithm was validated against a biomechanical measure that has been

identified as a significant predictor of ACL injuries (i.e. peak knee abduction

moment). However, this method requires extensive additional equipment and

time to complete and has only been validated in young (14-15 years old)

female athletes. The Clinic-Based Algorithm may be limited in its application

to a range of settings and athlete populations. The variants of the LESS appear

to be the most applicable community wide screening methods as the scoring

criteria target movement strategies associated with increases in ACL load or

injury risk, and minimal equipment and set-up/analysis time is required.


259

While the Tuck Jump Assessment requires minimal equipment and time to

undertake, further research linking deficiencies in tuck jump technique to

movement strategies that increase ACL loading or injury risk is required

before this method can be recommended. Limited studies were available

pertaining to the OSDKV and 2D-Cam Method, thus it is difficult to

recommend the use of these screening methods in evaluating ACL injury risk.

Differences between screening methods surrounding the equipment required,

time taken to screen athletes, and applicability across varying sports and

athletes were identified. These are important factors in determining the most

appropriate screening method to use in various situations. Further research is

required to determine the predictive ability of these screening methods in

identifying injuries, as well as their ability to identify athletes who perform

more sport-specific movements with high-risk lower limb biomechanics. This

may allow for a better identification as to which screening methods may be

more applicable to certain athletes, sports or athletic manoeuvres. The data

presented within this review can be used to determine which screening

method may be the best option or more applicable for use in certain settings,

or with specific athlete populations.

260

Chapter Eight: Efficacy of Field-Based Anterior Cruciate Ligament Injury Risk Screening Methods in Identifying High-Risk Landing Postures during a Sport-Specific Task

Adapted from: Fox AS, Bonacci J, McLean SG, Saunders N. Efficacy of ACL injury

risk screening methods in identifying high-risk landing patterns during a sport-

specific task. Scand J Med Sci Sports. doi: 10.1111/sms.12715.

8.1 Introduction

Laboratory-based measures (e.g. three-dimensional motion analysis, force

plates) are capable of estimating segment motions and joint kinetics during

athletic tasks, providing an option to evaluate movement strategies that may

predispose an athlete to anterior cruciate ligament (ACL) injury. This

approach requires extensive laboratory-based equipment and training to

complete, thus is likely inaccessible for the majority of the wider community.

To combat this, field-based screening methods for identifying ACL injury risk

have emerged.[431] The field-based screening methods examined in the

previous chapter, namely the variants of the Landing Error Scoring System

(LESS);[91, 409, 410] Clinic-Based Algorithm;[87, 90] Observational Screening of

Dynamic Knee Valgus (OSDKV);[85] 2D-Cam Method;[92] and Tuck Jump

Assessment;[86] propose to offer a simplistic method to identify dysfunction in

an individual that may lead to ACL injury. Limited research was available

pertaining to the OSDKV and 2D-Cam methods, while the significant

Screening and High-Risk Landing Postures

261

equipment required for the Clinic-Based Algorithm suggests this method may

not be applicable for use across all settings. In contrast, the LESS and Tuck

Jump Assessment have received significant attention in recent

literature,[86, 91, 94, 104, 256, 410, 413, 414] and have minimal equipment and time costs

associated with their use. These methods appear applicable for the majority of

wider community settings and will therefore be the focus of this chapter.

In order for an ACL injury risk screening method to be effective, it must be

linked to the mechanical aetiology of ACL injuries. Noncontact ACL injuries

are most frequently observed during unilateral sporting manoeuvres, such as

side-step cutting or landing.[97-99, 103, 432] Screening methods that are efficacious

in identifying athletes who employ movement strategies that increase ACL

loading and injury risk during these high-risk sporting tasks are likely to be

the most effective in detecting athletes at-risk of ACL injury. Both the LESS

and Tuck Jump Assessment appear to address key biomechanical components

of ACL injury risk in evaluating jump-landing technique.[86, 91, 94, 413] Despite this,

there is conflicting evidence regarding the predictive validity of ACL injury

risk screening methods in identifying ACL injuries.[431] While no prospective

studies have incorporated the Tuck Jump Assessment method, two

studies[104, 105] have examined the predictive ability of the LESS in identifying

future ACL injuries. LESS scores were capable of identifying elite youth soccer

athletes at-risk of future ACL injury with 86% and 64% sensitivity and

specificity, respectively.[104] In contrast, no relationship between future ACL

injury occurrences and LESS scores was found within a cohort of high school

and collegiate athletes from a range of sports.[105] The populations examined

across these studies differed in age, gender and the sports played, suggesting

these may be important factors in the ability of screening methods to identify

athletes at-risk of ACL injury. Currently, the design and validation of the LESS


262

and Tuck Jump Assessment screening methods have focused on identifying

high-risk movement strategies during generic, non-sport-specific bilateral

athletic tasks that are not typically associated with ACL injury

occurrences.[86, 91, 94] Previous work has shown these generic tasks do not

adequately represent those seen in a competitive sporting environment.[95, 96]

In addition, noncontact ACL injuries occur almost exclusively during

unilateral landing or plant-and-cut manoeuvres.[97, 99, 103] The efficacy of an ACL

injury risk screening method may be dependent on its capacity to identify

those who perform unilateral sporting manoeuvres associated with ACL

injuries with high-risk lower limb biomechanics. Currently, neither the LESS

or Tuck Jump Assessment have been evaluated for their ability to identify

lower limb biomechanics associated with increases in ACL loads or injury risk

during sport-specific manoeuvres known to cause ACL injury. Understanding

the efficacy of these screening methods in identifying athletes who are

potentially at greater risk of ACL injury during the performance of high-risk

sporting tasks will aid in determining the applicability of these methods to

specific sports or sporting movements.

The purpose of this study was to examine the efficacy of two field-based

screening methods for ACL injury risk in identifying lower limb biomechanics

consistent with increased ACL loading or injury risk during a high-risk sport-

specific landing task. Specifically, this study aimed to: (i) examine whether

increases in LESS/Tuck Jump Assessment scores were associated with the

presence of lower limb biomechanics associated with ACL loading and injury

risk; and (ii) examine the applicability of previously recommended LESS/Tuck

Jump Assessment cut-off scores for identifying ‘high-risk’ athletes in

identifying lower limb biomechanics consistent with increased ACL loading

and injury risk. Due to the generic bilateral nature of the screening


263

movements, it was hypothesised that the scores obtained would have no

relationship to the lower limb biomechanics from the sport-specific high-risk

sporting task. Subsequently, the cut-off scores would not be successful in

identifying ‘high-risk’ athletes.

8.2 Methodology




8.2.1 Participants

Thirty-two female netball players (age = 23.2 ± 3.1 years; height = 170.9 ± 7.7

cm; mass = 67.6 ± 8.2 kg; netball experience = 12.47 ± 4.58 years) volunteered

and provided written consent to participate in this study. All participants met

the inclusion/exclusion criteria detailed in Section 3.2 of Chapter Three.


Data from the bilateral drop vertical jump, tuck jump and netball-specific

landing task (i.e. the leap landing) as described in Section 3.3.2, 3.3.3 and 3.3.1,

respectively, performed during testing session one (i.e. baseline) were used

within this study. The order of these tasks were randomised within the testing

session across participants. Participants performed three successful trials of

the bilateral drop vertical jump task as per the LESS protocol (see Figure 8.1,

Page 265),[91] and one successful trial of the repeat tuck jump task as per the

Tuck Jump Assessment protocol (see Figure 8.2, Page 265).[86] The netball leap

landing was used as the high-risk sport-specific task within this study.

Participants performed 20 successful trials of the leap landing on a


264

predetermined ‘landing limb.’ The landing limb was determined by asking

participants which limb they would be more comfortable landing on when

catching a pass during a game. Further to this, players were given an

opportunity to trial landing on each limb to identify which they were more

comfortable landing on. Each landing trial was initiated by breaking and



a pass (see Figure 8.3, Page 265). Trials were repeated until the 20 successful

trials were recorded. Participants were allocated 60-90 seconds rest between

trials in an effort to minimise fatigue.[134, 255] Kinematic and kinetic data

collected from the participants ‘landing limb’ during all landing trials were

used within this study.


265

Figure 8.1 Sagittal view of bilateral drop vertical jump-landing task.

Figure 8.2 Sagittal view of the tuck jump task.



266


For the LESS and Tuck Jump Assessment, video data were collected as per the

procedures outlined in Section 3.4.3 of Chapter Three. Briefly, two commercial

digital video cameras (30 Hz, Mini DV MD225, Canon, Macquarie Park,

Sydney, Australia) placed 3.54 metres from a designated landing area

simultaneously recorded frontal and sagittal views of the bilateral drop

vertical jump and tuck jump tasks. Cameras were repositioned if the entire

body of the participant could not be viewed throughout the task from this

distance. The sagittal plane video camera was always placed on the side that

could view the participant’s predetermined ‘landing limb.’ Scoring for both

LESS and Tuck Jump Assessments were undertaken using ImageJ video

analysis software (US National Institutes of Health, Bethesda, MD, United

States). The same individual performed scoring for all participants.

Procedures and results for determining intra- and inter-rater reliability for

both scoring methods are outlined in a later section of this chapter (see Section

8.2.4.1, Page 272).

To score the LESS, frontal and sagittal video footage of participants’ drop

vertical jump-landing trials were assessed for movement ‘errors’ using the

LESS scoring criteria.[91, 104] The LESS involves 17 score items or ‘errors,’ with

19 being the maximum attainable score (see Table 8.1, Page 268).[91] Scoring

items are dedicated to the assessment of: lower extremity and trunk position

at initial ground contact; foot position at initial contact and the time between

initial contact and maximum knee flexion; lower extremity and trunk

movements between initial contact and maximum knee flexion; overall sagittal

plane movement; and the general perception of landing quality.[91] ‘Errors’

were allocated if the criteria was met on at least two of the three trials, and

were summed to calculate each participants’ LESS score.[91] Higher LESS scores


267

were representative of poorer jump-landing technique and used as an

indication of increased ACL injury risk.[91] Individuals who receive a LESS

score of five and above are proposed to be at greater risk of ACL injury.[104]

Therefore, participants were placed in a ‘high-’ or ‘low-risk’ group based on

whether their total LESS score met this threshold.

To score the Tuck Jump Assessment, frontal and sagittal video footage of

participants’ repeat tuck jump trial were assessed using the Tuck Jump

Assessment scoring criteria.[86, 256] Deficits in tuck jump technique are identified

via ten scoring items relating to knee and thigh motion, foot position during

landing, and plyometric technique (see Table 8.2, Page 269).[86] The number of

technique deficiencies were summed to give the participants’ Tuck Jump

Assessment score, resulting in a maximum attainable score of ten.[86] Higher

Tuck Jump Assessment scores were representative of more deficiencies in

jump-landing technique and used as an indication of increased ACL injury

risk.[86] Individuals who demonstrate six or more deficits in jump-landing

technique are proposed to be at greater risk of ACL injury.[86] Therefore,

participants were placed in a ‘high-’ or ‘low-risk’ group based on whether their

total Tuck Jump Assessment score met this threshold.


268

Table 8.1 Operational definitions for Landing Error Scoring System items. Reproduced with permission from Padua et al.[104]

LESS Item Operational Definition of Error Scoring

Knee Flexion Initial Contact

The knee is flexed less than 30O at initial contact. 0 = Absent 1 = Present

Hip Flexion Initial Contact

The thigh is in line with the trunk at initial contact. 0 = Absent 1 = Present

Trunk Flexion Initial Contact

The trunk is vertical or extended on the hips at initial contact. 0 = Absent 1 = Present

Ankle Plantarflexion Initial Contact

The foot lands heel to toe or with a flat foot at initial contact. 0 = Absent 1 = Present

Medial Knee Position Initial Contact

The center of the patella is medial to the midfoot at initial contact. 0 = Absent 1 = Present

Lateral Trunk Flexion Initial Contact

The midline of the trunk is flexed to the left or the right side of the body at initial contact.

0 = Absent 1 = Present

Stance Width Wide

The feet are positioned greater than a shoulder width apart (acromion processes) at initial contact.


Stance Width Narrow

The feet are positioned less than a shoulder width apart (acromion processes) at initial contact.


Foot Position External Rotation

The foot is internally rotated more than 30O between initial contact and maximum knee flexion.


Foot Position Internal Rotation

The foot is externally rotated more than 30O between initial contact and maximum knee flexion.


Symmetric Initial Foot Contact

One foot lands before the other foot or 1 foot lands heel to toe and the other foot lands toe to heel.


Knee Flexion Displacement

The knee flexes less than 45O between initial contact and maximum knee flexion.


Hip Flexion Displacement

The thigh does not flex more on the trunk between initial contact and maximum knee flexion.


Trunk Flexion Displacement

The trunk does not flex more between initial contact and maximum knee flexion.


Medial Knee Displacement

At the point of maximum medial knee position, the center of the patella is medial to the midfoot.


Joint Displacement Soft: the participant demonstrates a large amount of trunk, hip, and knee displacement. Average: the participant has some, but not a large amount of, trunk, hip, and knee displacement. Stiff: the participant goes through very little, if any, trunk, hip, and knee displacement.

0 = Soft 1 = Average 2 = Stiff

Overall Impression Excellent: the participant displays a soft landing with no frontal-plane or transverse plane motion. Poor: the participant displays large frontal-plane or transverse-plane motion, or the participant displays a stiff landing with some frontal-plane or transverse-plane motion. Average: all other landings

0 = Excellent 1 = Average 2 = Poor

LESS – Landing Error Scoring System


269

Table 8.2 Operational definitions for Tuck Jump Assessment items. Reproduced with permission from Myer et al.[86]

Tuck Jump Assessment Scoring Items Scoring

Is there knee valgus on landing? Are the hip, knee and foot aligned with no inward knee collapse?

0 = No Valgus 1 = Valgus Present

Do the thighs reach parallel at the peak of the jump? 0 = Parallel Thighs 1 = Thighs not Parallel

Are the thighs equal side-to-side during the flight phase of the jump? 0 = Equal Thighs 1 = Unequal Thighs

Are the feet placed shoulder width apart? 0 = Shoulder Width 1 = Narrow

Are the feet placed parallel front-to-back? 0 = Feet Parallel 1 = Feet not Parallel

Is foot contact timing equal? 0 = Symmetrical Contact 1 = Asymmetrical Contact

Is there a consistent point of landing? Does landing occur in the same foot print? 0 = Consistent 1 = Inconsistent

Is there excessive landing contact noise? 0 = Minimal Noise 1 = Excessive Noise

Is there a pause between jumps? 0 = No Pause 1 = Pause

Does technique decline over the 10 second period? 0 = Consistent Technique 1 = Technique Declines

Kinematic and kinetic data from the leap landing task were collected and

analysed as per the procedures outlined in Section 3.4.2 of Chapter Three.

Briefly, three-dimensional (3D) kinematics and kinetics of the lower limb were



Kingdom) sampling at 250 Hz, synchronised with a 600 mm by 900 mm in

ground force plate (Advanced Medical Technology Incorporated, Watertown,

MA, United States) sampling at 1000 Hz collected marker trajectory and

ground reaction force (GRF) data. Marker trajectory and GRF data were low-

pass filtered using a cubic smoothing spline with a cut-off frequency of 12 Hz.




270









moment data were extracted from the instance of initial contact (IC) through

to 150 ms post IC across all landing trials. This time window was chosen for

biomechanical variables as ACL injuries are thought to occur within the early

deceleration phase of landing.[161]


A waveform analysis, using one-dimensional statistical parametric mapping

(SPM1D) was implemented via open source code[369, 371] in Python 2.7 using

Canopy 1.5 (Enthought Incorporated, Austin, TX, United States). Specifically,

canonical correlations analysis (CCA),[362] the vector-field equivalent of SPM1D

linear regression, was used to examine the relationships between total

LESS/Tuck Jump Assessment scores and the 3D joint rotation and moment

waveform data extracted from the leap landing task. This analysis determined

the strength of the predictor variable, in this case LESS/Tuck Jump Assessment

score, at each time node across the joint rotation and moment waveforms.

Correlations between screening method scores were examined for the mean

hip and knee joint rotation (RXYZ) and joint moment (MXYZ) vector-fields from

each participants’ successful sport-specific landing trials. Planned post-hoc

analyses of joint rotation (RXY, RXZ, RYZ) and joint moment (MXY, MXZ, MYZ)

vector-field couples using CCA; and the individual joint rotation (RX, RY, RZ)


271

and joint moment (MX, MY, MZ) components using SPM1D linear regression

were also conducted.

The 3D kinematic and kinetic waveforms between ‘high-’ and ‘low-risk’

groups identified via both the LESS and Tuck Jump Assessment score

thresholds were compared using two-sample Hotelling’s T2 tests.[362]

Comparisons of ‘high-’ and ‘low-risk’ groups were undertaken for the mean

hip and knee joint rotation (RXYZ) and joint moment (MXYZ) vector-fields from

each participants’ successful landing trials. Planned post-hoc comparisons of

joint rotation (RXY, RXZ, RYZ) and joint moment (MXY, MXZ, MYZ) vector-field

couples using Hotelling’s T2 tests; and of the individual joint rotation (RX, RY,

RZ) and joint moment (MX, MY, MZ) components using SPM1D two-sample t-

tests were also conducted.

A Type I family-wise error rate of alpha = 0.05 was retained by calculating a

corrected threshold based on the number of vector components (i.e. joints = 2

and planes = 3) used for analysis.[362, 370] This correction resulted in an alpha

level of 0.0083 being used for all analyses. Subsequently, for each SPM statistic

a critical threshold was calculated at the corrected alpha level to determine

where statistically significant results were present. A statistically significant

result was identified where the SPM statistical curve exceeded this critical

threshold. A supra-threshold cluster indicated that the same result would be

produced by random curves in only 0.83% of repeated tests. The subsequent

probabilities (i.e. p-values) with which supra-threshold regions of the SPM

statistical curve could have resulted from repeated samplings of equally

smooth random curves were then calculated.


272

8.2.4.1 Intra- and Inter-Rater Reliability

High levels of intra- and inter-rater reliability have been previously

established for both the LESS and Tuck Jump Assessment methods.[86, 91, 94, 105,

256] Within this study, the same process was followed in assessing reliability for

both screening methods. A subset of ten participants’ videos were reviewed at

two time points, separated by one week to determine intra-rater (i.e. within

rater) reliability. A separate investigator also reviewed these ten participants’

videos, with scores compared across investigators to determine inter-rater (i.e.

between rater) reliability. Reliability for the individual scoring items/criterion

was assessed by calculating Cohen’s kappa (κ) in conjunction with the

percentage of exact agreement (PEA), using the equation:

Reliability of the total scores was assessed by calculating the mean difference

with 95% limits of agreement between raters.[433] For the LESS, the average PEA

and Cohen’s κ across all scoring criterion was 97.0% (κ = 0.933) (PEA range =

80.0 – 100.0%; κ range = 0.524 – 1.000) and 90.0% (κ = 0.745) (PEA range = 70.0

– 100.0%; κ range = 0.286 – 1.000) within- and between-raters, respectively. The

mean difference and 95% limits of agreement for total LESS scores were

calculated as 0.50 ± 1.03 and 0.40 ± 1.01 within- and between-raters,

respectively. For the Tuck Jump Assessment, the average PEA and Cohen’s κ

across all scoring criterion was 97.0% (κ = 0.940) (PEA range = 90.0 – 100.0%; κ

range = 0.737 – 1.000) and 84.0% (κ = 0.689) (PEA range = 70.0 – 100.0%; κ range

= 0.400 – 1.000) within- and between-raters, respectively. The mean difference

and 95% limits of agreement for total Tuck Jump Assessment scores were

calculated as 0.30 ± 0.94 and 1.00 ± 1.30 within- and between-raters,

respectively. The reliability values for both screening methods were similar to


273

those previously reported and accepted as reliable.[91, 94, 256] A full summary of

results from intra- and inter-rater reliability testing are presented in Appendix

I (see Page 417).

8.3 Results

8.3.1 LESS Results

The average LESS score across participants was 4.9 ± 2.3 (range = 0 – 9) (see

Figure 8.4). LESS scores were related to sagittal plane knee joint rotations (RX)

at 30 – 57 ms post IC (p = 0.003; see Figure 8.5, Page 274). Higher LESS scores

were associated with reduced knee flexion during this phase of the leap

landing task. CCA and SPM1D linear regression revealed no further

statistically significant relationships between LESS scores and the hip or knee

joint rotation or moment vector-field combinations or individual waveforms

from the leap landing task.

Figure 8.4 Histogram illustrating the distribution of Landing Error

Scoring System (LESS) scores across the participants

examined.


274

A.

B.

Figure 8.5 Relationship between Landing Error Scoring System (LESS)

score and knee flexion during the leap landing task. A) A

statistically significant relationship between LESS scores and knee flexion is indicated at 30 – 57 milliseconds (ms) after

initial contact (IC). B) Mean knee flexion during the leap

landing task of participants scoring in the lower (n = 4; solid line), upper (n = 7; dotted line) and middle (n = 21; dashed line)

quartiles for LESS score and highlights the relationship

between increasing LESS score and knee flexion.


275

The characteristics of participants allocated to the ‘high-’ and ‘low-risk’ groups

based on the LESS cut-off score of five and above are presented in Table 8.3.

Mean hip and knee joint rotations and moments from IC through to 150 ms

post IC during the leap landing task, respectively, from the ‘high-’ and ‘low-

risk’ groups are presented in Figure 8.6 (see Page 276) and Figure 8.7 (see Page

277).

Table 8.3 Characteristics of participants allocated to ‘high-’ and ‘low-

risk’ groups based on Landing Error Scoring System (LESS) score.

HIGH-RISK (n = 21)

LOW-RISK (n = 11)

AGE (y) 23.0 ± 3.1 23.5 ± 3.4

HEIGHT (cm) 170.3 ± 8.2 172.1 ± 6.6

MASS (kg) 67.2 ± 7.1 68.2 ± 10.3

LESS SCORE* 6.1 ± 1.4 2.5 ± 1.6

* - significantly higher scores observed for the ‘high-risk’ group

Two-sample t-tests revealed no statistically significant differences between the

‘high-’ and ‘low-risk’ groups for age, height and weight (p > 0.05). Significantly

higher LESS scores were observed for participants in the ‘high-risk’ group (p

< 0.001; 95% CI [2.48, 4.72]). SPM1D found no statistically significant

differences for hip and knee joint rotations or moments between groups.


276


through to 150 milliseconds (ms) post IC during the leap landing task of participants allocated as ‘high-’ and ‘low-risk’

based on Landing Error Scoring System (LESS) score. FLEX –

flexion; EXT – extension; ADD – adduction; ABD – adduction; IR – internal rotation; ER – external rotation.


277


contact (IC) through to 150 milliseconds (ms) post IC during the leap landing task of participants allocated as ‘high-‘ and

‘low-risk’ based on Landing Error Scoring System (LESS)

score. Mom. – moment; EXT – extension; FLEX – flexion; ABD – abduction; ADD – adduction; ER – external rotation; IR –

internal rotation.


278

8.3.2 Tuck Jump Assessment Results

The average Tuck Jump Assessment score across participants was 4.5 ± 1.6

(range = 2 – 8) (see Figure 8.8). CCA revealed no statistically significant

relationships between Tuck Jump Assessment scores at any time points across

the hip and knee joint rotation or moment vector-field combinations from the

leap landing task. SPM1D linear regression revealed no statistically significant

relationships between Tuck Jump Assessment scores at any time points across

the individual hip and knee joint rotation or moment waveforms from the leap

landing task.

Figure 8.8 Histogram illustrating the distribution of Tuck Jump

Assessment scores across the participants examined.


279

The characteristics of participants allocated to the ‘high-’ and ‘low-risk’ groups

based on the Tuck Jump Assessment cut-off score of six and above are

presented in Table 8.4. Mean hip and knee joint rotations and moments from

IC through to 150 ms post IC during the leap landing task, respectively, from

the ‘high-‘ and ‘low-risk’ groups are presented in Figure 8.9 (see Page 280) and


Table 8.4 Characteristics of participants distributed to ‘high-’ and ‘low-

risk’ groups based on Tuck Jump Assessment score.

HIGH-RISK (n = 9)

LOW-RISK (n = 23)

AGE (y) 21.6 ± 2.0 23.8 ± 3.3

HEIGHT (cm) 169.1 ± 4.9 171.6 ± 8.5

MASS (kg) 67.2 ± 10.7 67.7 ± 7.3

TUCK JUMP SCORE* 6.4 ± 0.7 3.8 ± 1.1

* - significantly higher scores observed for the ‘high-risk’ group

Two-sample t-tests revealed no statistically significant differences between the

‘high-’ and ‘low-risk’ groups for age, height and weight (p > 0.05). Significantly

higher Tuck Jump Assessment scores were observed for participants in the

‘high-risk’ group (p < 0.001; 95% CI [1.86, 3.47]). SPM1D found no statistically

significant differences for hip and knee joint rotations or moments between

groups.


280


through to 150 milliseconds (ms) post IC during the leap

landing task of participants allocated as ‘high-’ and ‘low-risk’

based on Tuck Jump Assessment (TJA) score. FLEX – flexion; EXT – extension; ADD – adduction; ABD – abduction; IR –

internal rotation; ER – external rotation.


281


contact (IC) through to 150 milliseconds (ms) post IC during

the leap landing task of participants allocated as ‘high-’ and

‘low-risk’ based on Tuck Jump Assessment (TJA) score. Mom. – moment; EXT – extension; FLEX – flexion; ABD – abduction;

ADD – adduction; ER – external rotation; IR – internal rotation.


282

8.4 Discussion

Screening methods that can identify individuals who employ high-risk

movement strategies during sporting tasks associated with ACL injuries are

likely to be the most effective in detecting those at-risk of ACL injury. The

purpose of this study was to examine the efficacy of the Landing Error Scoring

System (LESS) and Tuck Jump Assessment screening methods in identifying

high-risk lower limb biomechanics consistent with increased or ACL loading

or injury risk during a high-risk sport-specific landing task. Minimal

relationships between lower limb biomechanics during the leap landing and

scores received from either screening method were identified, suggesting

these screening methods may not be applicable for identifying ACL injury risk

within the netball-specific context examined.

This study found no relationships between participants’ hip and knee joint

biomechanics during a netball leap landing task linked with ACL injuries and

the scores they received from the Tuck Jump Assessment screening method.

Currently there are no studies that have validated the Tuck Jump Assessment

against ACL injury occurrences or laboratory-based 3D measures. The results

of this study and the paucity of literature validating the predictive ability of

the Tuck Jump Assessment makes it difficult to recommend this method for

ACL injury risk screening purposes.

A significant relationship between LESS scores and knee flexion during the

leap landing task was identified in this study. Higher LESS scores (i.e.

suggestive of increased ACL injury risk) were associated with reduced knee

flexion during an early portion (30 – 57 ms post IC) of the landing. Limited

knee flexion during landing has been linked to higher ACL loads[25] and ACL

injuries often occur with a more extended knee.[97, 99] In addition; ACL injuries


283

are thought to occur within the early deceleration phase of landing,[161]

highlighting the importance of sufficient knee flexion during this phase. The

link between higher LESS scores and reduced knee flexion identified within

this study suggests this screening method may have some utility in identifying

high-risk movement strategies during high-risk sporting tasks. However, no

additional relationships were identified between LESS scores and hip and

knee joint rotations or moments across other planes. Injuries to the ACL are

proposed to occur via a multiplanar mechanism,[33] with frontal and transverse

plane biomechanics acknowledged as key factors in modulating ACL

loads.[20-23, 28, 32, 160] Therefore, a screening method that detects biomechanical

deficits during high-risk sporting tasks in a single plane alone may not be

efficacious in identifying athletes at greater risk of ACL injury.

The relationship between LESS scores and knee flexion is consistent with

earlier work,[91] however the absence of a relationship across other joint

rotations and moments is in contrast to the findings of others.[91, 94] Lower LESS

scores have been associated with decreased hip flexion; increased hip

adduction and knee valgus angles and moments; and increased hip and knee

internal rotation angles during bilateral drop vertical jump-landings.[91] These

contrasting findings may have resulted from the different tasks used to

evaluate the ability of LESS scores to identify high-risk landing biomechanics.

Generic bilateral landing tasks may not adequately represent the lower limb

biomechanics experienced during high-risk sport-specific manoeuvres,[95, 96]

particularly those important for evaluating ACL injury risk. Kristianslund and

Krosshaug[96] found poor correlations for knee abduction moments between

bilateral drop vertical jumps and sport-specific side-step cutting. Knee

abduction angles during bilateral drop vertical jumps were also poorly

correlated to knee abduction moments during the side-step cutting task.[96]


284

Comparison of a bilateral drop landing to a sport-specific volleyball spike

jump-landing revealed differences in landing forces and lower limb motion

and alignment.[95] Even when applied to a bilateral sporting task, generic

landing movements still may not represent what is experienced in a

competitive sporting environment.[95]

The different measurement methods used across tasks within this study may

have also played a role in the limited relationships between screening method

scores and lower limb biomechanics from the leap landing task. Three-

dimensional methods, such as those used to monitor leap landing

performance, are more reliable and sensitive to high-risk lower limb

biomechanics in comparison to the two-dimensional observational

assessments used as part of the LESS and Tuck Jump Assessment.[85] It may be

that application of three-dimensional methods to the scoring criteria of the

LESS and Tuck Jump Assessment would improve their relationship to lower

limb biomechanics from the leap landing task. As the purpose of this study

was to test the efficacy of these screening methods in the manner in which they

are intended to be used (i.e. with two-dimensional visual observation), three-

dimensional data from the bilateral drop vertical jump and tuck jump tasks

was not used. Therefore, this study cannot confirm whether the different

measurement methods used across tasks was a factor in the limited

relationships found. However, as previous work has shown that LESS scores

are sensitive to three-dimensional measures during drop vertical jump-

landings,[91, 94] it is likely that the different landing task used in this study

played a major role in the findings. Nonetheless, the limited ability of generic

bilateral landing tasks to predict lower limb postures during high-risk sport-

specific manoeuvres poses the question as to whether these tasks are

appropriate for validating ACL injury risk screening methods. Examining


285

ACL injury risk screening methods against sport-specific movements

associated with ACL injuries may be a more appropriate approach in

determining the efficacy of these methods in detecting high-risk landing

postures.

It is also important to consider the population examined in this study and how

this may have impacted on the ability of the screening methods to detect high-

versus low-risk lower limb biomechanics. The use of a healthy cohort with no

previous history of serious lower limb injury may present a ‘biased sample.’

These participants may not have displayed aberrant lower-limb biomechanics,

with this impacting on the ability of the screening methods to detect higher-

versus lower-risk athletes. Despite the use of a healthy cohort, a large range of

values was observed across key biomechanical variables associated with ACL

injury risk (see Figure 8.11, 286). Similarly large ranges were also observed for

LESS (see Figure 8.4, Page 273) and Tuck Jump Assessment (see Figure 8.8,

Page 278) scores. In combination these data support the finding that both LESS

and Tuck Jump Assessment scores had minimal relationships to lower limb

biomechanics during the leap landing.


286

A.

B.

C.

D.

E.

F.

Figure 8.11 Histogram illustrating the distribution of values for A) knee flexion at initial contact; B) peak knee flexion; C) knee

abduction at initial contact; D) peak knee abduction; E) peak

knee abduction moment; and E) peak knee internal rotation moment during the leap landing task across the participants

examined.

Participants allocated to the ‘high-risk’ groups based on the recommended

cut-off scores from Padua et al.[104] and Myer et al.[86] had significantly higher

(i.e. suggestive of increased ACL injury risk) LESS and Tuck Jump Assessment

scores compared to the ‘low-risk’ groups. However, no differences in hip and

knee joint rotations and moments from the leap landing task were found

between these ‘high-’ and ‘low-risk’ groups. While it would be expected that


287

those in the ‘high-risk’ groups would exhibit biomechanical deficits associated

with ACL loading or injury risk, the cross-sectional nature of this study is a

limiting factor in determining whether the ‘high-risk’ individuals as

determined by the LESS and Tuck Jump Assessment methods were truly at a

greater risk of ACL injury. As no follow-up data pertaining to ACL injuries

was collected, these results cannot confirm that the screening methods

examined do not identify athletes’ at-risk of future ACL injury. Interestingly,

a substantially different proportion of participants were allocated to the ‘high-

’ and ‘low-risk’ groups for the two different screening methods. While almost

two-thirds of participants were considered ‘high-risk’ based on their LESS

score, a similar number of participants were deemed ‘low-risk’ based on their

Tuck Jump Assessment score (see Table 8.3, Page 275 and Table 8.4, Page 279,

respectively). Further examination revealed that agreement between the

methods for ‘high-’ and ‘low-risk’ was only found for six and eight

participants, respectively. The large number of participants deemed as ‘high-

risk’ by the LESS could suggest the cut-off value used was overly sensitive,

while the small number of participants deemed ‘high-risk’ by the Tuck Jump

Assessment could indicate the cut-off value was not sensitive enough. It may

also be that the screening methods and cut-off values used were not applicable

to the population examined. Padua et al.[104] found that LESS scores were

capable of identifying elite youth soccer athletes at-risk of future ACL injury

with 86% and 64% sensitivity and specificity, respectively. In contrast, Smith

et al.[105] found no relationship between future ACL injury occurrences and

LESS scores within a cohort of high school and collegiate athletes from a range

of sports. The populations examined across these previous studies and the

current study differ in age, gender and the sports played. This may be an

important factor in the efficacy of a screening method in identifying athletes

at-risk of ACL injury. While the mechanism for noncontact ACL injuries


288

appears consistent across sports, the tasks being performed in the lead up or

at the injury event may differ.[5, 101-103] Sport-specific variations in athletic

manoeuvres can also have a modulating effect on lower limb biomechanics.[154,

155] A ‘one-size fits all’ approach with regard to the use of screening methods,

and how cut-off values for determining injury risk are applied, across different

populations and sports may not be appropriate. A screening methods efficacy

in identifying high-risk movement strategies and/or individuals may vary

across different high-risk sport-specific tasks. Examination of screening

methods against a range of sport-specific tasks associated with ACL injuries

may be valid for identifying the most appropriate method(s) across different

sports.

Evaluating an athlete’s movement technique during more ecologically valid

tasks relevant to their sport, comparable to those seen in competitive

environments when injuries occur, is a potential strategy for improving the

efficacy of ACL injury risk screening. There are, however, advantages and

disadvantages to consider regarding this approach. Movements such as

unplanned side-stepping more closely emulate a competitive environment

and the potential ACL injury mechanism by limiting decision time and

increasing knee loading.[247] The inclusion of sport-specific factors, such as

carrying or catching a ball, have also been shown to influence knee angles and

moments during athletic manoeuvres.[154, 155] Screening athletes’ movement

technique in these more game-like situations could theoretically yield greater

details regarding their ability to avoid an injurious event in a competitive

context. However, exposing athletes’ to a potentially high-risk task for the

purpose of injury risk screening may be undesirable for coaches. In addition,

a number of factors pertaining to technique can influence important

biomechanical measures of injury risk during these tasks[304] which could


289

increase the analysis and time burdens associated with screening. Movements

that are safe to perform, and simplistic to observe and evaluate may be the

most applicable for injury risk screening. Investigating the relationships

between movement tasks that fit these criteria and high-risk sport-specific

manoeuvres could identify relevant movements for field-based ACL injury

risk screening protocols. This will be undertaken in the following chapter,

exploring the relationship between a simplistic unilateral task and a netball

leap landing in order to examine the utility of including a unilateral movement

within netball-specific ACL injury risk screening protocols.

It is important to acknowledge that this study focused on one sport-specific

manoeuvre associated with ACL injuries, namely, a netball leap landing. The

findings from this study, therefore, are only applicable to ACL injury risk

screening in the netball-specific context examined. The different landing task

examined within this study may be a potential reason for the dissimilar

findings to previous works[91, 94] that have examined the validity of screening

methods in identifying high-risk landing postures. This does, however,

highlight the need to explore the validity of screening methods against a range

of high-risk sporting tasks.


290

8.5 Conclusions

This study identified minimal relationships between lower limb biomechanics

during a netball leap landing and scores received from the LESS and Tuck

Jump Assessment ACL injury risk screening methods. These findings suggest

the LESS and Tuck Jump Assessment may not be applicable for ACL injury

risk screening in the netball-specific context examined. Exploring the

relationships between additional movement tasks and leap landings may

reveal more applicable approaches for identifying athletes’ at-risk of ACL

injury within this sport. Considering the unilateral nature of leap landings and

that the majority of noncontact ACL injuries occur via a unilateral mechanism,

screening methods incorporating unilateral movements may be more

appropriate for field-based ACL injury risk screening protocols.

291

Chapter Nine: The Relationship between Performance of a Single-Leg Squat and Leap Landing Task: Moving Towards a Netball-Specific Anterior Cruciate Ligament Injury Risk Screening Method

9.1 Introduction

Identifying athletes’ at-risk of anterior cruciate ligament (ACL) injury appears

to be an important step in effective and efficient ACL injury prevention

practice.[84] Athletes’ who demonstrate high-risk movement strategies appear

to receive the greatest improvements in biomechanical factors associated with

ACL injury following neuromuscular training.[71, 83, 84] Screening methods that

are capable of detecting athletes who utilise high-risk movement strategies

may therefore be useful in identifying those who will benefit the most from

intervention. Further, screening methods may be useful in identifying specific

biomechanical deficits that require targeting within ACL injury prevention

programs.[358] Noncontact ACL injuries occur at the highest rate in sports

which require frequent cutting and landing movements.[1] Therefore, the need

for effective ACL injury risk screening protocols within these sports is

imperative.

Netball is often recognised as a particularly high-risk sport for ACL

injuries.[234-236] Within netball, landings are often implicated as the high-risk

task for ACL injuries.[103, 238, 245] In particular, a leap landing while catching a

pass is a common task associated with noncontact ACL injuries within the

Single-Leg Squat and Leap Landings

292

sport.[103, 245] Screening methods that identify athletes who employ movement

strategies associated with ACL loading and injury risk during this type of

landing manoeuvre are likely to be the most effective in detecting netball

players at-risk of ACL injury. The previous study found two established field-

based screening methods for ACL injury risk, the Landing Error Scoring

System (LESS) and Tuck Jump Assessment, were ineffective in identifying

lower limb biomechanics consistent with increased ACL loading during a

netball-specific leap landing manoeuvre performed while catching a pass. It

was proposed that the bilateral athletic tasks (i.e. drop vertical jump-landing

and repeated tuck jumps) used within these screening methods were not

associated with the high-risk landing task examined, and was a contributing

factor to the results. The effectiveness of field-based screening methods for

ACL injury risk appears to vary across different sports and sporting

populations.[104, 105] Therefore, a ‘one-size fits all’ approach to the use of

screening methods across different sports and populations may not be

appropriate. The limited effectiveness of the LESS and Tuck Jump Assessment

methods highlighted in the previous study indicates that further exploration

of ACL injury risk screening methods in a netball-specific context is required.

Investigating athletic tasks that are better linked to leap landings may identify

more optimal movements for ACL injury risk screening within a netball-

specific context.

The single-leg squat is a simplistic, commonly used movement for assessing

lower limb function across a range of clinical contexts.[257, 434, 435] Numerous

studies have highlighted the ease of use and reliability of visual assessment of

the single-leg squat.[400, 422, 423, 434, 436, 437] Substantial agreement is often reported

within- and between-raters for visual rating criteria across a range of

movement characteristics, such as medio-lateral knee motion;[422, 437] frontal


293

plane knee angle and alignment;[400, 423, 436] pelvis motion and alignment;[423]

frontal plane foot motion;[437] and general overall movement quality.[434, 437] A

key factor in the reliability and ease of use for visual ratings of single-leg squat

performance may be the speed of movement.[423] Faster movements, such as

drop vertical jump-landings, are more difficult to visually rate.[438] Therefore,

the speed at which the single-leg squat is performed makes it an easy

movement to visually observe. This highlights a beneficial aspect of using the

single-leg squat as a potential screening movement for field-based assessment

of ACL injury risk.

There is also evidence that movement characteristics during single-leg squats

are related to lower limb biomechanics during unilateral cutting and landing

tasks.[439, 440] Strong correlations (Pearson’s r > 0.50) have been observed

between the magnitude of frontal and transverse plane knee angular

displacement between a single-leg squat and single-leg drop landing task.[439]

Further, moderate to strong associations (r = 0.38 – 0.76) between peak hip

flexion, hip internal rotation, knee flexion and knee abduction angles between

a single-leg squat and side-step cutting task have also been observed.[440]

However, in direct contrast to these results, Stensrud et al.[441] found poor

correlations for knee abduction angle between a single-leg squat and single-

leg drop vertical jump task. Marshall et al.[439] also found no association

between hip and knee biomechanical variables when examining the

relationship between a single-leg squat and side-step cut. Methodological

inconsistencies may be a reason for these conflicting findings. The landing and

cutting tasks examined largely varied across studies, ranging from single-leg

drop vertical jumps[441] and drop landings,[439] to side-step cutting tasks

performed at 75[439] and 90 degree angles.[440] Screening for ACL injury risk


294

using a single-leg squat may therefore only be relevant to sports where the

high-risk task(s) are associated with single-leg squat performance.

The relationship between single-leg squat performance and lower limb

biomechanics during a netball leap landing are yet to be examined. The single-

leg squat may be a particularly relevant movement for ACL injury risk

screening within a netball context. The hip and knee motion experienced

during the initial deceleration phase of leap landings, where unilateral weight

acceptance is required, closely resembles the lower limb positioning

experienced during the descent phase of a single-leg squat. Due to this,

movement dysfunctions that present during single-leg squat performance

may also present during leap landings. Should this be the case, the single-leg

squat may be a useful movement to incorporate within netball-specific field-

based ACL injury risk screening protocols.

The purpose of this study was to determine the relationship between the

performance of a single-leg squat and a netball-specific leap landing task, in

order to examine the utility of including a single-leg squat movement within

netball-specific ACL injury risk screening protocols. Specifically, this study

aimed to identify whether lower limb biomechanical patterns from a single-

leg squat task were associated with hip and knee lower limb biomechanical

patterns from a leap landing. Due to the similarities in lower limb postures

between tasks, it was hypothesised that lower limb biomechanical patterns

from the single-leg squat would be associated with hip and knee lower limb

biomechanical patterns from the leap landing task.


295

9.2 Methodology




9.2.1 Participants

Thirty-two female netball players (age = 23.2 ± 3.2 years; height = 171.3 ± 7.8

cm; mass = 68.0 ± 8.2 kg; netball experience = 12.45 ± 4.54 years) volunteered

and provided written consent to participate in this study. All participants met

the inclusion/exclusion criteria detailed in Section 3.2 of Chapter Three.


Data from the single-leg squat and netball-specific landing task (i.e. the leap

landing), as described in Sections 3.3.1 and 3.3.4, respectively, performed

during testing session one (i.e. baseline) were used within this study. The

order of these tasks were randomised within the testing session across

participants. Participants performed one successful trial of the single

repetition single-leg squat task (see Figure 9.1, Page 296). Only one trial was

used for analysis as it was thought this would replicate the use of the single-

leg squat within a field-based screening context. The trial was performed on

the same limb that was used for the leap landing task. Participants performed

20 successful trials of the leap landing on a predetermined ‘landing limb.’ The

landing limb was determined by asking participants which limb they would

be more comfortable landing on when catching a pass during a game. Further

to this, players were given an opportunity to trial landing on each limb to

identify which they were more comfortable landing on. Each landing trial was

initiated by breaking and accelerating from a static defender (model skeleton)


296

from a position six-metres from the landing area, followed by a run-up and

leap landing whilst catching a pass (see Figure 9.2). Trials were repeated until

the 20 successful trials were recorded. Participants were allocated 60-90

seconds rest between trials in an effort to minimise

fatigue.[134, 255] Kinematic and kinetic data collected from the participants

‘landing limb’ during the single-leg squat task and all landing trials were used

within this study.

Figure 9.1 Sagittal view of the single-leg squat task.



297


Kinematic and kinetic data from the single-leg squat and leap landing tasks

were collected and analysed as per the procedures outlined in Section 3.4.2 of

Chapter Three. Briefly, three-dimensional (3D) kinematics and kinetics of the

lower limb were computed using an established musculoskeletal model.[305, 310]

An eight camera 3D motion capture system (Vicon, Oxford Metrics Limited,

Oxford, United Kingdom) sampling at 250 Hz, synchronised with a 600 x 900

mm in ground force plate (Advanced Medical Technology Incorporated,

Watertown, MA, United States) sampling at 1000 Hz collected marker

trajectory and ground reaction force (GRF) data. Marker trajectory and GRF

data from the single-leg squat and leap landing tasks were low-pass filtered

using a cubic smoothing spline with cut-off frequencies of 9 and 12 Hz,

respectively. As this study aimed to identify biomechanical patterns from the

single-leg squat task that were relevant for use within a netball-specific field-

based screening method for ACL injury risk, only biomechanical patterns that

could be visually observed in this task were deemed applicable. Hence, only

kinematic data were extracted and examined from the single-leg squat task.

Three-dimensional hip, knee and ankle joint rotations were calculated for the

single-leg squat task from filtered marker trajectory data using an X-Y-Z

Cardan angles sequence; with the X, Y and Z axes representing joint rotations

in the sagittal, frontal and transverse planes, respectively.[138] Kinematic data

were expressed relative to each participant’s stationary (neutral) trial.[56, 310, 311]

Joint rotation data were extracted from the ‘squatting phase’ of the task. The

points where the musculoskeletal models centre of mass began descending

and returned to its original position defined the start and end points of the

‘squatting phase.’ These points tended to correspond with the initiation and

completion of hip and/or knee flexion, hence they were deemed as applicable


298

points for defining the squatting movement. Extracted data from the single-

leg squat task were time-normalised to 101 data points (i.e. 0-100% of the

squatting phase). Three-dimensional hip and knee joint rotations were

calculated for the leap landing task as per the methods outlined above. In

addition, three-dimensional hip and knee external joint moments from the

leap landing task were calculated by submitting the filtered marker trajectory

and GRF data to a conventional inverse dynamics analysis.[261] Joint moments

were expressed in the reference frames of the joint coordinate systems and

were normalised by dividing the total joint moment (Nm) by the product of

the participant’s weight and height (kgm). Joint rotation and moment data

were extracted from the instance of initial contact (IC) to 150 ms post IC across



landing.[161]

Both one-dimensional statistical parametric mapping (SPM1D) and principal

component analysis (PCA) have been used within previous studies of this

thesis to examine joint rotation and moment data. The purpose of this study

was to examine the relationship between lower limb biomechanical patterns

from the single-leg squat and leap landing task. As PCA takes into account

features that occur over specific phases of the waveform, it was deemed as the

more appropriate choice for identifying and correlating patterns across the

two tasks. Data were exported to MATLAB (The Mathworks Incorporated,

Natick, MA, United States) for further analyses. Kinematic waveform data

extracted from the single-leg squat task were submitted to PCA to identify

lower limb biomechanical patterns during squatting performance. Patterns

were extracted for hip flexion/extension (HFLEX/EXT), adduction/abduction

(HADD/ABD), and internal/external rotation (HIR/ER); knee extension/flexion


299

(KEXT/FLEX), adduction/abduction (KADD/ABD), and internal/external rotation

(KIR/ER); and ankle dorsiflexion/plantarflexion (ADF/PF), inversion/eversion

(AINV/EVE), and internal/external rotation (AIR/ER) joint rotations. The n number

of principal patterns (PP) that summed to explain at least 90% of the variation

in the waveform were retained and referred to as PP1, PP2, …, PPn.[276, 327] The

90% criterion was chosen as this value has been previously used when

examining lower limb biomechanical waveforms.[326, 335] Principal pattern

scores (PP-Scores) were then computed for each biomechanical pattern across

each joint rotation variable. Principal pattern interpretation was achieved by

examining the waveforms obtained by adding and subtracting a suitable

multiple of the patterns loading vector to the mean of the relevant

waveform.[333] An identical process was then followed to identify the principal

hip and knee biomechanical patterns from the leap landing task, where PCA

was applied to the extracted kinematic and kinetic data.


Pearson’s correlation coefficients (r) were calculated between the PP-Scores

from the single-leg squat and leap landing to identify relationships between

the two tasks. As a number of associations were tested, a stricter alpha level of

0.01 was set for identifying statistically significant correlations.[89] The strength

of correlations was guided by Cohen[384] (weak r = 0.1; moderate r = 0.3; strong

r = 0.5). PP-Scores that were statistically significant (p < 0.01) and strongly

correlated (r > 0.50) were considered indicative of a relationship between the

two tasks.


300

9.3 Results

Principal patterns were extracted for single-leg squat joint rotations, and leap

landing joint rotations and moments via PCA (see Table 9.1, Page 302; Table

9.2, Page 303; and Table 9.3, Page 304, respectively). A graphical display of the

patterns associated with high and low PP-scores for each PP from the single-

leg squat and leap landing tasks are provided in Appendix J (see Page 419)

and Appendix K (see Page 432), respectively. The Pearson’s correlation values

(r) between the kinematic principal patterns from the single-leg squat, and the

kinematic and kinetic principal patterns, respectively, from the leap landing

task are presented in Table 9.4 (see Page 305) and Table 9.5 (see Page 306). For

ease of interpretation, the strong statistically significant (i.e. r > 0.50; p < 0.01)

correlations are highlighted.

Hip internal rotation during the squat was associated with greater overall hip

internal rotation during landing, while hip external rotation during the squat

was associated with greater overall hip external rotation during landing. The

overall magnitude of knee flexion was related across the two tasks, whereby

greater knee flexion during the squat was associated with greater knee flexion

during landing. The timing of peak knee flexion during the squat was

associated with prolonged hip flexion and the magnitude of transverse plane

hip motion during landing. Earlier onset of peak knee flexion during the squat

was related to a shortened duration of hip flexion and greater overall hip

external rotation during landing. In contrast, delayed peak knee flexion during

the squat was related to prolonged hip flexion and greater overall hip internal

rotation. Knee abduction during the squat was associated with greater overall

knee abduction and reduced knee external rotation from IC to 30 ms post IC

during landing; while knee adduction during the squat was associated with


301

reduced knee abduction but greater knee adduction and greater knee internal

rotation from IC to 30 ms post IC. Knee internal rotation during the squat was

associated with greater overall knee internal rotation and greater knee

adduction moments during landing. Further, greater transverse plane knee

motion during the squat was associated with reduced knee external rotation

from IC to approximately 30 ms post IC, reduced hip abduction moments from

IC to approximately 30 ms post IC, and a delayed and reduced peak knee

abduction moment during landing. Greater overall ankle dorsiflexion during

the squat was associated with greater overall knee flexion during landing;

while reduced ankle dorsiflexion range of motion and delayed time to peak

ankle dorsiflexion during the squat was associated with prolonged hip flexion

and greater overall hip internal rotation during landing. Reduced ankle

eversion during the squat was associated with greater overall knee abduction,

greater overall knee internal rotation, and reduced knee adduction moments

during landing. In contrast, greater overall ankle eversion during the squat

was associated with reduced and greater overall knee adduction, respectively,

reduced overall knee internal rotation, and greater knee adduction moments

during landing.

302


Tabl

e 9.

1 D

escr

iptio

ns o

f pri

ncip

al p

atte

rns

(PPs

) ext

ract

ed fr

om s

ingl

e-le

g sq

uat h

ip, k

nee

and

ankl

e jo

int r

otat

ion

data

.

Join

t Rot

atio

n PP

D

escr

iptio

n V

aria

nce

Expl

aine

d

HFL

EX/E

XT

1 2 G

ener

al s

hape

and

ove

rall

mag

nitu

de –

hig

her s

core

= g

reat

er o

vera

ll hi

p fle

xion

Ph

ase

shift

in ti

min

g –

high

er sc

ore

= de

laye

d pe

ak h

ip fl

exio

n 79

.3%

14

.9%

H

AD

D/A

BD

1 2 3

Gen

eral

sha

pe a

nd o

vera

ll m

agni

tude

– h

ighe

r sco

re =

gre

ater

ove

rall

hip

addu

ctio

n M

agni

tude

at ~

0-25

% a

nd ~

75-1

00%

of s

quat

– h

ighe

r sco

re =

gre

ater

hip

add

uctio

n at

~0-

25%

and

~75

-100

%

Phas

e sh

ift in

tim

ing

– hi

gher

scor

e =

earl

ier p

eak

hip

addu

ctio

n

77.8

%

10.7

%

5.9%

H

IR/E

R

1 2 G

ener

al s

hape

and

ove

rall

mag

nitu

de –

hig

her s

core

= re

duce

d ov

eral

l hip

ext

erna

l rot

atio

n M

agni

tude

of p

eaks

– h

ighe

r sco

re =

gre

ater

hip

inte

rnal

and

ext

erna

l rot

atio

n pe

aks

81.3

%

9.9%

K

EXT/

FLEX

1 2

Gen

eral

sha

pe a

nd o

vera

ll m

agni

tude

– h

ighe

r sco

re =

redu

ced

over

all k

nee

flexi

on

Phas

e sh

ift in

tim

ing

– hi

gher

scor

e =

dela

yed

peak

kne

e fle

xion

72

.9%

17

.2%

K

AD

D/A

BD

1 2 G

ener

al s

hape

and

ove

rall

mag

nitu

de –

hig

her s

core

= g

reat

er o

vera

ll kn

ee a

dduc

tion

Mag

nitu

de o

f pea

ks –

hig

her s

core

= g

reat

er k

nee

addu

ctio

n an

d ab

duct

ion

peak

s 87

.3%

5.

7%

KIR

/ER

1 2 G

ener

al s

hape

and

ove

rall

mag

nitu

de –

hig

her s

core

= g

reat

er o

vera

ll kn

ee in

tern

al ro

tatio

n M

agni

tude

of t

rans

vers

e pl

ane

knee

ang

ular

dis

plac

emen

t – h

ighe

r sco

re =

gre

ater

ang

ular

dis

plac

emen

t 84

.9%

9.

3%

AD

F/PF

1 2 3

Gen

eral

sha

pe a

nd o

vera

ll m

agni

tude

– h

ighe

r sco

re =

gre

ater

ove

rall

ankl

e do

rsifl

exio

n D

orsi

flexi

on R

OM

and

tim

e to

pea

k do

rsifl

exio

n –

high

er s

core

= re

duce

d RO

M a

nd d

elay

ed p

eak

Dor

sifle

xion

RO

M a

nd ti

me

to p

eak

dors

iflex

ion

– hi

gher

sco

re =

redu

ced

ROM

and

ear

lier p

eak

71.0

%

16.0

%

8.7%

A

INV

/EV

E 1

Gen

eral

sha

pe a

nd o

vera

ll m

agni

tude

– h

ighe

r sco

re =

redu

ced

over

all a

nkle

eve

rsio

n 91

.6%

A

IR/E

R

1 2 3 4

Gen

eral

sha

pe a

nd o

vera

ll m

agni

tude

– h

ighe

r sco

re =

redu

ced

over

all a

nkle

ext

erna

l rot

atio

n M

agni

tude

at ~

0-20

% o

f squ

at –

hig

her s

core

= re

duce

d an

kle

exte

rnal

rota

tion

at ~

0-20

%

Mag

nitu

de a

t ~75

-100

% o

f squ

at –

hig

her s

core

= re

duce

d an

kle

exte

rnal

rota

tion

at ~

75-1

00%

M

agni

tude

at ~

60-7

5% o

f squ

at –

hig

her s

core

= g

reat

er a

nkle

ext

erna

l rot

atio

n at

~60

-75%

61.3

%

13.4

%

9.1%

7.

0%

HFL

EX/E

XT

– hi

p fle

xion

/ext

ensi

on;

HA

DD

/ABD

– h

ip a

dduc

tion/

abdu

ctio

n; H

IR/E

R –

hip

int

erna

l/ext

erna

l ro

tatio

n; K

EXT/

FLEX

– k

nee

exte

nsio

n/fle

xion

; K

AD

D/A

BD –

kne

e ad

duct

ion/

abdu

ctio

n; K

IR/E

R –

kne

e in

tern

al/e

xter

nal r

otat

ion;

AD

F/PF

– a

nkle

dor

sifle

xion

/pla

ntar

flexi

on; A

INV

/EV

E –

ankl

e in

vers

ion/

ever

sion

; AIR

/ER –

ank

le in

tern

al/e

xter

nal

rota

tion;

PP

– pr

inci

pal p

atte

rn; R

OM

– ra

nge

of m

otio

n.

303


Tabl

e 9.

2 D

escr

iptio

ns o

f pri

ncip

al p

atte

rns

(PPs

) ext

ract

ed fr

om le

ap la

ndin

g hi

p an

d kn

ee jo

int r

otat

ion

data

.

Join

t Rot

atio

n PP

D

escr

iptio

n V

aria

nce

Expl

aine

d

HFL

EX/E

XT

1 2 G

ener

al s

hape

and

ove

rall

mag

nitu

de –

hig

her s

core

= g

reat

er o

vera

ll hi

p fle

xion

D

urat

ion

of h

ip fl

exio

n –

high

er s

core

= p

rolo

nged

pos

ition

of h

ip fl

exio

n th

roug

h la

ndin

g 86

.4%

7.

7%

HA

DD

/ABD

1 2

Gen

eral

sha

pe a

nd o

vera

ll m

agni

tude

– h

ighe

r sco

re =

gre

ater

ove

rall

hip

addu

ctio

n M

agni

tude

of f

ront

al p

lane

hip

ang

ular

dis

plac

emen

t – h

ighe

r sco

re =

gre

ater

ang

ular

dis

plac

emen

t 86

.7%

9.

7%

HIR

/ER

1 2 G

ener

al s

hape

and

ove

rall

mag

nitu

de –

hig

her s

core

= g

reat

er o

vera

ll hi

p in

tern

al ro

tatio

n M

agni

tude

of h

ip e

xter

nal r

otat

ion

from

IC to

~30

ms

post

IC –

hig

her s

core

= re

duce

d hi

p ex

tern

al ro

tatio

n 74

.0%

17

.4%

K

EXT/

FLEX

1

Gen

eral

sha

pe a

nd o

vera

ll m

agni

tude

– h

ighe

r sco

re =

redu

ced

over

all k

nee

flexi

on

90.6

%

KA

DD

/ABD

1 2 3

Gen

eral

sha

pe a

nd o

vera

ll m

agni

tude

– h

ighe

r sco

re =

redu

ced/

grea

ter o

vera

ll kn

ee a

bduc

tion/

addu

ctio

n Ph

ase

shift

in ti

min

g –

high

er sc

ore

= ea

rlie

r kne

e ad

duct

ion

peak

Pr

esen

ce o

f bim

odal

pea

ks in

cur

ve –

hig

her s

core

= p

rese

nce

of s

ingu

lar p

eak

in fr

onta

l pla

ne k

nee

rota

tion

77.7

%

10.7

%

7.8%

K

IR/E

R

1 2 3

Gen

eral

sha

pe a

nd o

vera

ll m

agni

tude

– h

ighe

r sco

re =

gre

ater

ove

rall

knee

inte

rnal

rota

tion

Mag

nitu

de o

f kne

e ex

tern

al ro

tatio

n fr

om IC

to ~

30 m

s po

st IC

– h

ighe

r sco

re =

redu

ced

knee

ext

erna

l rot

atio

n M

agni

tude

of p

eak

– hi

gher

sco

re =

gre

ater

pea

k kn

ee in

tern

al ro

tatio

n

73.4

%

15.3

%

8.7%

H

FLEX

/EX

T –

hip

flexi

on/e

xten

sion

; H

AD

D/A

BD –

hip

add

uctio

n/ab

duct

ion;

HIR

/ER –

hip

int

erna

l/ext

erna

l ro

tatio

n; K

EXT/

FLEX

– k

nee

flexi

on/e

xten

sion

; K

AD

D/A

BD –

kne

e ad

duct

ion/

abdu

ctio

n; K

IR/E

R –

kne

e in

tern

al/e

xter

nal r

otat

ion;

PP

– pr

inci

pal p

atte

rn; m

s –

mill

isec

onds

; IC

– in

itial

con

tact

.

304


Tabl

e 9.

3 D

escr

iptio

ns o

f pri

ncip

al p

atte

rns

(PPs

) ext

ract

ed fr

om le

ap la

ndin

g hi

p an

d kn

ee jo

int m

omen

t dat

a.

Join

t Mom

ent

PP

Des

crip

tion

Var

ianc

e Ex

plai

ned

HEX

T/FL

EX M

om.

1 2 3

Gen

eral

sha

pe a

nd o

vera

ll m

agni

tude

– h

ighe

r sco

re =

redu

ced

hip

flexi

on m

omen

t M

agni

tude

of p

eak

– hi

gher

sco

re =

redu

ced

peak

hip

flex

ion

mom

ent

Mag

nitu

de fr

om ~

30-6

0 m

s pos

t IC

– h

ighe

r sco

re =

redu

ced

hip

flexi

on m

omen

t fro

m ~

30-6

0 m

s po

st IC

56.8

%

27.0

%

6.7%

H

ABD

/AD

D M

om.

1 2 3

Gen

eral

sha

pe a

nd o

vera

ll m

agni

tude

– h

ighe

r sco

re =

redu

ced

over

all h

ip a

dduc

tion

mom

ent

Mag

nitu

de fr

om IC

to ~

30m

s po

st IC

– h

ighe

r sco

re =

gre

ater

hip

abd

uctio

n m

omen

t fro

m IC

to ~

30 m

s po

st IC

Pr

esen

ce o

f bim

odal

pea

ks in

cur

ve –

hig

her s

core

= p

rese

nce

of h

ip a

bduc

tion

and

addu

ctio

n m

omen

t pea

ks

74.0

%

15.4

%

5.2%

H

ER/I

R M

om.

1 2 G

ener

al s

hape

and

ove

rall

mag

nitu

de –

hig

her s

core

= g

reat

er o

vera

ll hi

p in

tern

al ro

tatio

n m

omen

t M

agni

tude

of p

eak

hip

exte

rnal

rota

tion

mom

ent –

hig

her s

core

= re

duce

d pe

ak h

ip e

xter

nal r

otat

ion

mom

ent

83.0

%

10.2

%

KFL

EX/E

XT M

om.

1 2 3

Gen

eral

sha

pe a

nd o

vera

ll m

agni

tude

– h

ighe

r sco

re =

gre

ater

ove

rall

knee

flex

ion

mom

ent

Phas

e sh

ift in

tim

ing

– hi

gher

scor

e =

dela

yed

peak

kne

e fle

xion

mom

ent

Mag

nitu

de o

f sag

ittal

pla

ne m

omen

t pea

ks –

hig

her s

core

= re

duce

d kn

ee fl

exio

n an

d ex

tens

ion

mom

ent p

eaks

60.3

%

29.4

%

4.9%

K

ABD

/AD

D M

om.

1 2 3

Mag

nitu

de o

f kne

e ad

duct

ion

mom

ent –

hig

her s

core

= re

duce

d kn

ee a

dduc

tion

mom

ent

Mag

nitu

de o

f pea

k kn

ee a

bduc

tion

mom

ent –

hig

her s

core

= g

reat

er p

eak

mom

ent

Tim

ing

and

mag

nitu

de o

f pea

k kn

ee a

bduc

tion

mom

ent –

hig

her s

core

= e

arlie

r and

gre

ater

pea

k

73.3

%

14.4

%

8.0%

K

ER/IR

Mom

. 1 2

Gen

eral

sha

pe a

nd o

vera

ll m

agni

tude

– h

ighe

r sco

re =

gre

ater

ove

rall

knee

inte

rnal

rota

tion

mom

ent

Mag

nitu

de o

f kne

e in

tern

al ro

tatio

n m

omen

t pea

k –

high

er s

core

= g

reat

er p

eak

mom

ent

80.6

%

14.9

%

HEX

T/FL

EX

Mom

. –

hip

exte

nsio

n/fle

xion

m

omen

t; H

ABD

/AD

D

Mom

. –

hip

abdu

ctio

n/ad

duct

ion

mom

ent;

HER

/IR

Mom

. –

hip

exte

rnal

/inte

rnal

ro

tatio

n m

omen

t; K

FLEX

/EX

T M

om.

– kn

ee e

xten

sion

/flex

ion

mom

ent;

KA

BD/A

DD M

om.

– kn

ee a

bduc

tion/

addu

ctio

n m

omen

t; K

ER/IR

Mom

. –

knee

ext

erna

l/int

erna

l ro

tatio

n m

omen

t; PP

– p

rinc

ipal

pat

tern

; ms

– m

illis

econ

ds; I

C –

initi

al c

onta

ct.

305


Tabl

e 9.

4 Su

mm

ary

of c

orre

latio

ns b

etw

een

sing

le-le

g sq

uat j

oint

rot

atio

n an

d le

ap la

ndin

g jo

int r

otat

ion

prin

cipa

l pat

tern

s

(PPs

). St

rong

sta

tistic

ally

sig

nifi

cant

(r >

0.5

0, p

< 0

.01)

cor

rela

tions

are

hig

hlig

hted

.

Le

ap L

andi

ng Jo

int R

otat

ion

PPs

H

FLEX

/EXT

H

AD

D/A

BD

HIR

/ER

KFL

EX/E

XT

KA

DD

/ABD

K

IR/E

R

PP

1 PP

2 PP

1 PP

2 PP

1 PP

2 PP

1 PP

1 PP

2 PP

3 PP

1 PP

2 PP

3

Single Leg Squat Joint Rotation PPs

HFL

EX/E

XT

PP1

0.1

9 0

.26

-0.0

1 -0

.32

0.0

9 0

.16

-0.1

2 -0

.44

0.2

4 0

.17

-0.0

4 0

.03

0.0

8 PP

2 0

.40

0.4

7 0

.02

-0.1

1 0

.42

0.1

0 -0

.53

0.0

2 0

.32

-0.1

9 0

.29

-0.1

6 -0

.09

HA

DD

/ABD

PP

1 0

.18

0.0

7 0

.34

0.2

8 -0

.14

-0.3

1 0

.22

-0.1

4 0

.08

0.2

5 -0

.11

0.0

9 0

.13

PP2

0.4

5 0

.22

0.3

0 -0

.07

0.2

0 -0

.16

-0.2

7 0

.01

0.0

1 0

.02

0.2

2 0

.06

0.0

9 PP

3 -0

.38

0.0

9 0

.13

-0.0

6 0

.04

-0.1

1 0

.31

-0.1

3 -0

.07

0.1

1 -0

.14

0.2

4 0

.37

HIR

/ER

PP1

-0.0

5 0

.31

-0.0

1 -0

.28

0.6

5 -0

.07

-0.3

3 -0

.01

0.0

5 -0

.18

0.0

7 0

.14

0.0

1 PP

2 0

.14

-0.0

8 0

.24

0.2

0 0

.12

-0.0

6 0

.06

-0.0

3 0

.08

0.1

4 0

.19

0.1

3 -0

.04

KFL

EX/E

XT

PP1

-0.3

5 -0

.20

0.0

0 0

.31

-0.2

9 -0

.01

0.6

5 0

.12

-0.0

1 0

.03

-0.3

7 0

.08

0.0

5 PP

2 0

.15

0.5

4 0

.13

-0.0

2 0

.51

0.0

4 -0

.26

-0.1

1 0

.46

-0.2

9 0

.12

0.0

8 0

.03

KA

DD

/ABD

PP

1 0

.26

0.1

9 -0

.21

0.0

9 0

.19

0.1

7 -0

.06

0.7

8 0

.10

0.0

3 0

.39

-0.5

3 -0

.09

PP2

-0.0

6 -0

.05

-0.0

7 -0

.08

-0.2

5 0

.05

-0.0

3 -0

.27

-0.1

8 0

.10

-0.2

1 -0

.10

0.0

5

KIR

/ER

PP1

0.2

0 0

.19

0.0

0 -0

.13

0.4

0 -0

.08

-0.3

6 0

.34

0.1

4 -0

.14

0.9

4 -0

.12

-0.0

7 PP

2 -0

.23

0.0

2 0

.28

-0.2

4 -0

.05

-0.4

4 -0

.18

-0.4

3 -0

.02

-0.0

3 0

.13

0.5

9 0

.03

AD

F/PF

PP

1 0

.27

-0.1

2 -0

.15

-0.2

3 0

.04

0.0

1 -0

.59

0.1

3 -0

.16

-0.0

7 0

.26

-0.1

6 -0

.24

PP2

0.2

6 0

.55

0.1

2 -0

.16

0.6

1 -0

.09

-0.4

4 -0

.23

0.2

9 -0

.23

0.2

3 0

.13

0.1

4 PP

1 0

.07

-0.0

3 0

.07

0.0

0 0

.04

-0.2

6 0

.07

-0.1

0 -0

.36

0.0

3 -0

.07

0.2

3 0

.15

AIN

V/E

VE

PP2

-0.2

8 -0

.07

0.0

7 0

.04

-0.3

0 0

.13

0.4

5 -0

.59

0.0

5 0

.15

-0.7

0 0

.15

0.2

0

AIR

/ER

PP1

-0.1

3 -0

.17

-0.0

3 -0

.12

-0.3

0 0

.03

0.0

5 -0

.47

-0.0

6 0

.08

-0.4

2 -0

.04

0.1

3 PP

2 0

.05

-0.1

2 -0

.13

-0.0

7 -0

.09

0.2

4 0

.01

0.0

9 0

.04

-0.0

6 -0

.25

-0.1

5 -0

.21

PP3

0.1

3 -0

.29

-0.2

9 0

.11

-0.2

3 0

.10

0.1

2 -0

.02

-0.1

4 0

.12

-0.1

3 -0

.30

-0.0

5 PP

4 0

.21

-0.2

5 -0

.06

0.0

2 -0

.09

0.1

6 -0

.26

0.1

3 0

.13

0.3

3 0

.05

-0.3

2 -0

.15

HFL

EX/E

XT

– hi

p fle

xion

/ext

ensi

on;

HA

DD

/ABD

– h

ip a

dduc

tion/

abdu

ctio

n; H

IR/E

R –

hip

inte

rnal

/ext

erna

l ro

tatio

n; K

EXT/

FLEX

– k

nee

flexi

on/e

xten

sion

; K

AD

D/A

BD –

kne

e ad

duct

ion/

abdu

ctio

n; K

IR/E

R –

kne

e in

tern

al/e

xter

nal r

otat

ion;

PP

– pr

inci

pal p

atte

rn.

306


Tabl

e 9.

5 Su

mm

ary

of c

orre

latio

ns b

etw

een

sing

le-le

g sq

uat j

oint

rota

tion

and

leap

land

ing

join

t mom

ent p

rinc

ipal

pat

tern

s

(PPs

). St

rong

sta

tistic

ally

sig

nifi

cant

(r >

0.5

0, p

< 0

.01)

cor

rela

tions

are

hig

hlig

hted

.

Le

ap L

andi

ng Jo

int M

omen

t PPs

H

EXT/

FLEX

Mom

. H

ABD

/AD

D M

om.

HER

/IR M

om.

KFL

EX/E

XT M

om.

KA

BD/A

DD

Mom

. K

ER/IR

Mom

.

PP

1 PP

2 PP

3 PP

1 PP

2 PP

3 PP

1 PP

2 PP

1 PP

2 PP

3 PP

1 PP

2 PP

3 PP

1 PP

2

Single Leg Squat Joint Rotation PPs

HFL

EX/E

XT

PP1

0.0

6 -0

.06

0.0

4 0

.24

-0.1

1 -0

.21

-0.2

8 0

.04

0.0

1 0

.27

0.3

1 0

.42

-0.1

3 -0

.15

0.0

8 0

.05

PP2

-0.0

1 0

.26

-0.0

7 0

.06

0.0

0 0

.29

-0.0

8 -0

.04

0.4

5 -0

.16

-0.2

0 -0

.17

-0.0

9 0

.41

-0.2

1 0

.15

HA

DD

/ABD

PP

1 -0

.29

0.0

3 0

.08

-0.1

2 0

.17

-0.2

0 0

.00

-0.2

2 -0

.26

0.0

8 0

.05

0.1

3 0

.01

-0.1

5 -0

.03

-0.1

9 PP

2 -0

.14

0.0

0 0

.16

-0.1

7 0

.03

-0.0

7 -0

.04

-0.1

0 0

.13

0.0

8 -0

.24

-0.0

3 -0

.04

0.0

6 -0

.29

-0.2

3 PP

3 0

.21

0.0

1 -0

.03

-0.0

5 -0

.18

-0.2

9 -0

.14

-0.0

2 -0

.23

0.0

2 0

.36

0.0

8 -0

.39

-0.2

0 -0

.13

-0.4

5

HIR

/ER

PP1

0.1

4 0

.33

-0.1

2 0

.06

-0.1

7 0

.26

0.0

0 0

.33

0.4

4 -0

.17

0.3

0 -0

.10

-0.0

3 0

.09

0.0

7 0

.42

PP2

-0.0

1 -0

.16

-0.0

7 -0

.11

0.0

8 0

.09

0.0

0 -0

.11

-0.0

8 -0

.07

-0.1

9 -0

.10

0.1

2 -0

.02

-0.0

2 -0

.10

KFL

EX/E

XT

PP1

-0.1

0 -0

.15

-0.2

0 -0

.14

0.1

3 -0

.13

0.1

3 -0

.14

-0.3

8 -0

.04

0.0

3 0

.04

-0.1

0 -0

.01

0.2

4 -0

.27

PP2

-0.0

1 0

.10

-0.2

7 0

.10

-0.0

3 0

.08

-0.1

6 -0

.10

0.3

0 -0

.17

-0.1

8 -0

.04

-0.2

9 0

.49

-0.0

4 0

.01

KA

DD

/ABD

PP

1 -0

.33

0.2

1 -0

.17

0.0

9 0

.30

0.0

3 0

.09

-0.1

3 0

.11

-0.2

0 -0

.05

-0.4

9 0

.34

0.2

1 -0

.11

0.3

4 PP

2 -0

.03

0.1

1 0

.14

0.0

7 0

.05

-0.2

8 0

.05

0.0

7 -0

.02

0.3

9 0

.37

0.2

3 -0

.10

-0.1

3 0

.18

0.0

4

KIR

/ER

PP1

-0.1

2 0

.08

0.0

4 0

.09

-0.1

4 -0

.24

-0.1

3 -0

.03

0.0

7 -0

.22

0.0

2 -0

.60

0.1

7 0

.06

-0.3

9 0

.35

PP2

0.2

5 -0

.26

0.0

6 -0

.14

-0.5

1 -0

.26

-0.1

3 0

.08

-0.0

5 0

.23

0.1

8 0

.14

-0.2

7 -0

.55

-0.2

1 -0

.15

AD

F/PF

PP

1 0

.13

0.1

7 0

.18

-0.0

7 -0

.02

0.3

3 0

.16

0.1

2 0

.49

-0.0

2 -0

.20

-0.2

1 0

.28

-0.0

2 -0

.28

0.3

3 PP

2 -0

.07

0.2

0 -0

.08

0.1

6 -0

.19

-0.1

1 -0

.28

0.0

4 0

.27

-0.2

2 -0

.03

0.0

5 -0

.18

0.2

6 -0

.21

0.1

6 PP

1 -0

.13

0.1

5 0

.25

0.0

5 -0

.27

-0.1

3 -0

.24

0.1

7 -0

.32

-0.2

6 0

.28

0.2

5 0

.17

-0.3

9 -0

.14

0.0

3 A

INV

/EV

E PP

2 0

.17

-0.1

9 0

.00

0.0

1 -0

.05

-0.0

2 0

.00

0.1

3 -0

.19

0.3

9 0

.17

0.5

8 -0

.48

-0.1

0 0

.41

-0.4

4

AIR

/ER

PP1

0.1

7 -0

.09

0.0

7 0

.12

-0.1

2 -0

.24

-0.1

2 0

.04

-0.1

6 0

.12

0.0

6 0

.38

-0.2

9 -0

.03

0.0

6 -0

.26

PP2

0.0

0 -0

.02

-0.0

3 -0

.05

0.1

5 0

.31

0.2

7 0

.17

0.2

0 0

.11

-0.1

0 0

.00

0.1

4 0

.12

0.1

2 0

.22

PP3

-0.1

7 0

.46

0.0

9 0

.06

0.3

1 -0

.22

0.0

8 -0

.29

-0.1

0 -0

.25

0.2

7 0

.04

0.3

4 0

.10

0.0

2 0

.14

PP4

0.1

1 0

.00

0.0

2 0

.01

0.3

6 0

.19

0.2

1 0

.06

0.3

7 0

.31

-0.1

2 -0

.10

0.1

5 0

.04

0.1

5 0

.22

HEX

T/FL

EX M

om. –

hip

ext

ensi

on/fl

exio

n m

omen

t; H

ABD

/AD

D M

om. –

hip

abd

uctio

n/ad

duct

ion

mom

ent;

HER

/IR M

om. –

hip

ext

erna

l/int

erna

l rot

atio

n m

omen

t; K

FLEX

/EX

T M

om. –

kne

e ex

tens

ion/

flexi

on m

omen

t; K

ABD

/AD

D M

om. –

kne

e ab

duct

ion/

addu

ctio

n m

omen

t; K

ER/IR

Mom

. – k

nee

exte

rnal

/inte

rnal

rota

tion

mom

ent;

PP –

pri

ncip

al p

atte

rn.


307

9.4 Discussion

Identifying athletic tasks that are linked to sporting manoeuvres associated

with ACL injuries may reveal more optimal movements for ACL injury risk

screening within sport-specific contexts. The purpose of this study was to

determine the relationship between the performance of a single-leg squat and

a netball-specific leap landing task, in order to examine the utility of including

a single-leg squat movement within netball-specific ACL injury risk screening

protocols. A number of associations were found between biomechanical

patterns from the single-leg squat and leap landing task. Biomechanical

patterns from the single-leg squat were linked to a number of biomechanical

factors that have been proposed to increase ACL loads or injury risk during

landing movements. These findings suggest that a single-leg squat may be a

useful screening movement for identifying athletes’ who utilise movement

strategies associated with ACL loading or injury risk during high-risk netball-

specific landing manoeuvres.

The strongest associations between the single-leg squat and leap landing tasks

were evident in matching joint rotation patterns, particularly at the knee.

Participants who performed the single-leg squat with greater hip internal

rotation, knee flexion, knee abduction and knee internal rotation exhibited

these same biomechanical characteristics during the leap landing task. These

findings are similar to previous studies that have found a relationship between

single-leg squat and landing or cutting tasks, where similar biomechanical

characteristics were shared across the two movements.[439, 440] A limitation of

these previous studies is that they have only examined relationships between

corresponding biomechanical variables within the same joints and planes of

movement across the two tasks. In addition to examining associations between


308

corresponding biomechanical patterns, the present study identified further

relationships outside of the corresponding joint and/or movement plane. For

example, sagittal plane biomechanical patterns at the hip and ankle during the

single-leg squat were related to the magnitude of knee flexion during the leap

landing task. Further, sagittal plane movement strategies at the knee during

the single-leg squat were associated with sagittal plane hip motion during the

leap landing task. These results, combined with the direct relationship

between the magnitude of knee flexion between the two tasks, suggests the

overall sagittal plane movement strategies of the lower limb between the

single-leg squat and leap landing task are closely related. Cross-planar

relationships were also apparent between the two tasks, whereby frontal and

transverse plane knee motion during the single-leg squat was related to

transverse and frontal plane knee biomechanics, respectively, during the leap

landing task. Similarly, the magnitude of frontal plane ankle motion during

the single-leg squat was related to frontal and transverse plane biomechanical

patterns at the knee during the leap landing task.

The large number of associations observed between the two tasks likely stems

from the similarities in lower limb positioning. The hip and knee motion

experienced during the initial deceleration phase of netball leap landings,

where unilateral weight acceptance is required, closely resemble the lower

limb positioning experienced during the descent phase of a single-leg squat.

Due to the similarities between the tasks, biomechanical deficits that have the

potential to increase ACL injury risk could manifest across both movements.

Biomechanical patterns from the single-leg squat task were associated with a

number of factors linked to increased ACL loading and injury risk during the

leap landing task. ‘Dynamic valgus’ during landing has been identified as a

significant predictor of ACL injuries in female athletes,[160] and a valgus


309

collapse is often seen when ACL injuries occur.[97-99] Internal rotation at the hip,

combined with knee abduction and knee internal rotation is thought to

contribute to a ‘dynamic valgus’ position.[160, 162] Considering these

biomechanical patterns were directly related between the two tasks examined,

an individual displaying a ‘dynamic valgus’ position during the single-leg

squat may also do so during leap landing performances. Reduced knee flexion

has been linked to increases in ACL loads[20, 24, 25] and is also frequently

observed during ACL injuries.[97, 99] Earlier peak hip flexion, and reduced

overall knee flexion and ankle dorsiflexion during the single-leg squat were

all associated with reduced knee flexion during the leap landing task. The

presence of this sagittal plane movement strategy during a single-leg squat

may therefore be a useful indicator of individuals’ who perform leap landings

with reduced knee flexion. The peak knee abduction moment experienced

during athletic tasks is often acknowledged as a primary risk factor for ACL

injury due to larger peak moments being linked prospectively to future ACL

injuries[160] and higher ACL loads.[20-23, 28] Greater ankle eversion during the

single-leg squat tended to be associated with earlier and larger peak knee

abduction moments during landing, thus the degree of frontal plane ankle

motion during a single-leg squat may be a useful indicator of ACL injury risk

during leap landing performances.

The single-leg squat appears to be an appropriate movement screen for ACL

injury risk in the netball-specific context examined. Hip internal rotation; knee

flexion, abduction and internal rotation; and ankle eversion during the single-

leg squat were all linked to biomechanical factors associated with ACL loading

or injury risk during the leap landing task. The development of visual rating

criteria that accurately assess or identify these biomechanical patterns during

a single-leg squat may provide a simplistic method for identifying athletes


310

who employ high-risk movement strategies during leap landing

performances. These criteria could subsequently be used within a movement

screening protocol to identify netball players at potentially greater risk of ACL

injury.

Although a number of associations between single-leg squat and leap landing

performance were identified, these were obtained using laboratory-based

techniques (i.e. motion capture). Three-dimensional measures were used in

the identification of lower limb biomechanical patterns during the single-leg

squat, and as mentioned throughout this thesis, these methods are not

applicable to the majority of the wider community. Field-based injury risk

screening methods must function without the need for these laboratory-based

techniques. Visual rating using video-based or live observational methods are

commonly used within field-based movement assessments,[85, 86, 91, 410] as the

equipment (or lack thereof) required for these methods is minimal and often

readily available. Visual rating of single-leg squat performance must be

reliable and comparative to three-dimensional measures to ensure it is

appropriate for use within a netball-specific field-based ACL injury risk

screening method. Studies investigating visual rating of single-leg squat

performance have found encouraging results for reliability across a range of

different grading criteria.[422, 423, 434, 437] High levels of intra- and inter-rater

agreement have been reported for ratings of medio-lateral knee motion (i.e.

knee position medial to the toe),[422, 423, 437] frontal plane knee angle and

alignment,[400, 423, 436] and general overall movement quality[434, 437] during single-

leg squat performance. It appears that single-leg squat performance can be

reliably assessed from visual observation, which provides promise for its

incorporation within field-based ACL injury risk screening protocols.

However, it is also important that ratings obtained from visual observation are


311

valid against laboratory-based measures to ensure they are accurate in

measuring the construct of interest.

The majority of studies validating visual rating criteria against three-

dimensional measures during single-leg squat performance have focused on

frontal and transverse plane motion,[422, 423, 435, 442] with mixed findings across the

literature. Ageberg et al.[422] investigated the validity of observational rating of

medio-lateral knee motion during single-leg squat performance. Raters

graded whether or not participants displayed a knee position medial to their

toe during the single-leg squat. When compared against three-dimensional

measures, the hip was more internally rotated in those with a knee position

medial to the foot compared to those with a knee-over-foot position.[422]

However, no differences in three-dimensional knee abduction/adduction

angles were identified between the knee-medial-to-foot and knee-over-toe

groups.[422] Similarly, Whatman et al.[423] found a knee position medial to the

toe as deemed by video observation was indicative of greater hip adduction

and internal rotation, but not knee abduction when compared to three-

dimensional measures. In contrast, Mauntel et al.[442] found participants who

displayed a knee position medial to the toe determined from visual

observation during a single-leg squat demonstrated greater three-dimensional

knee abduction, with no differences in hip adduction or internal rotation.

Similar definitions for medial knee motion were used across these studies,

however different protocols relating to the single-leg squat task were

apparent. Differences in the control of squat depth and pace, and the number

of consecutive squat repetitions were noted between studies.[422, 423, 442] Careful

consideration surrounding the single-leg squat movement protocols may be

required if using visually observed medial knee motion as an indication of

three-dimensional hip and knee motion in the frontal and transverse planes.


312

The frontal plane projection angle (FPPA) at the knee (i.e. the angle formed at

the intersection of the thigh and shank segments)[443] has also been investigated

in visual rating of single-leg squat performance.[400, 435, 436] Willson and Davis[435]

found that the FPPA was associated with three-dimensional measures of hip

adduction, hip internal rotation and knee external rotation, but not frontal

plane knee angle during single-leg squat performance. Inconsistencies exist

across the literature with regard to the validity of visual ratings of single-leg

squat performance comparative to three-dimensional measures. Movement

screening protocols that incorporate visual assessment of a single-leg squat

must therefore carefully develop and validate the criteria used to grade

performance.

The use of only one successful single-leg squat trial for analysis in this study

is an important factor to consider. Studies examining the validity of single-leg

squat movements for screening purposes have often used multiple squat

repetitions or movement trials for performance measures.[422, 434, 435] The

purpose of this study was to investigate the potential of the single-leg squat to

be incorporated into field-based ACL injury risk screening protocols. Hence,

it was deemed important that the screening movement be simple and involve

minimal time to complete. A single repetition of the single-leg squat met these

criteria and was subsequently used. It is possible that an individual’s single-

leg squat performance would vary over multiple repetitions or trials, hence

these additions may have captured further information regarding their

performance of the task. However, the numerous strong associations observed

in this study suggests a single repetition of the single-leg squat may provide

sufficient information on an individual’s ability to control the lower limb

during a high-risk netball-specific landing task. Further work investigating the

relationship between simplistic movements and high-risk sporting tasks may


313

wish to include additional trials of the movement screening task. The added

time costs surrounding collection and analysis of these extra trials within

screening protocols must, however, also be considered.

9.5 Conclusions

This study found numerous strong associations for lower limb biomechanical

patterns between a single-leg squat and netball-specific leap landing task.

Lower limb motion during the single-leg squat was linked to a number of

biomechanical factors that have been proposed to increase ACL loads or injury

risk during landing movements. Specifically, hip internal rotation; knee

flexion, abduction and internal rotation; and ankle eversion during the single-

leg squat were linked to high-risk movement strategies during the leap

landing task. This study has provided the first step in the development of a

netball-specific ACL injury risk screening method incorporating a single-leg

squat, by establishing the key biomechanical patterns that should be observed

and rated due to their links to a high-risk landing task within the sport.

Validation of a set of visual rating criterion relating to the key biomechanical

patterns from the single-leg squat identified in this study will ensure they are

related to the three-dimensional measures they are targeting. These rating

criterion may then be evaluated against three-dimensional measures from the

leap landing task to ensure visual observation of the single-leg squat

movement is also related to leap landing performance. Further work linking

single-leg squat movement strategies to ACL injury occurrences may also

provide insight on the ability of a single-leg squat movement screen to

prospectively identify athletes’ at-risk of ACL injury.

314

Chapter Ten: General Discussion

10.1 Summary

This thesis aimed to explore three inter-related areas relating to anterior

cruciate ligament (ACL) injury prevention. The concepts of neuromuscular

anomalies and muscle activation profiles during the performance of a high-

risk landing task were investigated. In addition, the effect of an ACL injury

prevention program on individual neuromuscular and biomechanical

responses was explored. Lastly, examination of existing field-based ACL

injury risk screening methods and a proposed sport-specific screening

movement was undertaken to determine their applicability for use in wider

community and sport-specific settings. The experimental and review studies

undertaken have produced several key and novel findings that can contribute

to ACL injury prevention practices.

Study one (Chapter Four) examined the concept of intra-individual anomalies

in muscle activation profiles during a high-risk sporting task. Anomalies in

muscle activation profiles were found to occur during repeat performances of

the high-risk landing task examined. However, in contrast to the hypothesis,

these neuromuscular anomalies did not induce changes to lower limb

biomechanics consistent with increased ACL loads or injury risk. While the

results of this study support the theory that consistent task performance can

be achieved with anomalies in muscle activation profiles, it may be unwise to

dismiss the potential role of neuromuscular anomalies in ACL injuries.

Investigating the differential effects of specific neuromuscular anomalies may

elicit further understanding regarding their influence on ACL injury risk.

General Discussion

315

Findings from study one also demonstrated that some individuals utilised

multiple typical or ‘normal’ muscle activation profiles during the high-risk

landing task. Subsequently, study two (Chapter Five) explored this concept

and found participants’ utilised different muscle activation profiles across

repeat performance of the high-risk landing task. While these different profiles

were identified, individuals tended to persevere with a preferred profile

throughout the majority of landing trials. This study failed to identify ‘safe’ or

‘high-risk’ muscle activation profiles. Rather, specific muscle activation

characteristics or combinations of activation characteristics within profiles

were linked to the presence of biomechanical deficits associated with ACL

injury risk. Tailored injury prevention programs targeting these specific

muscle activation characteristics may induce more beneficial neuromuscular


Study three (Chapter Six) examined the effects of an ACL injury prevention

program at the group and individual level. The injury prevention program

induced neuromuscular and biomechanical adaptations at the group level.

However, examination of individual data revealed dissimilar responses to the

training program within the training group. Individuals who display a pre-

existing high-risk neuromuscular or biomechanical strategy may receive

greater benefits from the ACL injury prevention program only if the program

targets the relevant high-risk strategy. Injury prevention programs targeting

the specific neuromuscular or biomechanical deficits of the individual may

therefore be optimal for reducing ACL injury risk.

Study four (Chapter Seven) provided an extensive review of current field-

based screening methods for ACL injury risk and evaluated their applicability

for use in wider community settings. There was limited evidence supporting

the predictive validity of field-based screening methods for identifying future

General Discussion

316

ACL injuries. Furthermore, the capacity of screening methods to identify

athletes at-risk of ACL injury may vary across different sports or sporting

populations. Differences between screening methods surrounding the

equipment required, time taken to screen athletes, and applicability across

varying sports and athletes were identified. It was concluded that these are

important factors in determining the most appropriate screening method to

use in various situations. The Landing Error Scoring System and Tuck Jump

Assessment were deemed as two field-based screening methods that were

applicable for use in wider community settings, and were subsequently

examined in the following study.

Study five (Chapter Eight) examined the efficacy of the Landing Error Scoring

System (LESS) and Tuck Jump Assessment screening methods in identifying

lower limb biomechanics consistent with increased ACL loading or injury risk

during a high-risk sport-specific task, namely, a netball leap landing. Minimal

relationships between lower limb biomechanics during the leap landing and

scores received from the LESS and Tuck Jump Assessment were identified. It

was concluded that the LESS and Tuck Jump Assessment are unlikely to be

effective in identifying ACL injury risk in the netball-specific context

examined. It was proposed the bilateral nature of the movements used within

these screening methods were a contributing factor to these results. Due to

this, the final study of this thesis explored the potential for including a

unilateral movement within netball-specific ACL injury risk screening

protocols.

Study six (Chapter Nine) investigated the relationships between performance

of a single-leg squat and a netball-specific leap landing task that has been

associated with ACL injuries. Lower limb motion during the single-leg squat

was linked to a number of biomechanical factors that have been proposed to

General Discussion

317

increase ACL loads or injury risk during landing movements. These findings

highlighted the potential of a single-leg squat movement to be incorporated

within netball-specific ACL injury risk screening protocols.

When considering these findings collectively, there are several overarching

topics that warrant further discussion. These will be highlighted, along with

recommendations for future research, in the following sections.

10.2 Neuromuscular Contributions to ACL Injury Risk

Many studies have attempted to link neuromuscular contributions to lower

limb biomechanics associated with ACL loading and injury

risk.[44, 56, 131, 134, 171, 174-181] Despite the variable nature of neuromuscular control

during biomechanical tasks,[48-50, 254] there has been very little focus on the

impact of trial-to-trial differences in muscle activation characteristics on ACL

injury risk. The findings of this thesis further the notion that neuromuscular

control can vary from one performance to the next. Anomalies in

neuromuscular control were found across repeat trials of the high-risk landing

task examined, while the use of different muscle activation profiles across

trials was also identified. Future work examining neuromuscular

contributions to ACL injury risk may benefit from a greater focus on these

specific trial-to-trial differences. The ‘one-off’ nature with which ACL injuries

occur suggests a random neuromuscular dysfunction may play a role in these

injury occurrences. To truly understand these injuries, we need to identify the

underlying causes that lead to these isolated events. Further examination of

changes in neuromuscular control from a trial-to-trial basis, and linking these

with lower limb postures associated with increased ACL loads and injury risk

may assist in achieving this. Musculoskeletal modelling and computer

simulation methods are a viable approach for further examination of

General Discussion

318

neuromuscular contributions to ACL injury risk. These methods have been

used to identify how alterations in muscle activation and forces can impact

human movement,[379-381] calculate ACL loads during high-risk sporting

tasks,[419, 444] and identify potential mechanisms of ligament rupture via injury

simulations.[176, 309] Together, these methods may be used to investigate how

specific alterations in neuromuscular control contribute to ACL loading and

potentially cause ACL injury.

10.3 ACL Injury Prevention Programs

When combined with previous work, a number of findings across this thesis

further the argument that a generic universal approach to ACL injury

prevention may not be optimal. The level of adoption and compliance of

current ACL injury prevention programs is low[227, 228] and their deployment in

“real-world” situations may not be effective.[223, 229-231] Further to these practical

issues, generic ACL injury prevention programs may not be the most effective

approach in providing a benefit to all involved.

This thesis found that a number of different muscle activation profiles can be

utilised by individuals during performance of a high-risk sporting task. While

certain individuals showed a tendency to vary the activation profile

employed, the majority tended to persevere with a preferred profile

throughout repeat landing trials. Specific muscle activation characteristics

within each of these profiles were linked to biomechanical deficits associated

with ACL loading and injury risk. Therefore, a reduction in ACL injury risk

may be achieved with injury prevention programs that target these potentially

high-risk activation characteristics. Injury prevention programs tailored to the

relevant high-risk muscle activation characteristics of the individual’s

preferred profile may be optimal for inducing beneficial neuromuscular

General Discussion

319

adaptations and reducing ACL injury risk. In addition, these results support

the notion of previous studies[71, 82, 83] that variable individual responses to ACL

injury prevention programs are to be expected. Within this thesis, inconsistent

neuromuscular and biomechanical adaptations to the training program were

apparent across individuals. However, those who exhibited a pre-existing

high-risk movement strategy (i.e. large peak knee abduction moments)

received the largest benefits in reducing frontal plane knee moments during

landing. It was proposed that this stemmed from the injury prevention

program specifically targeting movement strategies associated with the pre-

existing risk pattern of large peak knee abduction moments. This provides

further evidence that tailored injury prevention programs targeting the high-

risk neuromuscular or biomechanical strategies of the individual may be the

most effective in reducing ACL injury risk.

While evidence points towards more beneficial outcomes with tailored ACL

injury prevention programs, the cost-effectiveness of this approach must be

considered. Highly individualised ACL injury prevention programs are

unlikely to be feasible from an implementation and economical perspective,

particularly within wider community settings where time and staff resources

may be limited. A sub-group approach to tailoring injury prevention

programs, whereby the training targets the high-risk characteristics relevant

to the individuals within the sub-group, has been suggested as a more effective

and efficient option over generic approaches.[358] Studies directly comparing

generic versus tailored injury prevention programs may determine whether

or not this approach is a better option. Furthermore; while multiple

studies[71, 82, 83] have shown that individual responses to injury prevention

programs differ, there is little understanding as to why this is the case. The

presence of a pre-existing high-risk neuromuscular or biomechanical strategy

General Discussion

320

may have a modulating effect on the response, whereby greater benefits are

received in the area these high-risk strategies exist.[71, 83, 84] Apart from this,

there is a limited understanding of what dictates the adaptations an individual

will exhibit following participation in an ACL injury prevention program.

There is large scope for future research investigating potential factors that

contribute to this phenomenon. Future work in these areas will aid in

improving the design of ACL injury prevention programs that provide a

benefit for all involved.

10.4 Field-Based Screening Methods for ACL Injury Risk

The results of this thesis emphasise the need for sport-specific ACL injury risk

screening methods. A key finding of the systematic review of field-based ACL

injury risk screening methods was their limited predictive validity in

identifying ACL injuries. While certain work has shown field-based screening

methods can identify athletes at-risk of future ACL injury, this appears to vary

across different sports or sporting populations.[104, 105] Furthermore; the two

field-based screening methods examined in this thesis, the Landing Error

Scoring System (LESS) and Tuck Jump Assessment, were unable to identify

high-risk movement strategies during a sport-specific task associated with

ACL injuries. It is important to note that only one sport-specific task, a netball

leap landing, was examined. The LESS and Tuck Jump Assessment screening

methods should not be disregarded based on these findings. While they

appear ineffective in the netball-specific context examined, these methods may

be applicable to other sports or sporting populations. Future studies

examining the capacity of field-based screening methods in identifying high-

risk movement strategies across a range of high-risk sport-specific tasks may

aid in determining the optimal method(s) across different sports or sporting

General Discussion

321

populations. The associations between the single-leg squat and netball leap

landing tasks found within this thesis also suggest the investigation and

development of sport-specific ACL injury risk screening methods is justified.

The generic bilateral movements used in current field-based ACL injury risk

screening methods[85, 87, 90, 91] have minimal relationships to lower limb

biomechanics experienced during sport-specific manoeuvres associated with

ACL injury.[95, 96] Simplistic movements that can be easily observed and reflect

the lower limb postures experienced during high-risk sport-specific tasks may

be the most viable options for investigation and inclusion with field-based

screening methods for ACL injury risk. Identifying movements that better

replicate sport-specific movements associated with ACL injuries may further

aid in the development of appropriate ACL injury risk screening methods.

10.5 Limitations

There are a number of limitations to this thesis that must be acknowledged

and considered when interpreting the findings. The cross-sectional nature of

studies may have been a limitation in identifying neuromuscular or

biomechanical factors that contributed to ACL injury risk. Prospective studies

provide the strongest evidence in detecting neuromuscular or biomechanical

factors that contribute to ACL injury risk[160, 177] or evaluating the capacity of

screening methods in identifying athletes at-risk of ACL injury.[104, 105, 412]

However, prospective studies investigating ACL injury risk require a large

cohort to achieve meaningful outcomes. The capacity to recruit such a cohort

was deemed to be outside the scope of this thesis. Prospective studies

investigating ACL injury risk often recruit upwards of 200 participants in

order to attain meaningful results.[105, 160, 412] Due to the limited participant pool

in the population examined (i.e. local community netball), these participant

General Discussion

322

numbers could not be achieved for these studies. Lower limb biomechanical

factors that have been linked to ACL loading were used as a surrogate measure

for identifying potential ACL injury risk across the studies in this thesis. While

this approach is commonly used for investigating potential neuromuscular

and biomechanical risk factors for ACL injury,[44, 56, 131, 134, 171, 174-181] it is important

to acknowledge this as a limitation. Prospective longitudinal studies are

needed to confirm current findings and further understand the underlying

neuromuscular and biomechanical risk factors for ACL injury. Multicentre

investigations may improve the feasibility of these studies by providing access

to larger participant cohorts, leading to more robust findings.[445]

Neuromuscular and biomechanical data from the trunk were not collected as

part of the studies in this thesis. ACL injuries often occur with a lateral and

extended trunk position,[159, 377] and trunk rotation and lateral flexion have been

shown to influence frontal and transverse plane knee moments during landing

and cutting tasks.[166, 167, 255, 372] These recent findings have confirmed that trunk

control plays an important role in ACL injury risk. Despite this, studies

investigating the biomechanics of netball leap landings with regard to ACL

injury risk have not included a focus on the trunk.[245, 248, 253] The leap landing

task performed within the studies in this thesis involved a straight running

approach with a pass received directly out in front of the body. Hence, it was

thought that minimal frontal and/or transverse plane motion of the trunk

would occur during landing performances and the inclusion of trunk-related

factors would not substantially improve the outcomes associated with this

research. Nonetheless, the inclusion of trunk-related factors could have

strengthened certain findings within this thesis. For example, activation

characteristics of trunk muscles may be an important aspect of the overall

muscle activation profile that contributes to ACL injury risk. In addition, the

General Discussion

323

efficacy of the screening methods examined in this thesis may have improved

with the inclusion of trunk biomechanics from the leap landing task. The LESS

includes criteria related to trunk motion during landing,[91] while certain

criteria within the Tuck Jump Assessment are purportedly linked to a ‘trunk-

dominant’ movement strategy proposed to increase ACL injury risk.[413] As

these methods consider trunk-related factors in their evaluation of ACL injury

risk, they could have an association to three-dimensional measures of trunk

biomechanics during high-risk landing tasks. Further relationships between

the single-leg squat and netball leap landing tasks may have also been

identified with the inclusion of trunk biomechanics from the single-leg squat.

If this were the case, trunk movement may be an additional component to

consider when evaluating single-leg squat performance in the context of a

netball-specific ACL injury risk screening method. Future investigations

pertaining to ACL injury risk or prevention should consider inclusion of

neuromuscular control and biomechanics of the trunk in addition to the lower

limb. This is particularly important for tasks where frontal and transverse

plane trunk movements are more prominent, such as side-step cutting or

overhead catch-and-land tasks,[166, 255, 372, 446] and may subsequently impact

loading at the knee.

It is also important to acknowledge that this study focused on one population

(i.e. female netball athletes) and sport-specific manoeuvre associated with

ACL injuries (i.e. a netball leap landing). The findings from this thesis,

therefore, are most applicable to ACL injury risk and prevention in the netball-

specific context examined. Applying similar methods to those used in this

thesis to additional high-risk populations or sporting tasks may further our

understanding of ACL injury risk and prevention across a range of settings.

Examining whether neuromuscular anomalies and multiple muscle activation

General Discussion

324

profiles occur in different populations and sporting tasks, and if so, how they

influence lower limb biomechanics may yield greater knowledge surrounding

neuromuscular contributions to ACL injury risk. It may be the case that

different neuromuscular risk factors present across different high-risk

populations or sporting tasks. If so, ACL injury prevention programs specific

to these settings may be required to effectively target neuromuscular control

and reduce ACL injury risk. Examining the efficacy of ACL injury risk

screening methods in different sports or sporting populations would also aid

in identifying the most applicable method(s) across these variable settings.

10.6 Conclusions

The findings of this thesis highlight the need for tailored ACL injury

prevention programs targeting neuromuscular and biomechanical deficits that

are relevant to the individual. It is apparent that the neuromuscular system

employs multiple muscle activation profiles to perform the same landing task,

and the neuromuscular and biomechanical adaptations to an ACL injury

prevention program vary across individuals. Importantly, it appears that

high-risk movement strategies can be targeted with appropriately designed

ACL injury prevention programs.

Identifying screening methods that are capable of detecting those at-risk of

injury during sporting tasks known to cause ACL injuries is a vital step in

improving the design of targeted ACL injury prevention programs. The single-

leg squat was identified as a potentially useful movement to be incorporated

within netball-specific ACL injury risk screening protocols. The single-leg

squat appears to be more effective in identifying high-risk lower limb

biomechanics during a netball-specific landing task than existing ACL injury

risk screening methods. Further examination of simplistic unilateral

General Discussion

325

movements against other sport-specific tasks associated with ACL injury may

aid in ascertaining the optimal method(s) for ACL injury risk screening across

different sports or sporting populations.

326

References

1. Hootman JM, Dick R, Agel J. Epidemiology of collegiate injuries for 15 sports: Summary and recommendations for injury prevention initiatives. J Athl Train. 2007; 42(2): 311-9.

2. Agel J, Arendt EA, Bershadsky B. Anterior cruciate ligament injury in National Collegiate Athletic Association basketball and soccer: A 13-year review. Am J Sports Med. 2005; 33(4): 524-30.

3. Arendt E, Dick R. Knee injury patterns among men and women in collegiate basketball and soccer: NCAA data and review of literature. Am J Sports Med. 1995; 23(6): 694-701.

4. Borowski LA, Yard EE, Fields SK, Comstock RD. The epidemiology of US high school basketball injuries, 2005-2007. Am J Sports Med. 2008; 36(12): 2328-35.

5. Myklebust G, Maehlum S, Holm I, Bahr R. A prospective cohort study of anterior cruciate ligament injuries in elite Norwegian team handball. Scand J Med Sci Sports. 1998; 8(3): 149-53.

6. Powell JW, Barber-Foss KD. Sex-related injury patterns among selected high school sports. Am J Sports Med. 2000; 28(3): 385-91.

7. Waldén M, Hägglund M, Magnusson H, Ekstrand J. Anterior cruciate ligament injury in elite football: A prospective three-cohort study. Knee Surg Sports Traumatol Arthrosc. 2011; 19(1): 11-9.

8. Bjordal JM, Arnoy F, Hannestad B, Strand T. Epidemiology of anterior cruciate ligament in soccer. Am J Sports Med. 1997; 25(3): 341-5.

9. Roos H, Ornell M, Gardsell P, Lohmander LS, Lindstrand A. Soccer after anterior cruciate ligament injury - An incompatible combination? A national survey of incidence and risk factors and a 7-year follow-up of 310 players. Acta Orthop Scand. 1995; 66(2): 107-12.

10. Kvist J. Rehabilitation following anterior cruciate ligament injury: Current recommendations for sports participation. Sports Med. 2004; 34(4): 269-80.

11. Myklebust G, Bahr R. Return to play guidelines after anterior cruciate ligament surgery. Br J Sports Med. 2005; 39(3): 127-31.

References

327

12. Swärd P, Kostogiannis I, Roos H. Risk factors for a contralateral anterior cruciate ligament injury. Knee Surg Sports Traumatol Arthrosc. 2010; 18(3): 277-91.

13. Wright RW, Magnussen RA, Dunn WR, Spindler KP. Ipsilateral graft and contralateral ACL rupture at five years or more following ACL reconstruction: A systematic review. J Bone Joint Surg. 2011; 93(12): 1159-65.

14. Paterno MV, Rauh MJ, Schmitt LC, Ford KR, Hewett TE. Incidence of contralateral and ipsilateral anterior cruciate ligament (ACL) injury after primary ACL reconstruction and return to sport. Clin J Sport Med. 2012; 22(2): 116-21.

15. Hewett T, Myer G, Ford K, Paterno M, Quatman C. The 2012 ABJS Nicolas Andry Award: The sequence of prevention: A systematic approach to prevent anterior cruciate ligament injury. Clin Orthop Relat Res. 2012; 470(10): 1-11.

16. Lohmander LS, Englund PM, Dahl LL, Roos EM. The long-term consequence of anterior cruciate ligament and meniscus injuries. Am J Sports Med. 2007; 35(10): 1756-69.

17. Lohmander LS, Östenberg A, Englund M, Roos H. High prevalence of knee osteoarthritis, pain, and functional limitations in female soccer players twelve years after anterior cruciate ligament injury. Arthritis Rheum. 2004; 50(10): 3145-52.

18. Øiestad BE, Holm I, Engebretsen L, Risberg MA. The association between radiographic knee osteoarthritis and knee symptoms, function and quality of life 10–15 years after anterior cruciate ligament reconstruction. Br J Sports Med. 2011; 45(7): 583-8.

19. Claes S, Hermie L, Verdonk R, Bellemans J, Verdonk P. Is osteoarthritis an inevitable consequence of anterior cruciate ligament reconstruction? A meta-analysis. Knee Surg Sports Traumatol Arthrosc. 2013; 21(9): 1967-76.

20. Markolf KL, Burchfield DM, Shapiro MM, Shephard MF, Finerman GAM, Slauterback JL. Combined knee loading states that generate high anterior cruciate ligament forces. J Orthop Res. 1995; 13(6): 930-5.

21. Oh YK, Lipps DB, Ashton-Miller JA, Wojtys EM. What strains the anterior cruciate ligament during a pivot landing? Am J Sports Med. 2012; 40(3): 574-83.

References

328

22. Shin CS, Chaudhari AM, Andriacchi TP. Valgus plus internal rotation moments increase anterior cruciate ligament strain more than either alone. Med Sci Sports Exerc. 2011; 43(8): 1484-91.

23. Withrow TJ, Huston LJ, Wojtys EM, Ashton-Miller JA. The effect of an impulsive knee valgus moment on in vitro relative ACL strain during a simulated jump landing. Clin Biomech. 2006; 21(9): 977-83.

24. Mizuno K, Andrish JT, van den Bogert AJ, McLean SG. Gender dimorphic ACL strain in response to combined dynamic 3D knee joint loading: Implications for ACL injury risk. Knee. 2009; 16(6): 432-40.

25. Taylor KA, Terry ME, Utturkar GM, Spritzer CE, Queen RM, Irribarra LA, Garrett WE, DeFrate LE. Measurement of in vivo anterior cruciate ligament strain during dynamic jump landing. J Biomech. 2011; 44(3): 365-71.

26. DeMorat G, Weinhold P, Blackburn T, Chudik S, Garrett Jr W. Aggressive quadriceps loading can induce noncontact anterior cruciate ligament injury. Am J Sports Med. 2004; 32(2): 477-83.

27. Cerulli G, Benoit DL, Lamontagne M, Caraffa A, Liti A. In vivo anterior cruciate ligament strain behaviour during a rapid deceleration movement: Case report. Knee Surg Sports Traumatol Arthrosc. 2003; 11(5): 307-11.

28. Kiapour AM, Quatman CE, Goel VK, Wordeman SC, Hewett TE, Demetropoulos CK. Timing sequence of multi-planar knee kinematics revealed by physiologic cadaveric simulation of landing: Implications for ACL injury mechanism. Clin Biomech. 2014; 29(1): 75-82.

29. Berns GS, Hull ML, Patterson HA. Strain in the anteriomedial bundle of the anterior cruciate ligament under combination loading. J Orthop Res. 1992; 10(2): 167-76.

30. Shultz SJ, Schmitz RJ, Benjaminse A, Collins M, Ford K, Kulas AS. ACL Research Retreat VII: An update on anterior cruciate ligament injury risk factor identification, screening, and prevention. J Athl Train. 2015; 50(10): 1076-93.

31. Quatman CE, Kiapour AM, Demetropoulos CK, Kiapour A, Wordeman SC, Levine JW, Goel VK, Hewett TE. Preferential loading of the ACL compared with the MCL during landing: A novel in sim approach yields the multiplanar mechanism of dynamic valgus during ACL injuries. Am J Sports Med. 2014; 42(1): 177-86.

References

329

32. Kiapour AM, Kiapour A, Goel VK, Quatman CE, Wordeman SC, Hewett TE, Demetropoulos CK. Uni-directional coupling between tibiofemoral frontal and axial plane rotation supports valgus collapse mechanism of ACL injury. J Biomech. 2015; 48(10): 1745-51.

33. Quatman CE, Quatman-Yates CC, Hewett TE. A 'plane' explanation of anterior cruciate ligament injury mechanisms: A systematic review. Sports Med. 2010; 40(9): 729-46.

34. Wojtys EM, Huston LJ, Taylor PD, Bastian SD. Neuromuscular adaptations in isokinetic, isotonic, and agility training programs. Am J Sports Med. 1996; 24(2): 187-92.

35. Cochrane J, Lloyd D, Besier T, Ackland T, Elliot B. The effect of lower limb training programs on muscular support of the knee and prevention of ACL injury. J Sci Med Sport. 2003; 6(S1): 89.

36. Chimera NJ, Swanik KA, Swanik CB, Straub SJ. Effects of plyometric training on muscle-activation strategies and performance in female athletes. J Athl Train. 2004; 39(1): 24-31.

37. Hewett TE, Stroupe AL, Nance TA, Noyes FR. Plyometric training in female athletes: Decreased impact forces and increased hamstring torques. Am J Sports Med. 1996; 24(6): 765-73.

38. Wilkerson GB, Colston MA, Short NI, Neal KL, Hoewischer PE, Pixley JJ. Neuromuscular changes in female collegiate athletes resulting from a plyometric jump-training program. J Athl Train. 2004; 39(1): 17-23.

39. Lephart SM, Abt JP, Ferris CM, Sell TC, Nagai T, Myers JB, Irrgang JJ. Neuromuscular and biomechanical characteristic changes in high school athletes: A plyometric versus basic resistance program. Br J Sports Med. 2005; 39(12): 932-8.

40. Zebis MK, Bencke J, Andersen LL, Døssing S, Alkjaer T, Magnusson SP, Kjaer M, Aagaard P. The effects of neuromuscular training on knee joint motor control during sidecutting in female elite soccer and handball players. Clin J Sport Med. 2008; 18(4): 329-37.

41. Nagano Y, Hirofumi I, Akai M, Fukubayashi T. Effects of jump and balance training on knee kinematics and electromyography of female basketball athletes during a single limb drop landing: Pre-post intervention study. Sports Med Arthrosc Rehabil Ther Tech. 2011; 3(1): 14-21.

References

330

42. Tsang KKW, DiPasquale AA. Improving the Q:H strength ratio in women using plyometric exercises. J Strength Cond Res. 2011; 25(10): 2740-5.

43. Hurd WJ, Chmielewski TL, Snyder-Mackler L. Perturbation-enhanced neuromuscular training alters muscle activity in female athletes. Knee Surg Sports Traumatol Arthrosc. 2006; 14(1): 60-9.

44. Wilderman DR, Ross SE, Padua DA. Thigh muscle activity, knee motion, and impact force during side-step pivoting in agility-trained female basketball players. J Athl Train. 2009; 44(1): 14-25.

45. Holcomb WR, Rubley MD, Lee HJ, Guadagnoli MA. Effect of hamstring-emphasized resistance training on hamstring:quadriceps strength ratios. J Strength Cond Res. 2007; 21(1): 41-7.

46. Hashemi J, Chandrashekar N, Jang T, Karpat F, Oseto M, Ekwaro-Osire S. An alternative mechanism of non-contact anterior cruciate ligament injury during jump-landing: In-vitro simulation. Exp Mech. 2007; 47(3): 347-54.

47. Griffin LY, Agel J, Albohm MJ, Arendt EA, Dick RW, Garrett WE, Garrick JG, Hewett TE, Huston L, Ireland ML, Johnson RJ, Kibler WB, Lephart S, Lewis JL, Lindenfeld TN, Mandelbaum BR, Marchak P, Teitz CC, Wojtys EM. Noncontact anterior cruciate ligament injuries: Risk factors and prevention strategies. J Am Acad Orthop Surg. 2000; 8(3): 141-50.

48. Hug F, Turpin NA, Couturier A, Dorel S. Consistency of muscle synergies during pedaling across different mechanical constraints. J Neurophysiol. 2011; 106(1): 91-103.

49. Hug F, Turpin NA, Guével A, Dorel S. Is interindividual variability of EMG patterns in trained cyclists related to different muscle synergies? J Appl Physiol. 2010; 108(6): 1727-36.

50. Frère J, Hug F. Between-subject variability of muscle synergies during a complex motor skill. Front Comput Neurosci. 2012; 6: 99.

51. Latash ML, Scholz JP, Schoener G. Motor control strategies revealed in the structure of motor variability. Exerc Sport Sci Rev. 2002; 30(1): 26-31.

52. Slifkin AB, Newell KM. Is variability in human performance a reflection of system noise? Curr Dir Psychol Sci. 1998; 7(6): 170-7.

References

331

53. Winter DA, Yack HA. EMG profiles during normal human walking: Stride-to-stride and inter-subject variability. Electroencephalogr Clin Neurophysiol. 1987; 67(5): 402-11.

54. Cowling EJ, Steele JR. The effect of upper-limb motion on lower-limb muscle synchrony. J Bone Joint Surg. 2001; 83(1): 35-41.

55. Cowling EJ, Steele JR. Is lower limb muscle synchrony during landing affected by gender? Implications for variations in ACL injury rates. J Electromyogr Kinesiol. 2001; 11(4): 263-8.

56. McLean SG, Borotikar B, Lucey SM. Lower limb muscle pre-motor time measures during a choice reaction task associate with knee abduction loads during dynamic single leg landings. Clin Biomech. 2010; 25(6): 563-9.

57. Carcia C, Eggen J, Shultz S. Hip-abductor fatigue, frontal-plane landing angle, and excursion during a drop jump. J Sport Rehabil. 2005; 14(4): 321-31.

58. Simonsen EB, Magnusson SP, Bencke J, Nasborg H, Havkrog M, Ebstrup JF, Sorensen H. Can the hamstring muscles protect the anterior cruciate ligament during a side-cutting maneuver? Scand J Med Sci Sports. 2000; 10(2): 78-84.

59. Sigward SM, Powers CM. The influence of gender on knee kinematics, kinetics and muscle activation patterns during side-step cutting. Clin Biomech. 2006; 21(1): 41-8.

60. Myer GD, Ford KR, Hewett TE. The effects of gender on quadriceps muscle activation strategies during a maneuver that mimics a high ACL injury risk position. J Electromyogr Kinesiol. 2005; 15(2): 181-9.

61. DeMont RG, Lephart SM, Giraldo JL, Swanik CB, Fu FH. Muscle preactivity of anterior cruciate ligament-deficient and -reconstructed females during functional activities. J Athl Train. 1999; 34(2): 115-20.

62. Shultz SJ, Carcia CR, Perrin DH. Knee joint laxity affects muscle activation patterns in the healthy knee. J Electromyogr Kinesiol. 2004; 14(4): 475-83.

63. Chappell JD, Creighton A, Giuliani C, Yu B, Garrett WE. Kinematics and electromyography of landing preparation in vertical stop-jump: Risks for noncontact anterior cruciate ligament injury. Am J Sports Med. 2007; 35(2): 235-41.

References

332

64. Colby S, Francisco A, Yu B, Kirkendall D, Finch M, Garrett W. Electromyographic and kinematic analysis of cutting maneuvers: Implications for anterior cruciate ligament injury. Am J Sports Med. 2000; 28(2): 234-40.

65. Mandelbaum BR, Silvers HJ, Watanabe DS, Knarr JF, Thomas SD, Griffin LY, Kirkendall DT, Garrett Jr W. Effectiveness of a neuromuscular and proprioceptive training program in preventing anterior cruciate ligament injuries in female athletes: 2-year follow-up. Am J Sports Med. 2005; 33(7): 1003-10.

66. Olsen OE, Myklebust G, Engebretsen L, Holm I, Bahr R. Exercises to prevent lower limb injuries in youth sports: Cluster randomised controlled trial. Br Med J. 2005; 330: 449.

67. Petersen W, Braun C, Bock W, Schmidt K, Weimann A, Drescher W, Eiling E, Stange R, Fuchs T, Hedderich J, Zantop T. A controlled prospective case control study of a prevention training program in female team handball players: The German experience. Arch Orthop Trauma Surg. 2005; 125(9): 614-21.

68. Chappell JD, Limpisvasti O. Effect of a neuromuscular training program on the kinetics and kinematics of jumping tasks. Am J Sports Med. 2008; 36(6): 1081-6.

69. Lim BO, Lee YS, Kim JG, An KO, Yoo J, Kwon YH. Effects of sports injury prevention training on the biomechanical risk factors of anterior cruciate ligament injury in high school female basketball players. Am J Sports Med. 2009; 37(9): 1728-34.

70. Myer GD, Ford KR, Palumbo JP, Hewett TE. Neuromuscular training improves performance and lower-extremity biomechanics in female athletes. J Strength Cond Res. 2005; 19(1): 51-60.

71. Pollard CD, Sigward SM, Ota S, Langford K, Powers CM. The influence of in-season injury prevention training on lower-extremity kinematics during landing in female soccer players. Clin J Sport Med. 2006; 16(3): 223-7.

72. Myer GD, Ford KR, McLean SG, Hewett TE. The effects of plyometric versus dynamic stabilization and balance training on lower extremity biomechanics. Am J Sports Med. 2006; 34(3): 445-55.

73. Onate JA, Guskiewicz KM, Marshall SW, Giuliani C, Yu B, Garrett WE. Instruction of jump-landing technique using videotape feedback: Altering lower extremity motion patterns. Am J Sports Med. 2005; 33(6): 831-42.

References

333

74. Brown TN, Palmieri-Smith RM, McLean SG. Comparative adaptations of lower limb biomechanics during unilateral and bilateral landings after different neuromuscular-based ACL injury prevention protocols. J Strength Cond Res. 2014; 28(10): 2859-71.

75. Celebrini RG, Eng JJ, Miller WC, Ekegren CL, Johnston JD, Depew TA, Maclntyre DL. Effect of a novel movement strategy in decreasing ACL risk factors in female adolescent soccer players: A randomized controlled trial. Clin J Sport Med. 2014; 24(2): 134-41.

76. de Marche Baldon R, Moreira Lobato DF, Yoshimatsu AP, dos Santos AF, Francisco AL, Santiago PRP, Serrãp FV. Effect of plyometric training on lower limb biomechanics in females. Clin J Sport Med. 2014; 24(1): 44-50.

77. Bell DR, Oates DC, Clark MA, Padua DA. Two- and 3-dimensional knee valgus are reduced after an exercise intervention in young adults with demonstrable valgus during squatting. J Athl Train. 2013; 48(4): 442-9.

78. Pfile KR, Hart JM, Herman DC, Hertel J, Kerrigan DC, Ingersoll CD. Different exercise training interventions and drop landing biomechanics in high school female athletes. J Athl Train. 2013; 48(4): 450-62.

79. Tate JJ, Milner CE, Fairbrother JT, Zhang S. The effects of a home-based instructional program aimed at improving frontal plane knee biomechanics during a jump-landing task. J Orthop Sports Phys Ther. 2013; 43(7): 486-94.

80. Greska EK, Cortes N, Van Lunen BL, Oñate JA. A feedback inclusive neuromuscular training program alters frontal plane kinematics. J Strength Cond Res. 2012; 26(6): 1609-19.

81. Benjaminse A, Gokeler A, Dowling AV, Faigenbaum A, Ford KR, Hewett TE, Onate JA, Otten B, Myer GD. Optimization of the anterior cruciate ligament injury prevention paradigm: Novel feedback techniques to enhance motor learning and reduce injury risk. J Orthop Sports Phys Ther. 2015; 45(3): 170-82.

82. Cowling E, Steele J, McNair P. Training quadriceps-hamstring muscle synchrony: Do group data mask individual lower limb muscle coordination strategies? J Sci Med Sport. 2003; 6(S1): 90.

83. Dempsey AR, Elliott BC, Munro BJ, Steele JR, Lloyd DG. Can technique modification training reduce knee moments in a landing task? J Appl Biomech. 2014; 30(2): 231-6.

References

334

84. Myer GD, Ford KR, Brent JL, Hewett TE. Differential neuromuscular training effects on ACL injury risk factors in "high-risk" versus "low-risk" athletes. BMC Musc Disord. 2007; 8: 39-45.

85. Ekegren CL, Miller WC, Celebrini RG, Eng JJ, MacIntyre DL. Reliability and validity of observational risk screening in evaluating dynamic knee valgus. J Orthop Sports Phys Ther. 2009; 39(9): 665-74.

86. Myer GD, Ford KR, Hewett TE. Tuck jump assessment for reducing anterior cruciate ligament injury risk. Athl Ther Today. 2008; 13(5): 39-44.

87. Myer GD, Ford KR, Hewett TE. New method to identify athletes at high risk of ACL injury using clinic-based measurements and freeware computer analysis. Br J Sports Med. 2011; 45(4): 238-44.

88. Myer GD, Ford KR, Khoury J, Succop P, Hewett TE. Development and validation of a clinic-based prediction tool to identify female athletes at high risk for anterior cruciate ligament injury. Am J Sports Med. 2010; 38(10): 2025-33.

89. Myer GD, Ford KR, Khoury J, Succop P, Hewett TE. Clinical correlates to laboratory measures for use in non-contact anterior cruciate ligament injury risk prediction algorithm. Clin Biomech. 2010; 25(7): 693-9.

90. Myer GD, Ford KR, Khoury J, Succop P, Hewett TE. Biomechanics laboratory-based prediction algorithm to identify female athletes with high knee loads that increase risk of ACL injury. Br J Sports Med. 2011; 45(4): 245-52.

91. Padua DA, Marshall SW, Boling MC, Thigpen CA, Garrett Jr WE, Beutler AI. The Landing Error Scoring System (LESS) is a valid and reliable clinical assessment tool of jump-landing biomechanics: The JUMP-ACL study. Am J Sports Med. 2009; 37(10): 1996-2002.

92. McLean SG, Walker K, Ford KR, Myer GD, Hewett TE, van den Bogert AJ. Evaluation of a two dimensional analysis method as a screening and evaluation tool for anterior cruciate ligament injury. Br J Sports Med. 2005; 39(6): 355-62.

93. Myer GD, Ford KR, Khoury J, Hewett TE. Three-dimensional motion analysis validation of a clinic-based nomogram designed to identify high ACL injury risk in female athletes. Physician Sportsmed. 2011; 39(1): 19-28.

References

335

94. Onate J, Cortes N, Welch C, van Lunen B. Expert versus novice interrater reliability and criterion validity of the Landing Error Scoring System. J Sport Rehabil. 2010; 19(1): 41-56.

95. Edwards S, Steele JR, McGhee DE. Does a drop landing represent a whole skill landing and is this moderated by fatigue? Scand J Med Sci Sports. 2010; 20(3): 516-23.

96. Kristianslund E, Krosshaug T. Comparison of drop jumps and sport-specific sidestep cutting: Implications for anterior cruciate ligament injury risk screening. Am J Sports Med. 2013; 41(3): 684-8.

97. Boden BP, Dean GS, Feagin Jr JA, Garrett Jr WE. Mechanisms of anterior cruciate ligament injury. Orthopedics. 2000; 23(6): 573-8.

98. Krosshaug T, Nakamae A, Boden BP, Engebretsen L, Smith G, Slauterback JR, Hewett TE, Bahr R. Mechanisms of anterior cruciate ligament injury in basketball: Video analysis of 39 cases. Am J Sports Med. 2007; 35(3): 359-67.

99. Olsen OE, Myklebust G, Engebretsen L, Bahr R. Injury mechanisms for anterior cruciate ligament injuries in team handball: A systematic video analysis. Am J Sports Med. 2004; 32(4): 1002-12.

100. Waldén M, Krosshaug T, Bjørneboe J, Einar Andersen T, Faul O, Hägglund M. Three distinct mechanisms predominate in noncontact anterior cruciate ligament injuries in male professional football players: A systematic video analysis of 39 cases. Br J Sports Med. 2015; 49(22): 1452-60.

101. Monfort SM, Comstock RD, Collins CL, Onate JA, Best TM, Chaudhari AMW. Association between ball-handling versus defending actions and acute noncontact lower extremity injuries in high school basketball and soccer. Am J Sports Med. 2015; 43(4): 802-7.

102. Cochrane J, Lloyd D, Buttfield A, Seward H, McGivern J. Characteristics of anterior cruciate ligament injuries in Australian football. J Sci Med Sport. 2007; 10(2): 96-104.

103. Stuelcken MC, Mellifont DB, Gorman AD, Sayers MGL. Mechanisms of anterior cruciate ligament injuries in elite women's netball: A systematic video analysis. J Sports Sci. 2016; 34(16): 1516-22.

References

336

104. Padua DA, DiStefano LJ, Beutler AI, de la Motte SJ, DiStefano MJ, Marshall SW. The Landing Error Scoring System as a screening tool for an anterior cruciate ligament injury-prevention program in elite-youth soccer athletes. J Athl Train. 2015; 50(6): 589-95.

105. Smith HC, Johnson RJ, Shultz SJ, Tourville T, Holterman LA, Slauterbeck J, Vacek PM, Beynnon BD. A prospective evaluation of the Landing Error Scoring System (LESS) as a screening tool for anterior cruciate ligament injury risk. Am J Sports Med. 2012; 40(3): 521-6.

106. Hewett TE, Ford KR, Hoogenboom BJ, Myer GD. Understanding and preventing ACL injuries: Current biomechanical and epidemiologic considerations - Update 2010. N Am J Sports Phys Ther. 2010; 5(4): 234-51.

107. Hewett TE, Myer GD, Ford KR. Anterior cruciate ligament injuries in female athletes Part 1: Mechanisms and risk factors. Am J Sports Med. 2006; 34(2): 299-311.

108. Gillquist J, Messner K. Anterior cruciate ligament reconstruction and the long term incidence of gonarthrosis. Sports Med. 1999; 27(3): 143-56.

109. Barenius B, Ponzer S, Shalabi A, Bujak R, Norlén L, Eriksson K. Increased risk of osteoarthritis after anterior cruciate ligament reconstruction: A 14-year follow-up study of a randomized controlled trial. Am J Sports Med. 2014; 42(5): 1049-57.

110. Conaghan PG. Update on osteoarthritis, Part 1: Current concepts and the relation to exercise. Br J Sports Med. 2002; 36(5): 330-3.

111. Andriacchi TP, Dyrby CO. Interactions between kinematics and loading during walking for the normal and ACL deficient knee. J Biomech. 2005; 38(2): 293-8.

112. Maletius W, Messner K. Eighteen- to twenty-four-year follow-up after complete rupture of the anterior cruciate ligament. Am J Sports Med. 1999; 27(6): 711-7.

113. Øiestad B, Holm I, Engebretsen L, Aune A, Gunderson R, Risberg M. The prevalence of patellofemoral osteoarthritis 12 years after anterior cruciate ligament reconstruction. Knee Surg Sports Traumatol Arthrosc. 2013; 21(4): 942-9.

References

337

114. Harris KP, Driban JB, Sitler MR, Cattano NM, Balasubramanian E, Hootman JM. Tibiofemoral osteoarthritis after surgical or nonsurgical treatment of anterior cruciate ligament rupture: A systematic review. J Athl Train. 2015; In Press.

115. Grindem H, Eitzen I, Engebretsen L, Snyder-Mackler L, Risberg MA. Nonsurgical or surgical treatment of ACL injuries: Knee function, sports participation, and knee reinjury. J Bone Joint Surg. 2014; 96(15): 1233-41.

116. Filbay SR, Ackerman IN, Russell TG, Macri EM, Crossley KM. Health-related quality of life after anterior cruciate ligament reconstruction: A systematic review. Am J Sports Med. 2014; 42(5): 1247-55.

117. Filbay SR, Culvenor AG, Ackerman IN, Russell TG, Crossley KM. Quality of life in anterior cruciate ligament-deficient individuals: A systematic review and meta-analysis. Br J Sports Med. 2015; 49(16): 1033-41.

118. Paterno MV, Rauh MJ, Schmitt LC, Ford KR, Hewett TE. Incidence of second ACL injuries 2 years after primary ACL reconstruction and return to sport. Am J Sports Med. 2014; 42(7): 1567-73.

119. Flood L, Harrison JE. Epidemiology of basketball and netball injuries that resulted in hospital admission in Australia, 2000-2004. Med J Aust. 2009; 190(2): 87-90.

120. Bencke J, Zebis MK. The influence of gender on neuromuscular pre-activity during side-cutting. J Electromyogr Kinesiol. 2011; 21(2): 371-5.

121. Decker MJ, Torry MR, Wyland DJ, Sterett WI, Richard Steadman J. Gender differences in lower extremity kinematics, kinetics and energy absorption during landing. Clin Biomech. 2003; 18(7): 662-9.

122. Ford KR, Shapiro R, Myer GD, van den Bogert AJ, Hewett TE. Longitudinal sex differences during landing in knee abduction in young athletes. Med Sci Sports Exerc. 2010; 42(10): 1923-31.

123. Garrison JC, Hart JM, Palmieri RM, Kerrigan DC, Ingersoll CD. Comparison of knee-joint moments in male and female college soccer players during a single-leg landing. J Sport Rehabil. 2005; 14(4): 332-7.

124. Garrison JC, Hart JM, Palmieri RM, Kerrigan DC, Ingersoll CD. Lower extremity EMG in male and female college soccer players during single-leg landing. J Sport Rehabil. 2005; 14(1): 48-57.

References

338

125. James CR, Sizer PS, Starch DW, Lockhart TE, Slauterbeck J. Gender differences among sagittal plane knee kinematic and ground reaction force characteristics during a rapid sprint and cut maneuver. Res Q Exerc Sport. 2004; 75(1): 31-8.

126. McLean SG, Walker KB, van den Bogert AJ. Effect of gender on lower extremity kinematics during rapid direction changes: An integrated analysis of three sports movements. J Sci Med Sport. 2005; 8(4): 411-22.

127. Schmitz RJ, Kulas AS, Perrin DH, Riemann BL, Shultz SJ. Sex differences in lower extremity biomechanics during single leg landings. Clin Biomech. 2007; 22(6): 681-8.

128. Hughes G, Watkins J, Owen N. Differences between the sexes in knee kinetics during landing from volleyball block jumps. Eur J Sport Sci. 2010; 10(1): 1-11.

129. Nagano Y, Ida H, Akai M, Fukubayashi T. Gender differences in knee kinematics and muscle activity during single limb drop landing. Knee. 2007; 14: 218-23.

130. Hughes G, Dally N. Gender difference in lower limb muscle activity during landing and rapid change of direction. Sci Sports. 2015; 30(3): 163-8.

131. Beaulieu ML, Lamontagne M, Xu L. Lower limb muscle activity and kinematics of an unanticipated cutting manoeuvre: A gender comparison. Knee Surg Sports Traumatol Arthrosc. 2009; 17(8): 968-76.

132. Kibler WB, Chandler TJ, Cabell L, Shapiro R. Gender differences in muscle activation around the hip, knee, and ankle in cutting maneuvers. Clin Biomech. 2001; 16(10): 945.

133. Iguchi J, Tateuchi H, Taniguchi M, Ichihashi N. The effect of sex and fatigue on lower limb kinematics, kinetics and muscle activity during unanticipated side-step cutting. Knee Surg Sports Traumatol Arthrosc. 2014; 22(1): 41-8.

134. Beaulieu ML, Lamontagne M, Xu L. Gender differences in time-frequency EMG analysis of unanticipated cutting maneuvers. Med Sci Sports Exerc. 2008; 40(10): 1795-804.

135. Kernozek TW, Torry MR, Van Hoof H, Cowley H, Tanner S. Gender differences in frontal and sagittal plane biomechanics during drop landings. Med Sci Sports Exerc. 2005; 37(6): 1003-12.

References

339

136. Miranda DL, Fadale PD, Hulstyn MJ, Shalvoy RM, Machan JT, Fleming BC. Knee biomechanics during a jump-cut maneuver: Effects of sex and ACL surgery. Med Sci Sports Exerc. 2013; 45(5): 942-51.

137. Hughes G, Watkins J, Owen N. Gender differences in lower limb frontal plane kinematics during landing. Sports Biomech. 2008; 7(3): 333-41.

138. McLean SG, Lipfert SW, van den Bogert AJ. Effect of gender and defensive opponent on the biomechanics of sidestep cutting. Med Sci Sports Exerc. 2004; 36(6): 1008-16.

139. Kernozek TW, Torry MR, Iwasaki M. Gender differences in lower extremity landing mechanics caused by neuromuscular fatigue. Am J Sports Med. 2008; 36(3): 554-65.

140. Schmitz RJ, Shultz SJ, Anh-Dung N. Dynamic valgus alignment and functional strength in males and females during maturation. J Athl Train. 2009; 44(1): 26-32.

141. Ford KR, Myer GD, Hewett TE. Valgus knee motion during landing in high school female and male basketball players. Med Sci Sports Exerc. 2003; 35(10): 1745-50.

142. Gehring D, Melnyk M, Gollhofer A. Gender and fatigue have influence on knee joint control strategies during landing. Clin Biomech. 2009; 24(1): 82-7.

143. Hughes G, Watkins J, Owen N. The effects of opposition and gender on knee kinematics and ground reaction force during landing from volleyball block jumps. Res Q Exerc Sport. 2010; 81(4): 384-91.

144. Pappas E, Sheikhzadeh A, Hagins M, Nordin M. The effect of gender and fatigue on the biomechanics of bilateral landings from a jump: Peak values. J Sports Sci Med. 2007; 6(1): 77-84.

145. Russell KA, Palmieri RM, Zinder SM, Ingersoll CD. Sex differences in valgus knee angle during a single-leg drop jump. J Athl Train. 2006; 41(2): 166-71.

146. Walsh MS, Waters J, Kersting UG. Gender bias on the effects of instruction on kinematic and kinetic jump parameters of high-level athletes. Res Sports Med. 2007; 15(4): 283-95.

References

340

147. Pollard CD, Davis IM, Hamill J. Influence of gender on hip and knee mechanics during a randomly cued cutting maneuver. Clin Biomech. 2004; 19(10): 1022-31.

148. Orishimo KF, Liederbach M, Kremenic IJ, Hagins M, Pappas E. Comparison of landing biomechanics between male and female dancers and athletes, Part 1: Influence of sex on risk of anterior cruciate ligament injury. Am J Sports Med. 2014; 42(5): 1082-8.

149. McLean SG, Huang X, van den Bogert AJ. Association between lower extremity posture at contact and peak knee valgus moment during sidestepping: Implications for ACL injury. Clin Biomech. 2005; 20(8): 863-70.

150. Norcross MF, Lewek MD, Padua DA, Shultz SJ, Weinhold PS, Blackburn JT. Lower extremity energy absorption and biomechanics during landing, Part II: Frontal-plane energy analyses and interplanar relationships. J Athl Train. 2013; 48(6): 757-63.

151. Kiriyama S, Sato H, Takahira N. Gender differences in rotation of the shank during single-legged drop landing and its relation to rotational muscle strength of the knee. Am J Sports Med. 2009; 37(1): 167-74.

152. McLean SG. The ACL injury enigma: We can't prevent what we don't understand. J Athl Train. 2008; 43(5): 538-40.

153. Kirkendall DT, Garrett Jr WE. The anterior cruciate ligament enigma: Injury mechanisms and prevention. Clin Orthop Relat Res. 2000; 372: 64-8.

154. Chaudhari AM, Hearn BK, Andriacchi TP. Sport-dependent variations in arm position during single-limb landing influence knee loading: Implications for anterior cruciate ligament injury. Am J Sports Med. 2005; 33(6): 824-30.

155. Fedie R, Carlstedt K, Willson JD, Kernozek TW. Effect of attending to a ball during a side-cut maneuver on lower extremity biomechanics in male and female athletes. Sports Biomech. 2010; 9(3): 165-77.

156. Fox A, Spittle M, Otago L, Saunders N. Offensive agility techniques performed during international netball competition. Int J Sports Sci Coach. 2014; 9(3): 543-52.

157. Fox AS, Spittle M, Otago L, Saunders N. Descriptive analysis of landings during international netball competition: Enhancing ecological validity of laboratory testing environments. Int J Perform Anal Sport. 2013; 13(3): 690-702.

References

341

158. Lloyd DG. Rationale for training programs to reduce anterior cruciate ligament injuries in Australian football. J Orthop Sports Phys Ther. 2001; 31(11): 645-54.

159. Hewett TE, Torg JS, Boden BP. Video analysis of trunk and knee motion during non-contact anterior cruciate ligament injury in female athletes: Lateral trunk and knee abduction motion are combined components of the injury mechanism. Br J Sports Med. 2009; 43(6): 417-22.

160. Hewett TE, Myer GD, Ford KR, Heidt Jr RS, Colosimo AJ, McLean SG, van den Bogert AJ, Paterno MV, Succop P. Biomechanical measures of neuromuscular control and valgus loading of the knee predict anterior cruciate ligament injury risk in female athletes: A prospective study. Am J Sports Med. 2005; 33(4): 492-501.

161. Donnelly CJ, Elliott BC, Ackland TR, Doyle TLA, Beiser TF, Finch CF, Cochrane JL, Dempsey AR, Lloyd DG. An anterior cruciate ligament injury prevention framework: Incorporating the recent evidence. Res Sports Med. 2012; 20(3-4): 239-62.

162. Powers CM. The influence of abnormal hip mechanics on knee injury: A biomechanical perspective. J Orthop Sports Phys Ther. 2010; 40(2): 42-51.

163. Kipp K, McLean SG, Palmieri-Smith RM. Patterns of hip flexion motion predict frontal and transverse plane knee torques during a single-leg land-and-cut maneuver. Clin Biomech. 2011; 26(5): 504-8.

164. Jones PA, Herrington LC, Graham-Smith P. Technique determinants of knee joint loads during cutting in female soccer players. Hum Movement Sci. 2015; 42: 203-11.

165. Brown TN, Palmieri-Smith RM, McLean SG. Knee kinematics during single- and double-legged jump landings following six weeks of neuromuscular training. J Athl Train. 2010; 45(5): 530.

166. Dempsey AR, Lloyd DG, Elliott BC, Steele JR, Munro BJ. Changing sidestep cutting technique reduces knee valgus loading. Am J Sports Med. 2009; 37(11): 2194-200.

167. Dempsey AR, Lloyd DG, Elliott BC, Steele JR, Munro BJ, Russo KA. The effect of technique change on knee loads during sidestep cutting. Med Sci Sports Exerc. 2007; 39(10): 1765-73.

References

342

168. Sigward SM, Ota S, Powers CM. Predictors of frontal plane knee excursion during a drop land in young female soccer players. J Orthop Sports Phys Ther. 2008; 38(11): 661-7.

169. Wahlstedt C, Rasmussen-Barr E. Anterior cruciate ligament injury and ankle dorsiflexion. Knee Surg Sports Traumatol Arthrosc. 2015; 23(11): 3202-7.

170. Lloyd DG, Buchanan TS. Strategies of muscular support of varus and valgus isometric loads at the human knee. J Biomech. 2001; 34(10): 1257-67.

171. Besier TF, Lloyd DG, Ackland TR. Muscle activation strategies at the knee during running and cutting maneuvers. Med Sci Sports Exerc. 2003; 35(1): 119-27.

172. Neptune RR, Wright IC, van den Bogert AJ. Muscle coordination and function during cutting movements. Med Sci Sports Exerc. 1999; 31(2): 294-302.

173. Zhang SN, Bates BT, Dufek JS. Contributions of lower extremity joints to energy dissipation during landings. Med Sci Sports Exerc. 2000; 32(4): 812-9.

174. Brown TN, McLean SG, Palmieri-Smith RM. Associations between lower limb muscle activation strategies and resultant multi-planar knee kinetics during single leg landings. J Sci Med Sport. 2014; 17(4): 408-13.

175. Padua DA, Bell DR, Clark MA. Neuromuscular characteristics of individuals displaying excessive medial knee displacement. J Athl Train. 2012; 47(5): 525-36.

176. McLean SG, Huang X, van den Bogert AJ. Investigating isolated neuromuscular control contributions to non-contact anterior cruciate ligament injury risk via computer simulation methods. Clin Biomech. 2008; 23(7): 926-36.

177. Zebis MK, Andersen LL, Bencke J, Kjær M, Aagaard P. Identification of athletes at future risk of anterior cruciate ligament ruptures by neuromuscular screening. Am J Sports Med. 2009; 37(10): 1967-73.

178. Walsh M, Boling MC, McGrath M, Blackburn JT, Padua DA. Lower extremity muscle activation and knee flexion during a jump-landing task. J Athl Train. 2012; 47(4): 406-13.

179. Lee S, Chang J, Choi Y. Low limb muscle activation and joint angle in the sagittal plane during drop landing from various heights. J Phys Ther Sci. 2011; 23(2): 303-5.

References

343

180. Peng HT, Kernozek TW, Song CY. Quadricep and hamstring activation during drop jumps with changes in drop height. Phys Ther Sport. 2011; 12(3): 127-32.

181. Hanson AM, Padua DA, Blackburn JT, Prentice WE, Hirth CJ. Muscle activation during side-step cutting maneuvers in male and female soccer athletes. J Athl Train. 2008; 43(2): 133-43.

182. Li G, Rudy TW, Sakane M, Kanamori A, Ma CB, Woo SL. The importance of quadriceps and hamstring muscle loading on knee kinematics and in-situ forces in the ACL. J Biomech. 1999; 32(4): 395-400.

183. Withrow TJ, Huston LJ, Wojtys EM, Ashton-Miller JA. Effect of varying hamstring tension on anterior cruciate ligament strain during in vitro impulsive knee flexion and compression loading. J Bone Joint Surg. 2008; 90(4): 815-23.

184. Mokhtarzadeh H, Chen Hua Y, Goh JCH, Oetomo D, Malekipour F, Lee PVS. Contributions of the soleus and gastrocnemius muscles to the anterior cruciate ligament loading during single-leg landing. J Biomech. 2013; 46(11): 1913-20.

185. Morgan KD, Donnelly CJ, Reinbolt JA. Elevated gastrocnemius forces compensate for decreased hamstrings forces during the weight-acceptance phase of single-leg jump landing: Implications for anterior cruciate ligament injury risk. J Biomech. 2014; 47(13): 3295-302.

186. Saunders N, McLean SG, Fox AS, Otago L. Neuromuscular dysfunction that may predict ACL injury risk: A case report. Knee. 2014; 21(3): 789-92.

187. Palmieri-Smith RM, Wojtys EM, Ashton-Miller JA. Association between preparatory muscle activation and peak valgus knee angle. J Electromyogr Kinesiol. 2008; 18(6): 973-9.

188. Shultz SJ, Nguyen A, Leonard MD, Schmitz RJ. Thigh strength and activation as predictors of knee biomechanics during a drop jump task. Med Sci Sports Exerc. 2009; 41(4): 857-66.

189. Malfait B, Dingenen B, Staes F, Vanrenterghem J, Verschueren S. Differences in neuromuscular activity of quadriceps and hamstrings with respect to different landing patterns in female athletes. Br J Sports Med. 2014; 48(7): 631.

References

344

190. Hollman JH, Galardi CM, Lin IH, Voth BC, Whitmarsh CL. Frontal and transverse plane hip kinematics and gluteus maximus recruitment correlate with frontal plane knee kinematics during single-leg squat tests in women. Clin Biomech. 2014; 29(4): 468-74.

191. Hollman JH, Ginos BE, Kozuchowski J, Vaughn AS, Krause DA, Youdas JW. Relationships between knee valgus, hip-muscle strength, and hip-muscle recruitment during a single-limb step-down. J Sport Rehabil. 2009; 18(1): 104-17.

192. Hashemi J, Breighner R, Jang T-H, Chandrashekar N, Ekwaro-Osire S, Slauterbeck JR. Increasing pre-activation of the quadriceps muscle protects the anterior cruciate ligament during the landing phase of a jump: An in vitro simulation. Knee. 2010; 17(3): 235-41.

193. Ford KR, Myer GD, Toms HE, Hewett TE. Gender differences in the kinematics of unanticipated cutting in young athletes. Med Sci Sports Exerc. 2005; 37(1): 124-9.

194. Steams KM, Powers CM. The influence of hip abductor muscle performance on the knee valgus moment during landing: A gender comparison. J Athl Train. 2012; 47(5): e4.

195. Stearns KM, Powers CM. Improvements in hip muscle performance result in increased use of the hip extensors and abductors during a landing task. Am J Sports Med. 2014; 42(3): 602-9.

196. Hollman JH, Hohl JM, Kraft JL, Strauss JD, Traver KJ. Effects of hip extensor fatigue on lower extremity kinematics during a jump-landing task in women: A controlled laboratory study. Clin Biomech. 2012; 27(9): 903-9.

197. Hollman JH, Hohl JM, Kraft JL, Strauss JD, Traver KJ. Modulation of frontal-plane knee kinematics by hip-extensor strength and gluteus maximus recruitment during a jump-landing task in healthy women. J Sport Rehabil. 2013; 22(3): 184-90.

198. Khayambashi K, Ghoddosi N, Straub RK, Powers CM. Hip muscle strength predicts noncontact anterior cruciate ligament injury in male and female athletes: A prospective study. Am J Sports Med. 2016; 44(2): 355-61.

199. Zazulak BT, Ponce PL, Straub SJ, Medvecky MJ, Avedisian L, Hewett TE. Gender comparison of hip muscle activity during single-leg landing. J Orthop Sports Phys Ther. 2005; 35(5): 292-9.

References

345

200. Simpson CS, Sohn MH, Allen JL, Ting LH. Feasible muscle activation ranges based on inverse dynamics analyses of human walking. J Biomech. 2015; 48(12): 2990-7.

201. Myer GD, Ford KR, Brent JL, Hewett TE. The effects of plyometric versus dynamic stabilization and balance training on power, balance, and landing force in female athletes. J Strength Cond Res. 2006; 20(2): 345-53.

202. Caraffa A, Cerulli G, Projetti M, Aisa G, Rizzo A. Prevention of anterior cruciate ligament injuries in soccer. Knee Surg Sports Traumatol Arthrosc. 1996; 4(1): 19-21.

203. Ettlinger CF, Johnson RJ, Shealy JE. A method to help reduce the risk of serious knee sprains incurred in alpine skiing. Am J Sports Med. 1995; 23(5): 531-7.

204. Gilchrist J, Mandelbaum BR, Melancon H, Ryan GW, Silvers HJ, Griffin LY, Watanabe DS, Dick RW, Dvorak J. A randomized controlled trial to prevent non contact anterior cruciate ligament injury in female collegiate soccer players. Am J Sports Med. 2008; 36(8): 1476-83.

205. Heidt Jr RS, Sweeterman LM, Carlonas RL, Traub JA, Tekulve FX. Avoidance of soccer injuries with preseason conditioning. Am J Sports Med. 2000; 28(5): 659-62.

206. Hewett TE, Lindenfeld TN, Riccobene JV, Noyes FR. The effect of neuromuscular training on the incidence of knee injury in female athletes: A prospective study. Am J Sports Med. 1999; 27(6): 699-706.

207. Kiani A, Hellquist E, Ahlqvist K, Gedeborg R, Michaëlsson K, Byberg L. Prevention of soccer-related knee injuries in teenaged girls. Arch Intern Med. 2010; 170(1): 43-9.

208. Pfeiffer RP, Shea KG, Roberts D, Grandstrand S, Bond L. Lack of effect of a knee ligament injury prevention program on the incidence of noncontact anterior cruciate ligament injury. J Bone Joint Surg. 2006; 88(8): 1769-74.

209. Soderman K, Werner S, Pietila T, Engstrom B, Alfredson H. Balance board training: Prevention of traumatic injuries of the lower extremities in female soccer players? A prospective randomized intervention study. Knee Surg Sports Traumatol Arthrosc. 2000; 8(6): 356-63.

References

346

210. Steffen K, Myklebust G, Olsen OE, Holme I, Bahr R. Preventing injuries in female youth football: A cluster-randomized controlled trial. Scand J Med Sci Sports. 2008; 18(5): 605-14.

211. Pasanen K, Parkkani J, Pasanen M, Hiilloskorpi H, Mäkinen T, Järvinen M, Kannus P. Neuromuscular training and the risk of leg injuries in female floorball players: Cluster randomised controlled study. Br J Sports Med. 2008; 42(10): 502-5.

212. Myklebust G, Engebretsen L, Braekken IH, Skjolbert A, Olsen OE, Bahr R. Prevention of anterior cruciate ligament injuries in female team handball players: A prospective intervention study over three seasons. Clin J Sport Med. 2003; 13(2): 71-8.

213. Herman DC, Oñate JA, Weinhold PS, Guskiewicz KM, Garrett WE, Yu B, Padua DA. The effects of feedback with and without strength training on lower extremity biomechanics. Am J Sports Med. 2009; 37(7): 1301-8.

214. Aerts I, Cumps E, Verhagen E, Wuyts B, Van De Gucht S, Meeusen R. The effect of a 3-month prevention program on the jump-landing technique in basketball: A randomized controlled trial. J Sport Rehabil. 2015; 24(1): 21-30.

215. McCurdy K, Walker J, Saxe J, Woods J. The effect of short-term resistance training on hip and knee kinematics during vertical drop jumps. J Strength Cond Res. 2012; 26(5): 1257-64.

216. Cochrane JL, Lloyd DG, Besier TF, Elliott BC, Doyle TLA, Ackland TR. Training affects knee kinematics and kinetics in cutting maneuvers in sport. Med Sci Sports Exerc. 2010; 42(8): 1535-44.

217. Donnelly CJ, Elliott BC, Doyle TLA, Finch CF, Dempsey AR, Lloyd DG. Changes in knee joint biomechanics following balance and technique training and a season of Australian football. Br J Sports Med. 2012; 46(13): 917-22.

218. Gagnier JJ, Morgenstern H, Chess L. Interventions designed to prevent anterior cruciate ligament injuries in adolescents and adults: A systematic review and meta-analysis. Am J Sports Med. 2013; 41(8): 1952-62.

219. Stevenson JH, Beattie CS, Schwartz JB, Busconi BD. Assessing the effectiveness of neuromuscular training programs in reducing the incidence of anterior cruciate ligament injuries in female athletes: A systematic review. Am J Sports Med. 2015; 43(2): 482-90.

References

347

220. Steffen K, Emery CA, Romiti M, Jian K, Bizzini M, Dvorak J, Finch CF, Meeuwisse WH. High adherence to a neuromuscular injury prevention programme (FIFA 11+) improves functional balance and reduces injury risk in Canadian youth female football players: A cluster randomised trial. Br J Sports Med. 2013; 47(12): 794-802.

221. Hägglund M, Atroshi I, Wagner P, Waldén M. Superior compliance with a neuromuscular training programme is associated with fewer ACL injuries and fewer acute knee injuries in female adolescent football players: Secondary analysis of an RCT. Br J Sports Med. 2013; 47(15): 974-9.

222. Junge A, Roesch D, Peterson L, Graf-Baumann T, Dvorak J. Prevention of soccer injuries: A prospective intervention study in youth amateur players. Am J Sports Med. 2002; 30(5): 652-9.

223. Sugimoto D, Myer GD, Bush HM, Klugman MF, McKeon JMM, Hewett TE. Compliance with neuromuscular training and anterior cruciate ligament injury risk reduction in female athletes: A meta-analysis. J Athl Train. 2012; 47(6): 714-23.

224. Carow SD, Haniuk EM, Cameron KL, DiStefano LJ, Padua DA, Marshall SW, Beutler AI, Gerber JP. Risk of knee sprain following the implementation of an injury prevention program is associated with level of supervision. J Athl Train. 2012; 47(5): e8.

225. Fortington LV, Donaldson A, Lathlean T, Young WB, Gabbe BJ, Lloyd D, Finch CF. When ‘just doing it’ is not enough: Assessing the fidelity of player performance of an injury prevention exercise program. J Sci Med Sport. 2015; 18(3): 272-7.

226. Swart E, Redler L, Fabricant PD, Mandelbaum BR, Ahmad CS, Wang YC. Prevention and screening programs for anterior cruciate ligament injuries in young athletes: A cost-effectiveness analysis. J Bone Joint Surg. 2014; 96(9): 705-11.

227. Joy EA, Taylor JR, Novak MA, Chen M, Fink BP, Porucznik CA. Factors influencing the implementation of anterior cruciate ligament injury prevention strategies by girls soccer coaches. J Strength Cond Res. 2013; 27(8): 2263-9.

228. Keats MR, Emery CA, Finch CF. Are we having fun yet? Fostering adherence to injury preventive exercise recommendations in young athletes. Sports Med. 2012; 42(3): 175-84.

References

348

229. Frank B, Register-Mihalik J, Marshall S, Padua D. A coaching workshop improves coach intention but does not translate to implementation of a ACL injury prevention program. Br J Sports Med. 2014; 48(7): 595.

230. Myklebust G, Skjølberg A, Bahr R. ACL injury incidence in female handball 10 years after the Norwegian ACL prevention study: Important lessons learned. Br J Sports Med. 2013; 47(8): 476-9.

231. Renström PA. Eight clinical conundrums relating to anterior cruciate ligament (ACL) injury in sport: Recent evidence and a personal reflection. Br J Sports Med. 2013; 47(6): 1-7.

232. Australian Bureau of Statistics. Participation in sports and physical recreation, Australia, 2009-10. Canberra: ABS; 2010.

233. McGrath AC, Ozanne-Smith J. Attacking the goal of netball injury prevention: A review of the literature. Victoria: Monash University - Accident Research Centre, 1998.

234. Steele JR. Biomechanical factors affecting performance in netball: Implications for improving performance and injury reduction. Sports Med. 1990; 10(2): 88-102.

235. Hume PA, Steele JR. A preliminary investigation of injury prevention strategies in netball: Are players heeding the advice? J Sci Med Sport. 2000; 4(3): 406-13.

236. Hohmann E, Bryant A, Livingstone E, Reaburn P, Tetsworth K, Imhoff A, Bryant AL. Tibial acceleration profiles during the menstrual cycle in female athletes. Arch Orthop Trauma Surg. 2015; 135(10): 1419-27.

237. Hopper D, Elliott B. Lower-limb and back injury patterns of elite netball players. Sports Med. 1993; 16(2): 148-62.

238. Hopper D, Elliott B, Lalor J. A descriptive epidemiology of netball injuries during competition: A five year study. Br J Sports Med. 1995; 29(4): 223-8.

239. Otago L, Saunders N. An injury report to Netball Australia on the Commonwealth Bank Trophy competition 2006. University of Ballarat, 2007.

240. Otago L, Saunders N. An injury report to Netball Australia on the Commonwealth Bank Trophy competition 2007. University of Ballarat, 2009.

241. Otago L, Saunders N. An injury report to Netball Australia on the ANZ Championship competition 2008. University of Ballarat, 2010.

References

349

242. Saunders N, Otago L. Elite netball injury surveillance: Implications for injury prevention. J Sci Med Sport. 2009; 12(S1): S63.

243. Saunders N, Otago L. 2010 ANZ Championships injury report. University of Ballarat, 2010.

244. Otago L, Peake J. The role of insurance data in setting priorities for netball injury prevention strategies. J Sci Med Sport. 2007; 10(2): 105-9.

245. Otago L. Kinetic analysis of landings in netball: is a footwork rule change required to decrease ACL injuries? J Sci Med Sport. 2004; 7(1): 85-95.

246. Hopper D, Lo SK, Kirkham C, Elliott B. Landing patterns in netball: analysis of an international game. Br J Sports Med. 1992; 26(2): 101-6.

247. Brown S, Brughelli M, Hume P. Knee mechanics during planned and unplanned sidestepping: A systematic review and meta-analysis. Sports Med. 2014; 44(11): 1573-88.

248. Stuelcken M, Greene A, Smith R, Vanwanseele B. Knee loading patterns in a simulated netball landing task. Eur J Sport Sci. 2013; 13(5): 475-82.

249. Milner CE, Fairbrother JT, Srivatsan A, Zhang S. Simple verbal instruction improves knee biomechanics during landing in female athletes. Knee. 2012; 19(4): 399-403.

250. O'Connor KM, Bottum MC. Differences in cutting knee mechanics based on Principal Components Analysis. Med Sci Sports Exerc. 2009; 41(4): 867-78.

251. Sanna G, O'Connor KM. Fatigue-related changes in stance leg mechanics during sidestep cutting maneuvers. Clin Biomech. 2008; 23(7): 946-54.

252. Weinhandl JT, Joshi M, O'Connor KM. Gender comparisons between unilateral and bilateral landings. J Appl Biomech. 2010; 26(4): 444-53.

253. Vanwanseele B, Stuelcken M, Greene A, Smith R. The effect of external ankle support on knee and ankle joint movement and loading in netball players. J Sci Med Sport. 2014; 17(5): 511-5.

254. Saunders N. Characteristics of the female landing pattern. University of Ballarat: Ballarat, Victoria, Australia; 2006.

255. Dempsey AR, Elliott BC, Munro BJ, Steele JR, Lloyd DG. Whole body kinematics and knee moments that occur during an overhead catch and landing task in sport. Clin Biomech. 2012; 27(5): 466-74.

References

350

256. Herrington L, Myer GD, Munro A. Intra and inter-tester reliability of the tuck jump assessment. Phys Ther Sport. 2013; 14(3): 152-5.

257. Perrott MA, Pizzari T, Opar M, Cook J. Development of clinical rating criteria for tests of lumbopelvic stability. Rehabil Res Pract. 2012; 2012: 803637-43.

258. Saunders N, Otago L, Peoples J. Down to Earth: A practical guide to safe and effective landing in netball. University of Ballarat: School of Human Movement and Sport Sciences, 2006.

259. Saunders N, Otago L, Romiti M, Donaldson A, White P, Finch CF. Coaches' perspectives on implementing an evidence-informed injury prevention programme in junior community netball. Br J Sports Med. 2010; 44(15): 1128-32.

260. Enoka RM. Neuromechanics of Human Movement. 5th ed. Human Kinetics: Champaign, IL; 2015.

261. Winter DA. Biomechanics and Motor Control of Human Movement. 4th ed. John Wiley & Sons: New Jersey; 2009.

262. Hug F, Hodges PW, Tucker K. Muscle force cannot be directly inferred from muscle activation: Illustrated by the proposed imbalance of force between the vastus medialis and vastus lateralis in people with patellofemoral pain. J Orthop Sports Phys Ther. 2015; 45(5): 360-5.

263. Konrad P. The ABC of EMG: A practical introduction to kinesiological electromyography. 2005.

264. Hermens HJ, Freriks B, Disselhorst-Klug C, Rau G. Development of recommendations for SEMG sensors and sensor placement procedures. J Electromyogr Kinesiol. 2000; 10(5): 361-74.

265. Rainoldi A, Melchiorri G, Caruso I. A method for positioning electrodes during surface EMG recordings in lower limb muscles. J Neurosci Methods. 2004; 134(1): 37-43.

266. Standards for reporting EMG data. J Electromyogr Kinesiol. 2016; 28(6): I-II.

267. Cowling EJ, Steele JR, McNair PJ. Effect of verbal instructions on muscle activity and risk of injury to the anterior cruciate ligament during landing. Br J Sports Med. 2003; 37(2): 126-30.

References

351

268. Russell PJ, Croce RV, Swartz EE, Decoster LC. Knee-muscle activation during landings: Developmental and gender comparisons. Med Sci Sports Exerc. 2007; 39(1): 159-69.

269. Donnelly CJ, Elliott BC, Doyle TLA, Finch CF, Dempsey AR, Lloyd DG. Changes in muscle activation following balance and technique training and a season of Australian football. J Sci Med Sport. 2015; 18(3): 348-52.

270. Cram JR, Kasman GS, Holtz J. Introduction to Surface Electromyography. Aspen Publishers, Inc.: Gaithersburg; 1998.

271. Soderberg GL, Knutson LM. A guide for use and interpretation of kinesiologic electromyographic data. Phys Ther. 2000; 80(5): 485-98.

272. Chowdhury RH, Reaz MBI, Ali MABM, Bakar AAA, Chellappan K, Chang TG. Surface electromyography signal processing and classification techniques. Sensors. 2013; 13(9): 12432-66.

273. Kadaba MP, Ramakrishnan HK, Wootten ME, Gainey J, Gorton G, Cochran GV. Repeatability of kinematic, kinetic, and electromyographic data in normal adult gait. J Orthop Res. 1989; 7(6): 849-60.

274. Chapman A, Vicenzino B, Blanch P, Hodges P. Do differences in muscle recruitment between novice and elite cyclists reflect different movement patterns or less skilled muscle recruitment? J Sci Med Sport. 2009; 12(1): 31-4.

275. Fauth ML, Petushek EJ, Feldmann CR, Hsu BE, Garceau LR, Lutsch BN, Ebben WP. Reliability of surface electromyography during maximal voluntary isometric contractions, jump landings, and cutting. J Strength Cond Res. 2010; 24(4): 1131-7.

276. Hubley-Kozey CL, Robbins SM, Rutherford DJ, Stanish WD. Reliability of surface electromyographic recordings during walking in individuals with knee osteoarthritis. J Electromyogr Kinesiol. 2013; 23(2): 334-41.

277. Murley GS, Menz HB, Landorf KB, Bird AR. Reliability of lower limb electromyography during overground walking: a comparison of maximal- and sub-maximal normalisation techniques. J Biomech. 2010; 43(4): 749-56.

278. Bogey R, Cerny K, Mohammed O. Repeatability of wire and surface electrodes in gait. Am J Phys Med Rehabil. 2003; 82(5): 338-44.

References

352

279. Reaz MBI, Hussain MS, Mohd-Yasin F. Techniques of EMG signal analysis: Detection, processing, classification and applications. Biol Proced. 2006; 8(1): 11-35.

280. Homan KJ, Norcross MF, Goerger BM, Prentice WE, Blackburn JT. The influence of hip strength on gluteal activity and lower extremity kinematics. J Electromyogr Kinesiol. 2013; 23(2): 411-5.

281. Carcia CR, Martin RL. The influence of gender on gluteus medius activity during a drop jump. Phys Ther Sport. 2007; 8(4): 169-76.

282. Palmieri-Smith RM, McLean SG, Ashton-Miller JA, Wojtys EM. Association of quadriceps and hamstrings cocontraction patterns with knee joint loading. J Athl Train. 2009; 44(3): 256-63.

283. Perotto AO. Anatomical guide for the electromyographer: The limbs and trunk. Charles C. Thomas: Springfield, IL; 1994.

284. Edwards S, Steele JR, Cook JL, Purdam C, McGhee DE. Lower limb movement symmetry cannot be assumed when investigating the stop-jump landing. Med Sci Sports Exerc. 2012; 44(6): 1123-30.

285. Edwards S, Steele JR, Purdam CR, Cook JL, McGhee DE. Alterations to landing technique and patellar tendon loading in response to fatigue. Med Sci Sports Exerc. 2014; 46(2): 330-40.

286. Ambegaonkar JP, Shultz SJ, Perrin DH. A subsequent movement alters lower extremity muscle activity and kinetics in drop jumps vs. drop landings. J Strength Cond Res. 2011; 25(10): 2781-8.

287. Urabe Y, Kobayashi R, Sumida S, Tanaka K, Yoshida N, Nishiwaki GA, Tsutsumi E, Ochi M. Electromyographic analysis of the knee during jump landing in male and female athletes. Knee. 2005; 12(2): 129-34.

288. Lepley LK, Thomas AC, McLean SG, Palmieri-Smith RM. Fatigue's lack of effect on thigh-muscle activity in anterior cruciate ligament-reconstructed patients during a dynamic-landing task. J Sport Rehabil. 2013; 22(2): 83-92.

289. Tsai LC, McLean S, Colletti PM, Powers CM. Greater muscle co-contraction results in increased tibiofemoral compressive forces in females who have undergone anterior cruciate ligament reconstruction. J Orthop Res. 2012; 30(12): 2007-14.

References

353

290. Kipp K, Pfeiffer R, Sabick M, Harris C, Sutter J, Kuhlman S, Shea K. Muscle synergies during a single-leg drop-landing in boys and girls. J Appl Biomech. 2014; 30(2): 262-8.

291. Cappellini G, Ivanenko YP, Poppele RE, Lacquaniti F. Motor patterns in human walking and running. J Neurophysiol. 2006; 95(6): 3426-37.

292. Ivanenko YP, Cappellini G, Dominici N, Poppele RE, Lacquaniti F. Coordination of locomotion with voluntary movements in humans. J Neurosci. 2005; 25(31): 7238-53.

293. Ivanenko YP, Poppele RE, Lacquaniti F. Five basic muscle activation patterns account for muscle activity during human locomotion. J Physiol. 2004; 556(1): 267-82.

294. Barroso FO, Torricelli D, Moreno JC, Taylor J, Gomez-Soriano J, Bravo-Esteban E, Piazza S, Santos C, Pons JL. Shared muscle synergies in human walking and cycling. J Neurophysiol. 2014; 112(8): 1984-98.

295. Hug F. Can muscle coordination be precisely studied by surface electromyography? J Electromyogr Kinesiol. 2011; 21(1): 1-12.

296. Serrancoli G, Monllau JC, Font-Llagunes JM. Analysis of muscle synergies and activation-deactivation patterns in subjects with anterior cruciate ligament deficiency during walking. Clin Biomech. 2016; 31(1): 65-73.

297. Tresch MC, Cheung VC, d'Avella A. Matrix factorization algorithms for the identification of muscle synergies: evaluation on simulated and experimental data sets. J Neurophysiol. 2006; 95(4): 2199-212.

298. Lee DD, Seung HS. Algorithms for nonnegative matrix factorization. Advances in Neural Processing Systems; 2001; Cambridge, MA.

299. Beaulieu ML, McLean SG. Sex-dimorphic landing mechanics and their role within the noncontact ACL injury mechanism: Evidence, limitations and directions. Sports Med Arthrosc Rehabil Ther Tech. 2012; 4(1): 10-22.

300. McLean SG, Felin RE, Suedekum N, Calabrese G, Passerallo A, Joy S. Impact of fatigue on gender-based high-risk landing strategies. Med Sci Sports Exerc. 2007; 39(3): 502-14.

301. McLean SG, van den Bogert AJ. Knee kinematics during sidestep cutting: Potential for ACL injury in females. Clin Biomech. 2001; 16(10): 954-5.

References

354

302. Besier TF, Lloyd DG, Ackland TR, Cochrane JL. Anticipatory effects on knee joint loading during running and cutting maneuvers. Med Sci Sports Exerc. 2001; 33(7): 1176-81.

303. Besier TF, Lloyd DG, Cochrane JL, Ackland TR. External loading of the knee joint during running and cutting maneuvers. Med Sci Sports Exerc. 2001; 33(7): 1168-75.

304. Kristianslund E, Faul O, Bahr R, Myklebust G, Krosshaug T. Sidestep cutting technique and knee abduction loading: Implications for ACL prevention exercises. Br J Sports Med. 2014; 48(9): 779-83.

305. McLean SG, Su A, van den Bogert AJ. Development and validation of a 3-D model to predict knee joint loading during dynamic movement. J Biomech Eng. 2003; 125(6): 864-74.

306. Bell AL, Pedersen DR, Brand RA. A comparison of the accuracy of several hip center location prediction methods. J Biomech. 1990; 23(6): 617-21.

307. van den Bogert AJ, Smith GD, Nigg BM. In vivo determination of the anatomical axes of the ankle joint complex: An optimization approach. J Biomech. 1994; 27(12): 1477-88.

308. Vaughan CL, Davis BL, O'Connor JC. Dynamics of Human Gait. Human Kinetics: Champaign, IL; 1990.

309. McLean SG, Huang X, Su A, van den Bogert AJ. Sagittal plane biomechanics cannot injure the ACL during sidestep cutting. Clin Biomech. 2004; 19(8): 828-38.

310. McLean SG, Samorezov JE. Fatigue-induced ACL injury risk stems from a degradation in central control. Med Sci Sports Exerc. 2009; 41(8): 1661-72.

311. Borotikar BS, Newcomer R, Koppes R, McLean SG. Combined effects of fatigue and decision making on female lower limb landing postures: Central and peripheral contributions to ACL injury risk. Clin Biomech. 2008; 23(1): 81-92.

312. Leardini A, Chiari L, Croce UD, Cappozzo A. Human movement analysis using stereophotogrammetry, Part 3: Soft tissue artifact assessment and compensation. Gait Posture. 2005; 21(2): 212-25.

References

355

313. Wearing SC, Urry SR, Smeathers JE. The effect of visual targeting on ground reaction force and temporospatial parameters of gait. Clin Biomech. 2000; 15(8): 583-91.

314. Verniba D, Vergara ME, Gage WH. Force plate targeting has no effect on spatiotemporal gait measures and their variability in young and healthy population. Gait Posture. 2015; 41(2): 551-6.

315. Kristianslund E, Krosshaug T, van den Bogert AJ. Effect of low pass filtering on joint moments from inverse dynamics: Implications for injury prevention. J Biomech. 2012; 45(4): 666-71.

316. Chiari L, Croce UD, Leardini A, Cappozzo A. Human movement analysis using stereophotogrammetry, Part 2: Instrumental errors. Gait Posture. 2005; 21(2): 197-211.

317. Roewer BD, Ford KR, Myer GD, Hewett TE. The 'impact' of force filtering cut-off frequency on the peak knee abduction moment during landing: Artefact or 'artifiction'? Br J Sports Med. 2014; 48(6): 464-8.

318. Kristianslund E, Krosshaug T, Kam-Ming M, McLean S, van den Bogert AJ. Expressing the joint moments of drop jumps and sidestep cutting in different reference frames - does it matter? J Biomech. 2014; 47(1): 193-9.

319. Lee MJC, Lloyd DG, Lay BS, Bourke PD, Alderson JA. Effects of different visual stimuli on postures and knee moments during sidestepping. Med Sci Sports Exerc. 2013; 45(9): 1740-8.

320. Kristianslund E, Krosshaug T, van den Bogert AJ. Artefacts in measuring joint moments may lead to incorrect clinical conclusions: The nexus between science (biomechanics) and sports injury prevention! Br J Sports Med. 2013; 47(8): 470-3.

321. Woltring HJ. A Fortran package for generalized, cross-validatory spline smoothing and differentiation. Adv Eng Softw. 1986; 8(2): 104-13.

322. Yu B, Gabriel D, Noble L, An KN. Estimate of the optimum cutoff frequency for the Butterworth low-pass digital filter. J Appl Biomech. 1999; 15(3): 318-29.

323. Schache AG, Baker R. On the expression of joint moments during gait. Gait Posture. 2007; 25(3): 440-52.

References

356

324. Moisio KC, Sumner DR, Shott S, Hurwitz DE. Normalization of joint moments during gait: A comparison of two techniques. J Biomech. 2003; 36(4): 599-603.

325. Wannop JW, Worobets JT, Stefanyshyn DJ. Normalization of ground reaction forces, joint moments, and free moments in human locomotion. J Appl Biomech. 2012; 28(6): 665-76.

326. Deluzio KJ, Astephen JL. Biomechanical features of gait waveform data associated with knee osteoarthritis: An application of principal component analysis. Gait Posture. 2007; 25(1): 86-93.

327. Hubley-Kozey CL, Deluzio KJ, Landry SC, McNutt JS, Stanish WD. Neuromuscular alterations during walking in persons with moderate knee osteoarthritis. J Electromyogr Kinesiol. 2006; 16(4): 365-78.

328. Landry SC, McKean KA, Hubley-Kozey CL, Stanish WD, Deluzio KJ. Neuromuscular and lower limb biomechanical differences exist between male and female elite adolescent soccer players during an unanticipated run and crosscut maneuver. Am J Sports Med. 2007; 35(11): 1901-11.

329. Landry SC, McKean KA, Hubley-Kozey CL, Stanish WD, Deluzio KJ. Neuromuscular and lower limb biomechanical differences exist between male and female elite adolescent soccer players during an unanticipated side-cut maneuver. Am J Sports Med. 2007; 35(11): 1888-900.

330. Rutherford DJ, Hubley-Kozey CL, Stanish WD. Changes in knee joint muscle activation patterns during walking associated with increased structural severity in knee osteoarthritis. J Electromyogr Kinesiol. 2013; 23(3): 704-11.

331. Phinyomark A, Osis S, Hettinga BA, Ferber R. Kinematic gait patterns in healthy runners: A hierarchical cluster analysis. J Biomech. 2015; 48(14): 3897-904.

332. Phinyomark A, Hettinga BA, Osis S, Ferber R. Do intermediate- and higher-order principal components contain useful information to detect subtle changes in lower extremity biomechanics during running? Hum Movement Sci. 2015; 44: 91-101.

333. Donoghue OA, Harrison AJ, Coffey N, Hayes K. Functional data analysis of running kinematics in chronic Achilles tendon injury. Med Sci Sports Exerc. 2008; 40(7): 1323-35.

References

357

334. Foch E, Milner CE. The influence of iliotibial band syndrome history on running biomechanics examined via principal components analysis. J Biomech. 2014; 47(1): 81-6.

335. Astephen JL, Deluzio KJ. Changes in frontal plane dynamics and the loading response phase of the gait cycle are characteristic of severe knee osteoarthritis application of a multidimensional analysis technique. Clin Biomech. 2005; 20(2): 209-17.

336. Landry SC, McKean KA, Hubley-Kozey CL, Stanish WD, Deluzio KJ. Gender differences exist in neuromuscular control patterns during the pre-contact and early stance phase of an unanticipated side-cut and cross-cut maneuver in 15-18 years old adolescent soccer players. J Electromyogr Kinesiol. 2009; 19(5): e370-e9.

337. Astephen JL, Deluzio KJ, Caldwell GE, Dunbar MJ, Hubley-Kozey CL. Gait and neuromuscular pattern changes are associated with differences in knee osteoarthritis severity levels. J Biomech. 2008; 41(4): 868-76.

338. Richter C, O'Connor NE, Marshall B, Moran K. Comparison of discrete-point vs. dimensionality-reduction techniques for describing performance-related aspects of maximal vertical jumping. J Biomech. 2014; 47(12): 3012-7.

339. Donà G, Preatoni E, Cobelli C, Rodano R, Harrison AJ. Application of functional principal component analysis in race walking: An emerging methodology. Sports Biomech. 2009; 8(4): 284-301.

340. Neal RJ, Sydney-Smith M. The effects of footfall pattern and passing height on ground reaction forces in netball. Aust J Sci Med Sport. 1992; 24(3): 73-8.

341. O'Connor KM, Monteiro SK, Hoelker IA. Comparison of selected lateral cutting activities used to assess ACL injury risk. J Appl Biomech. 2009; 25(1): 9-21.

342. Bates BT. Single-subject methodology: An alternative approach. Med Sci Sports Exerc. 1996; 28(5): 631-8.

343. Richter C. The search for performance related factors in biomechanics. Dublin City University: Dublin, Ireland; 2014.

344. Brown C, Bowser B, Simpson KJ. Movement variability during single leg jump landings in individuals with and without chronic ankle instability. Clin Biomech. 2012; 27(1): 52-63.

References

358

345. Brown CN, Padua DA, Marshall SW, Guskiewicz KM. Variability of motion in individuals with mechanical or functional ankle instability during a stop jump maneuver. Clin Biomech. 2009; 24(9): 762-8.

346. Kipp K, Palmieri-Smith RM. Principal component based analysis of biomechanical inter-trial variability in individuals with chronic ankle instability. Clin Biomech. 2012; 27(7): 706-10.

347. Pollard CD, Heiderscheit BC, van Emmerik REA, Hamill J. Gender differences in lower extremity coupling variability during an unanticipated cutting maneuver. J Appl Biomech. 2005; 21(2): 143-52.

348. Richter C, O'Connor NE, Marshall B, Moran K. Clustering vertical ground reaction force curves produced during countermovement jumps. J Biomech. 2014; 47(10): 2385-90.

349. Carriero A, Zavatsky A, Stebbins J, Theologis T, Shefelbine SJ. Determination of gait patterns in children with spastic diplegic cerebral palsy using principal components. Gait Posture. 2009; 29(1): 71-5.

350. O'Byrne JM, Jenkinson A, O'Brien TM. Quantitative analysis and classification of gait patterns in cerebral palsy using a three-dimensional motion analyzer. J Child Neurol. 1998; 13(3): 101-8.

351. Toro B, Nester CJ, Farren PC. Cluster analysis for the extraction of sagittal gait patterns in children with cerebral palsy. Gait Posture. 2007; 25(2): 157-65.

352. O'Malley MJ, Abel MF, Damiano DL, Vaughan CL. Fuzzy clustering of children with cerebral palsy based on temporal-distance gait parameters. IEEE Trans Rehabil Eng. 1997; 5(4): 300-9.

353. Staudenmann D, Kingma I, Daffertshofer A, Stegeman DF, van Dieën JH. Heterogeneity of muscle activation in relation to force direction: A multi-channel surface electromyography study on the triceps surae muscle. J Electromyogr Kinesiol. 2009; 19(5): 882-95.

354. Cheung VCK, Piron L, Agostini M, Silvoni S, Turolla A, Bizzi E. Stability of muscle synergies for voluntary actions after cortical stroke in humans. Proc Natl Acad Sci USA. 2009; 106(46): 19563-8.

355. Daszykowski M, Walczak B, Massart DL. Looking for natural patterns in data. Part 1: Density-based approach. Chemometr Intell Lab. 2001; 56(2): 83-92.

References

359

356. Karami A, Johansson R. Choosing DBSCAN parameters automatically using differential evolution. Int J Comput Appl. 2014; 91(7).

357. Ester M, Kriegel HP, Sander J, Xu X. A density-based algorithm for discovering clusters in large spatial databases with noise. 2nd International Conference on Knowledge Discovery and Data Mining; 1996; Portland, OR, United States.

358. Pappas E, Shiyko MP, Ford KR, Myer GD, Hewett TE. Biomechanical deficit profiles associated with ACL injury risk in female athletes. Med Sci Sports Exerc. 2016; 48(1): 107-13.

359. Han J, Kamber M, Pei J. Data Mining: Concepts and Techniques. Morgan Kaufmann: Burlington, MA; 2006.

360. Fraley C, Raftery AE. How many clusters? Which clustering method? Answers via model-based cluster analysis. Comput J. 1998; 41(8): 578-88.

361. Pataky TC. Generalized n-dimensional biomechanical field analysis using statistical parametric mapping. J Biomech. 2010; 43(10): 1976-82.

362. Pataky TC, Robinson MA, Vanrenterghem J. Vector field statistical analysis of kinematic and force trajectories. J Biomech. 2013; 46(14): 2394-401.

363. Friston KJ, Ashburner JT, Kiebel SJ, Nichols TE, Penny WD. Statistical Parametric Mapping: The Analysis of Functional Brain Images. Elsevier Academic Press: Amsterdam; 2007.

364. De Ridder R, Willems T, Vanrenterghem J, Robinson MA, Roosen P. Lower limb landing biomechanics in subjects with chronic ankle instability. Med Sci Sports Exerc. 2015; 47(6): 1225-31.

365. De Ridder R, Willems T, Vanrenterghem J, Robinson MA, Palmans T, Roosen P. Multi-segment foot landing kinematics in subjects with chronic ankle instability. Clin Biomech. 2015; 30(6): 585-92.

366. Robinson MA, Donnelly C, Vanrenterghem J, Pataky T. Sagittal plane knee kinematics predict non-sagittal knee joint moments in unplanned sidestepping. XXV Congress of the International Society of Biomechanics; 2015; Glasgow, Scotland.

367. Robinson MA, Vanrenterghem J, Pataky TC. Statistical Parametric Mapping (SPM) for alpha-based statistical analyses of multi-muscle EMG time-series. J Electromyogr Kinesiol. 2015; 25(1): 14-9.

References

360

368. Adler RJ, Taylor JE. Random Fields and Geometry. Springer-Verlag: New York; 2007.

369. Pataky TC. One-dimensional statistical parametric mapping in Python. Comput Methods Biomech Biomed Engin. 2012; 15(3): 295-301.

370. Dingenen B, Malfait B, Vanrenterghem J, Robinson MA, Verschueren SMP, Staes FF. Can two-dimensional measured peak sagittal plane excursions during drop vertical jumps help identify three-dimensional measured joint moments? Knee. 2015; 22(2): 73-9.

371. Pataky T. SPM1D. 2015. Available from: spm1d.org.

372. Donnelly CJ, Lloyd DG, Elliott BC, Reinbolt JA. Optimizing whole-body kinematics to minimize valgus knee loading during sidestepping: Implications for ACL injury risk. J Biomech. 2012; 45(8): 1491-7.

373. Bernstein N. Coordination and regulation of movements. Pergamon: New York; 1967.

374. Houck JR, Duncan A, Haven KED. Comparison of frontal plane trunk kinematics and hip and knee moments during anticipated and unanticipated walking and side step cutting tasks. Gait Posture. 2006; 24(3): 314-22.

375. Pollard CD, Sigward SM, Powers CM. Gender differences in hip joint kinematics and kinetics during side-step cutting maneuver. Clin J Sport Med. 2007; 17(1): 38-42.

376. Kaila R. Influence of modern studded and bladed soccer boots and sidestep cutting on knee loading during match play conditions. Am J Sports Med. 2007; 35(9): 1528-36.

377. Sheehan FT, Sipprell WH, Boden BP. Dynamic sagittal plane trunk control during anterior cruciate ligament injury. Am J Sports Med. 2012; 40(5): 1068-74.

378. McLean SG, Neal RJ, Myers PT, Walters MR. Knee joint kinematics during the sidestep cutting maneuver: Potential for injury in women. Med Sci Sports Exerc. 1999; 31(7): 959-68.

379. Thompson JA, Chaudhari AMW, Schmitt LC, Best TM, Siston RA. Gluteus maximus and soleus compensate for simulated quadriceps atrophy and activation failure during walking. J Biomech. 2013; 46(13): 2165-72.

References

361

380. Reinbolt JA, Fox MD, Arnold AS, Õunpuu S, Delp SL. Importance of preswing rectus femoris activity in stiff-knee gait. J Biomech. 2008; 41(11): 2362-9.

381. Piazza SJ, Delp SL. The influence of muscles on knee flexion during the swing phase of gait. J Biomech. 1996; 29(6): 723-33.

382. Zebis MK, Bencke J, Andersen LL, Alkjær T, Suetta C, Mortensen P, Kjær M, Aagaard P. Acute fatigue impairs neuromuscular activity of anterior cruciate ligament-agonist muscles in female team handball players. Scand J Med Sci Sports. 2011; 21(6): 833-40.

383. Miles J, Shevlin M. Applying Regression and Correlation: A Guide for Students and Researchers. Sage Academic Press: London, United Kingdom; 2001.

384. Cohen J. Statistical Power Analysis for the Behavioral Sciences. 2nd ed. Lawrence Erlbaum Associates: Hillsdale, NJ; 1988.

385. Lloyd DG, Buchanan TS. A model of load sharing between muscles and soft tissues at the human knee during static tasks. J Biomech Eng. 1996; 118(3): 367-76.

386. Lloyd DG, Buchanan TS, Besier TF. Neuromuscular biomechanical modeling to understand knee ligament loading. Med Sci Sports Exerc. 2005; 37(11): 1939-47.

387. Claiborne TL, Armstrong CW, Gandhi V, Pincivero DM. Relationship between hip and knee strength and knee valgus during a single leg squat. J Appl Biomech. 2006; 22(1): 41-50.

388. Müller S, Brenton J, Dempsey AR, Harbaugh AG, Reid C. Individual differences in highly skilled visual perceptual-motor striking skill. Atten Percept Psychophys. 2015; 77(5): 1726-36.

389. Fleming BC, Renstrom PA, Ohlen G, Johnson RJ, Peura GD, Beynnon BD, Badger GJ. The gastrocnemius muscle is an antagonist of the anterior cruciate ligament. J Orthop Res. 2001; 19(6): 1178-84.

390. Shin CS, Chaudhari AM, Andriacchi TP. The influence of deceleration forces on ACL strain during single-leg landing: A simulation study. J Biomech. 2007; 40(5): 1145-52.

References

362

391. Celebrini RG, Eng JJ, Miller WC, Ekegren CL, Johnston JD, MacIntyre DL. The effect of a novel movement strategy in decreasing ACL risk factors in female adolescent soccer players. J Strength Cond Res. 2012; 26(12): 3406-17.

392. Bien DP. Rationale and implementation of anterior cruciate ligament injury prevention warm-up programs in female athletes. J Strength Cond Res. 2011; 25(1): 271-85.

393. Paterno MV, Myer GD, Ford KR, Hewett TE. Neuromuscular training improves single-limb stability in young female athletes. J Orthop Sports Phys Ther. 2004; 34(6): 305-16.

394. Glover TA, Albers CA. Considerations for evaluating universal screening assessments. J School Psychol. 2007; 45(2): 117-35.

395. Devillé WL, Buntinx F, Bouter LM, Montori VM, De Vet HCW, Van der Windt DAWM, Bezemer PD. Conducting systematic reviews of diagnostic studies: Didactic guidelines. BMC Med Res Methodol. 2002; 3: 2-9.

396. Barber-Westin SD, Smith ST, Campbell T, Noyes FR. The drop-jump video screening test: Retention of improvement in neuromuscular control in female volleyball players. J Strength Cond Res. 2010; 24(11): 3055-62.

397. Chorba RS, Chorba DJ, Bouillon LE, Overmyer CA, Landis JA. Use of a functional movement screening tool to determine injury risk in female collegiate athletes. N Am J Sports Phys Ther. 2010; 5(2): 47-54.

398. Dowling AV, Favre J, Andriacchi TP. A wearable system to assess risk for anterior cruciate ligament injury during jump landing: Measurements of temporal events, jump height, and sagittal plane kinematics. J Biomech Eng. 2011; 133(7): 071008.

399. Hamilton M, Velasquez JR. Ankle flexibility and jump landing mechanics: Implications for ACL injury risk. Int J Athl Ther Training. 2011; 16(6): 14-6.

400. Munro A, Herrington L, Carolan M. Reliability of 2-dimensional video assessment of frontal-plane dynamic knee valgus during common athletic screening tasks. J Sport Rehabil. 2012; 21(1): 7-11.

401. Nilstad A, Kristianslund E, Krosshaug T, Steffen K, Myklebust G, Bahr R, Andersen TE. A simple clinical screening test can identify elite female football players with increased knee valgus angles. Br J Sports Med. 2011; 45(4): 310.

References

363

402. Noyes FR, Barber-Westin SD, Fleckenstein C, Walsh C, West J. The drop-jump screening test: Difference in lower limb control by gender and effect of neuromuscular training in female athletes. Am J Sports Med. 2005; 33(2): 197-207.

403. Padua DA, Marshall SW, Beutler AI, DiStefano LJ, DiStefano MJ, de la Motte SJ, Goss D. The Landing Error Scoring System (LESS) prospectively identifies ACL injury. J Athl Train. 2010; 45(5): 539.

404. Steffen K, Nilstad A, Kristianslund E, Krosshaug T, Myklebust G, Bahr R. A comparison between the star excursion balance test and subjective assessment of knee stability in a single-leg squat. Br J Sports Med. 2011; 45(4): 375.

405. Brumitt J, Heiderscheit BC, Manske RC, Niemuth PE, Rauh MJ. Lower extremity functional tests and risk of injury in Division III collegiate athletes. Int J Sports Phys Ther. 2013; 8(3): 216-27.

406. Jöllenbeck T, Neuhaus D, Beck K, Wojtowicz S, Röckel M. Screening test for the potential risk of ACL rupture of female and male soccer players. 28th International Symposium on Biomechanics in Sports; 2010; Marquette, United States.

407. Markbreiter JG, Sagon BK, McLeod TCV, McCarty CW. Clinician scoring of the Landing Error Scoring System is reliable to assess jump-landing movement patterns. J Sport Rehabil. 2015; 24(2): 214-8.

408. Petushek EJ, Cokely ET, Ward P, Durocher JJ, Wallace SJ, Myer GD. Injury risk estimation expertise: Assessing the ACL injury risk estimation quiz. Am J Sports Med. 2015; 43(7): 1640-7.

409. Cortes N, Onate J. Clinical assessment of drop-jump landing for determination of risk for knee injury. Int J Athl Ther Training. 2013; 18(3): 10-3.

410. Padua DA, Boling MC, DiStefano LJ, Onate JA, Beutler AI, Marshall SW. Reliability of the Landing Error Scoring System-real time, a clinical assessment tool of jump-landing biomechanics. J Sport Rehabil. 2011; 20(2): 145-56.

411. Myer GD, Khoury J, Succop P, Hewett TE, Goetschius J, Smith HC, Vacek PM, Holterman LA, Tourville TW, Slauterbeck J, Johnson RJ, Beynnon BD, Shultz SJ. Clinic-based algorithm to identify female athletes at risk for anterior cruciate ligament injury: Letter to the editor. Am J Sports Med. 2013; 41(1): NP1-NP6.

References

364

412. Goetschius J, Smith HC, Vacek PM, Holterman LA, Shultz SJ, Tourville TW, Slauterbeck J, Johnson RJ, Beynnon BD. Application of a clinic-based algorithm as a tool to identify female athletes at risk for anterior cruciate ligament injury: A prospective cohort study with a nested, matched case-control analysis. Am J Sports Med. 2012; 40(9): 1978-84.

413. Stroube BW, Myer GD, Brent JL, Ford KR, Heidt Jr RS, Hewett TE. Effects of task-specific augmented feedback on deficit modification during performance of the tuck-jump exercise. J Sport Rehabil. 2013; 22(1): 7-18.

414. Dudley LA, Smith CA, Olson BK, Chimera NJ, Schmitz B, Warren M. Interrater and intrarater reliability of the tuck jump assessment by health professionals of varied educational backgrounds. J Sports Med. 2013; 2013: 483503.

415. Myer GD, Brent JL, Ford KR, Hewett TE. Real-time assessment and neuromuscular training feedback techniques to prevent anterior cruciate ligament injury in female athletes. Strength Cond J. 2011; 33(3): 21-35.

416. Ekegren C, Donaldson A, Gabbe B, Finch C. Implementing injury surveillance systems alongside injury prevention programs: Evaluation of an online surveillance system in a community setting. Inj Epidemiol. 2014; 1(19): 1-15.

417. Myer GD, Ford KR, Barber Foss KD, Chunyan L, Nick TG, Hewett TE. The relationship of hamstrings and quadriceps strength to anterior cruciate ligament injury in female athletes. Clin J Sport Med. 2009; 19(1): 3-8.

418. Hewett TE, Myer GD, Ford KR. Decrease in neuromuscular control about the knee with maturation in female athletes. J Bone Joint Surg. 2004; 86(8): 1601-8.

419. Weinhandl JT, Earl-Boehm JE, Ebersole KT, Huddleston WE, Armstrong BSR, O'Connor KM. Reduced hamstring strength increases anterior cruciate ligament loading during anticipated sidestep cutting. Clin Biomech. 2014; 29(7): 752-9.

420. Quatman-Yates CC, Myer GD, Ford KR, Hewett TE. A longitudinal evaluation of maturational effects on lower extremity strength in female adolescent athletes. Pediatr Phys Ther. 2013; 25(3): 271-6.

421. Frank B, Bell DR, Norcross MF, Blackburn JT, Goerger BM, Padua DA. Trunk and hip biomechanics influence anterior cruciate loading mechanisms in physically active participants. Am J Sports Med. 2013; 41(11): 2676-83.

References

365

422. Ageberg E, Bennell KL, Hunt MA, Simic M, Roos EM, Creaby MW. Validity and inter-rater reliability of medio-lateral knee motion observed during a single-limb mini squat. BMC Musc Disord. 2010; 11: 265.

423. Whatman C, Hume P, Hing W. The reliability and validity of physiotherapist visual rating of dynamic pelvis and knee alignment in young athletes. Phys Ther Sport. 2013; 14(3): 168-74.

424. Myer GD, Ford KR, Hewett TE. Rationale and clinical techniques for anterior cruciate ligament injury prevention among female athletes. J Athl Train. 2004; 39(4): 352-.

425. Sigward SM, Powers CM. Loading characteristics of females exhibiting excessive valgus moments during cutting. Clin Biomech. 2007; 22(7): 827-33.

426. Theiss JL, Gerber JP, Cameron KL, Beutler AI, Marshall SW, Distefano LJ, Padua DA, de la Motte SJ, Miller JM, Yunker CA. Jump-landing differences between varsity, club, and intramural athletes: The jump-ACL study. J Strength Cond Res. 2014; 28(4): 1164-71.

427. Padua DA, DiStefano LJ, Marshall SW, Beutler AI, de la Motte SJ, DiStefano MJ. Retention of movement pattern changes after a lower extremity injury prevention program is affected by program duration. Am J Sports Med. 2012; 40(2): 300-6.

428. Bahr R. Why screening tests to predict injury do not work - and probably never will...: A critical review. Br J Sports Med. 2016; 50(13): 776-80.

429. Lee H, Sullivan SJ, Schneiders AG, Ahmed OH, Balasundaram AP, Williams D, Meeuwisse WH, McCrory P. Smartphone and tablet apps for concussion road warriors (team clinicians): A systematic review for practical users. Br J Sports Med. 2015; 49(8): 499-505.

430. van Mechelen DM, van Mechelen W, Verhagen EALM. Sports injury prevention in your pocket?! Prevention apps assessed against the available scientific evidence: A review. Br J Sports Med. 2014; 48(11): 871-7.

431. Fox AS, Bonacci J, McLean SG, Spittle M, Saunders N. A systematic evaluation of field-based screening methods for the assessment of anterior cruciate ligament (ACL) injury risk. Sports Med. 2016; 46(5): 715-35.

432. Waldén M, Atroshi I, Magnusson H, Wagner P, Hägglund M. Prevention of acute knee injuries in adolescent female football players: Cluster randomised controlled trial. Br Med J. 2012; 344: e3042.

References

366

433. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1986; i: 307-10.

434. Crossley KM, Zhang WJ, Schache AG, Bryant A, Cowan SM. Performance on the single-leg squat task indicates hip abductor muscle function. Am J Sports Med. 2011; 39(4): 866-73.

435. Willson JD, Davis IS. Utility of the frontal plane projection angle in females with patellofemoral pain. J Orthop Sports Phys Ther. 2008; 38(10): 606-15.

436. Tate J, Dale B, Baker C. Expert versus novice interrater and intrarater reliability of the frontal plane projection angle during a single-leg squat. Int J Athl Ther Training. 2015; 20(4): 23-7.

437. Whatman C, Hing W, Hume P. Physiotherapist agreement when visually rating movement quality during lower extremity functional screening tests. Phys Ther Sport. 2012; 13(2): 87-96.

438. Knudson D. What can professionals qualitatively analyze? J Phys Educ Recreation Dance. 2000; 71(2): 19-23.

439. Marshall BM, Franklyn-Miller AD, Moran KA, King EA, Strike SC, Falvey EC. Can a single-legged squat provide insight into movement control and loading during dynamic sporting actions in athletic groin pain patients? J Sport Rehabil. 2016; 25(2): 117-25.

440. Alenezi F, Herrington L, Jones P, Jones R. Relationships between lower limb biomechanics during single leg squat with running and cutting tasks: A preliminary investigation. Br J Sports Med. 2014; 48(7): 560-1.

441. Stensrud S, Myklebust G, Kristianslund E, Bahr R, Krosshaug T. Correlation between two-dimensional video analysis and subjective assessment in evaluating knee control among elite female team handball players. Br J Sports Med. 2011; 45(7): 589-95.

442. Mauntel TC, Frank BS, Begalle RL, Blackburn JT, Padua DA. Kinematic differences between those with and without medial knee displacement during a single-leg squat. J Appl Biomech. 2014; 30(6): 707-12.

443. Jones D, Tillman SM, Tofte K, Mizner RL, Greenberg S, Moser MW, Chmielewski TL. Observational ratings of frontal plane knee position are related to the frontal plane projection angle but not the knee abduction angle during a step-down task. J Orthop Sports Phys Ther. 2014; 44(12): 973-8.

References

367

444. Weinhandl JT, Earl-Boehm JE, Ebersole KT, Huddleston WE, Armstrong BSR, O'Connor KM. Anticipatory effects on anterior cruciate ligament loading during sidestep cutting. Clin Biomech. 2013; 28(6): 655-63.

445. Hewett TE, Roewer B, Ford K, Myer G. Multicenter trial of motion analysis for injury risk prediction: Lessons learned from prospective longitudinal large cohort combined biomechanical - epidemiological studies. Braz J Phys Ther. 2015; 19(5): 398-409.

446. Jamison ST, Pan X, Chaudhari AMW. Knee moments during run-to-cut maneuvers are associated with lateral trunk positioning. J Biomech. 2012; 45(11): 1881-5.

368

Appendices

Appendix A Statement of Ethical Approval ................................................ 369

Appendix B Plain Language Statement and Consent ................................ 370

Appendix C Injury History Questionnaire .................................................. 378

Appendix D Down 2 Earth (D2E) Player Booklet ....................................... 379

Appendix E Maximal Voluntary Isometric Contraction (MVIC)

Procedures. Adapted from Konrad.[263]. ................................. 390

Appendix F Approach Speed Data Analysis Results ................................ 391

Appendix G Descriptions of the Principal Patterns Extracted from the

Netball-Specific Landing Task used in Chapter Six ............. 396

Appendix H Methodological Quality Assessment for Studies Included in

Chapter Seven ............................................................................ 415

Appendix I Full Results of Intra- and Inter-Rater Reliability Analyses for

the Landing Error Scoring System (LESS) and Tuck Jump

Assessment ................................................................................. 417

Appendix J Descriptions of the Principal Patterns Extracted from the

Single-Leg Squat Task used in Chapter Nine ....................... 419

Appendix K Descriptions of the Principal Patterns Extracted from the

Netball-Specific Landing Task used in Chapter Nine ......... 432

369

Appendix A Statement of Ethical Approval

370

Appendix B Plain Language Statement and Consent

PLAIN LANGUAGE STATEMENT AND CONSENT FORM To: Female netball players (aged 18 and above)

Plain Language Statement

Date:

Full Project Title: Neuromechanical control of the lower limb: Implications for ACL injury risk and prevention

Principal Researcher: Dr. Natalie Saunders Associate Researcher(s): Mr. Aaron Fox, Dr. Jason Bonacci, Dr. Scott McLean, Assoc. Prof. Michael Spittle Thank you for taking the time to read this Information Statement. This Participant Information Statement contains detailed information about the research project. Its purpose is to explain to you as clearly as possible all the procedures involved in the project before you decide whether or not to take part in it. Please read this information carefully. Feel free to ask questions about anything in the document, and discuss the project with relatives or friends. Once you understand what is involved in the project and if you agree to take part in it, you will be asked to sign the Consent Form. By signing the Consent Form, you indicate that you understand the information and that you give your consent to participate in the project. You will be given a copy of the Participant Information Sheet and Consent Form to keep as a record. Participation in this study is voluntary. If you do not wish to take part you are not obliged to. If you decide to take part and later change your mind, you are free to withdraw from the project at any stage. Your decision whether or not to participate will not affect your relationship with Deakin University.

Appendix B

371

1. What is the reason for this project?

The anterior cruciate ligament (ACL) is one of the major stabilising ligaments of the knee. Sports-related ACL injuries are a traumatic experience and often occur in the absence of contact with another player (i.e. noncontact). Sports that require frequent changes of direction, landing from jumps and/or rapid deceleration (such as netball) often incur the greatest incidence of noncontact ACL injuries. While this type of injury is not exclusive to a specific population, females incur a higher rate of injury in comparison to male counterparts. Despite extensive attempts, the mechanisms responsible for ACL injury remain elusive. Recently, research has begun to focus on how the muscles of the lower limb control the knee during sporting tasks, and whether dysfunctional patterns of muscle activation may influence ACL injury risk. Identifying these patterns that increase the risk of injury will aid in further explaining the mechanisms responsible for ACL injuries. An important aspect of injury prevention is the ability to identify athletes who are at-risk of future injury. With regard to ACL injuries, various clinical screening tools have been designed that attempt to identify those at-risk for future ACL injury. These screening tools are relatively new and further examination to confirm their accuracy and sensitivity for use in the wider community setting is required. In addition, injury intervention programs are also an essential aspect of injury prevention. It is possible that different individuals do not respond similarly to the same intervention program. Gaining a greater understanding of individual responses will aid in the design of targeted intervention programs that maximise the benefits for all athletes. 2. What will I be asked to do?

When you accept to participate in this study, you will be screened by one of the investigators to ensure that you meet all the criteria as a participant in this study.

You will be invited to participate in this study if: You are currently participating in competitive netball and complete at least

one training session for this per week. You are above 18 years of age. You do not currently have an injury to either lower limb. You have no previous history of a serious knee injury or lower limb surgery. You are free from any neuromuscular or musculoskeletal disorders that

affect the lower limb. Participation in this study will involve a familiarisation session and two laboratory testing sessions, and may require you to participate in a six-week injury intervention program (see below for details).

Appendix B

372

Location All testing sessions will be conducted at Deakin University (Waurn Ponds campus) Biomechanics Laboratory. Testing Sessions Participation will involve attending the Deakin University Biomechanics Laboratory at the Waurn Ponds campus on three separate occasions for one familiarisation and two testing sessions. At the familiarisation session, all tasks and analysis techniques being performed in the subsequent testing sessions will be described to you and you will also be given the opportunity to ask any questions of the researchers. Individual demographics (i.e. age, height, weight) will be collected at both testing sessions. The two testing sessions will be identical and involve the performance of a range of tasks, those being (1) a series of maximum effort voluntary muscle contractions of the leg against resistance; (2) multiple trials of a netball landing task; (3) two ACL clinical screening tools involving jump-landing tasks; and (4) three athletic performance measures (see below for details). All tasks will be observed by a trained researcher and recorded using digital video cameras. Prior to actual testing you will be given time to familiarise yourself with and practice the tasks required of you. If at any stage during laboratory testing you begin to feel uncomfortable or feel you may be developing an injury you should instruct the researcher and testing will stop. During the performance of all tasks during the laboratory testing sessions you will be monitored using kinematic and electromyography (EMG) analysis techniques. Thirty-four reflective markers will be attached to your lower limbs via tape and tracked using high speed cameras for kinematic analysis. EMG measures the level of activation of specific muscles via wireless surface electrodes placed on the lower limbs. Prior to the placement of the surface electrodes on the skin, each attachment site will be shaven to remove any hair, followed by gentle abrasion and sterilisation of the area with alcohol swabs. All preparation and analysis techniques are completely painless and non-invasive.

The familiarisation sessions will last for approximately 30 minutes. You will attend and perform your two laboratory testing sessions at the same time as one of your team mates. The laboratory testing sessions will run for approximately two to two-and-a-half hours, with testing procedures/tasks interspersed over this time period.

Maximal Effort Voluntary Muscle Contractions At each testing sessions, five separate maximal voluntary muscle contraction tasks will be performed. The maximal contractions will be performed against a resistance band, and each of the five contraction tasks will be repeated three times. The performance of these tasks requires you to apply maximal effort against the applied

Appendix B

373

resistance for a period of five seconds. Netball Landing Task The netball landing task you will be required to perform at both testing sessions is one you would commonly perform during netball matches. During laboratory testing you will perform ten ‘leap’ landing trials on each leg. The leap landing involves a run-up and single leg landing on the opposite foot used for take-off. You will be required to catch a chest height pass while performing the land. ACL Clinical Screening Tools Two clinical screening tools for ACL injury risk will be performed at each of the testing sessions, those being the Landing Error Scoring System (LESS) and Tuck Jump Assessment. The LESS involves the performance of a jump-landing task off a 30-cm high box. You will be required to perform a horizontal jump off the box to a distance 50% of your height and then immediately rebound off the ground performing a maximal vertical jump. Six repetitions of this task will be undertaken during testing. The Tuck Jump Assessment involves the performance of repeated tuck jumps on a spot for a ten-second period. Tuck jumps involve pulling the knees up as high as possible during each jump. Two ten-second repetitions of this task will be undertaken during testing. Athletic Performance Measures Three athletic performance measures will be performed at both testing sessions, those being a vertical jump test, single leg squat test and cross-over hop for distance test. The vertical jump test is a standard measure of lower limb muscular power. Two types of vertical jump tests will be performed, those being a stationary and in-motion test. The stationary test will involve the performance of a vertical jump, reaching for maximum height with no run-up. The in-motion test is identical with the exception of a two-step run-up being allowed. You will be given three attempts for both tests to reach your maximal jump height. The single leg squat test will involve the performance of five single leg squat repetitions while being observed by a researcher. During the test you will be required to keep your arms crossed in front of your body at chest height. The five repetitions will be undertaken on both the left and right legs. The cross-over hop for distance test involves the performance of four diagonal hops with the goal of travelling the greatest distance forward from the starting point.

Appendix B

374

Each hop must cross over a 20 centimetre zone marked out on the floor. The test will be performed on both the left and right legs. You will be given three attempts for both legs to reach your maximal hop distance. Training Program A random selection of participants will also be required to participate in a six-week injury intervention program between the two laboratory testing sessions. The injury intervention program being used in this research is the ‘Down to Earth’ program, which aims to teach netball specific safe and effective landing technique, and is promoted as a coaching resource by Netball Australia. The injury intervention program involves the performance of balance, strength, plyometric and agility based exercises three times per week (one training and two home-based sessions). You will be required to fill out a training diary detailing when you completed each of the training sessions. While completing the training program you will also be required to fill in an activity diary detailing any physical activity you complete over the six-week period. This is to monitor for any activities that may have an influence on the results of the intervention program. NOTE: participants who do not complete the training program are also required to complete a physical activity diary over the six-week period between laboratory testing sessions. 3. What will I gain by participating?

By participating in this research you are contributing to an improved understanding of ACL injury risk and prevention within the sport of netball. You may also gain a greater understanding of your risk for future ACL injury, where researchers will be happy to provide you with any of your results (see Section 6) or ask any questions you may have. If required you may be provided with further information on how to reduce the chance of sustaining an ACL injury in the future. Participants who complete the injury intervention program may in fact reduce their risk of future injury via the training program. 4. Are there any risks involved?

There is minimal risk associated with participation in this research. As all jump-landing tasks have the ability to cause injury there is the risk of incurring a lower limb injury during laboratory testing. The sporting tasks that you will be required to perform are performed in a controlled laboratory environment and present the same, if not lower risk of injury than those performed in a normal netball game. The research team includes an accredited exercise physiologist and physiotherapist, and in the unlikely event of an injury occurring research staff will provide initial care. However, accessing ongoing care and any treatment relating to the injury will be at your expense.

Appendix B

375

5. Will the information I provide be kept private?

Any information obtained in connection with this research project that can identify you will remain confidential and will only be used for the purpose of this research project. Information will only be disclosed with your permission, except as required to by law. If you give us permission by signing the Consent Form, we plan to publish the results at conferences and in scientific journals. In any publication or presentation, your information will be provided in such a way that you cannot be identified. 6. Results of the Project

Results of this project will be available to you by mail or e-mail, at your request. 7. Further Information or Any Problems

If you require further information or if you have any problems concerning this project you can contact one of the principal researchers. The researchers responsible for this project are:

Dr. Natalie Saunders 9246-8284 Mr. Aaron Fox 9251-7783 or 0411-962-169

Complaints If you have any complaints about any aspect of the project, the way it is being conducted or any questions about your rights as a research participant, then you may contact:

The Manager, Research Integrity, Deakin University, 221 Burwood Highway, Burwood Victoria 3125, Telephone: 9251 7129, Facsimile: 9244 6581; [email protected]

Please quote project number [2013-007].

Appendix B

376


Consent Form

Date:


Reference Number: 2013-007

I have read and I understand the attached Plain Language Statement.

I freely agree to participate in this project according to the conditions in the Plain Language Statement.

I have been given a copy of the Plain Language Statement and Consent Form to keep.

The researcher has agreed not to reveal my identity and personal details, including where information about this project is published, or presented in any public form.

Participant’s Name (printed) ……………………………………………………………………

Signature ……………………………………………………… Date …………………………

Appendix B

377


Revocation of Consent Form

(To be used for participants who wish to withdraw from the project)

Date:


Reference Number: 2013-007

I hereby wish to WITHDRAW my consent to participate in the above research project and understand that such withdrawal WILL NOT jeopardise my relationship with Deakin University.

Participant’s Name (printed) ……………………………………………………. Signature ………………………………………………………………. Date …………………… Please mail this form to: Dr. Natalie Saunders Centre for Exercise and Sport Science School of Exercise and Nutrition Sciences Deakin University, Melbourne Campus 221 Burwood Hwy, Burwood, VIC 3125 Telephone: +61 3 9246-8284

378

App

endi

x C

In

jury

His

tory

Que

stio

nnai

re

Plea

se fi

ll in

you

r inj

ury

hist

ory

for t

he fo

llow

ing

body

sect

ions

: N

ame:

___

____

____

____

____

____

____

____

____

____

____

____

Pa

rtic

ipan

t ID:

___

____

____

____

_ Bo

dy

Sect

ion

Type

of i

njur

y

(e.g

. mus

cle st

rain

, sp

rain

, fra

ctur

e)

Mon

th &

yea

r of

you

r inj

ury

(e.g

. Nov

201

2)

Was

you

r inj

ury

sust

aine

d pl

ayin

g ne

tbal

l?

(YES

/ N

O)

Did

your

inju

ry p

reve

nt y

ou fr

om p

layi

ng

or tr

aini

ng (i

f YES

, app

rox.

for h

ow lo

ng?)

(e

.g. Y

ES –

2 tr

aini

ng se

ssio

ns)

Low

er

Back

…

……

……

……

……

……

. …

……

……

……

……

……

. …

……

……

……

……

……

.

……

……

……

……

. …

……

……

……

….

……

……

……

……

.

……

……

……

……

……

……

……

……

……

……

……

……

……

…

……

……

……

……

……

……

……

……

……

……

……

……

…

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

...

……

……

……

……

……

……

……

……

……

……

……

……

...

……

……

……

……

……

……

……

……

……

……

……

……

...

Hip

……

……

……

……

……

….

……

……

……

……

……

….

……

……

……

……

……

….

……

……

……

……

. …

……

……

……

….

……

……

……

……

.

……

……

……

……

……

……

……

……

……

……

……

……

……

…

……

……

……

……

……

……

……

……

……

……

……

……

…

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

...

……

……

……

……

……

……

……

……

……

……

……

……

...

……

……

……

……

……

……

……

……

……

……

……

……

...

Knee

…

……

……

……

……

……

. …

……

……

……

……

……

. …

……

……

……

……

……

.

……

……

……

……

. …

……

……

……

….

……

……

……

……

.

……

……

……

……

……

……

……

……

……

……

……

……

……

…

……

……

……

……

……

……

……

……

……

……

……

……

…

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

...

……

……

……

……

……

……

……

……

……

……

……

……

...

……

……

……

……

……

……

……

……

……

……

……

……

...

Ankl

e …

……

……

……

……

……

. …

……

……

……

……

……

. …

……

……

……

……

……

.

……

……

……

……

. …

……

……

……

….

……

……

……

……

.

……

……

……

……

……

……

……

……

……

……

……

……

……

…

……

……

……

……

……

……

……

……

……

……

……

……

…

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

...

……

……

……

……

……

……

……

……

……

……

……

……

...

……

……

……

……

……

……

……

……

……

……

……

……

...

Foot

…

……

……

……

……

……

. …

……

……

……

……

……

. …

……

……

……

……

……

.

……

……

……

……

. …

……

……

……

….

……

……

……

……

.

……

……

……

……

……

……

……

……

……

……

……

……

……

…

……

……

……

……

……

……

……

……

……

……

……

……

…

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

……

...

……

……

……

……

……

……

……

……

……

……

……

……

...

……

……

……

……

……

……

……

……

……

……

……

……

...

379

Appendix D Down 2 Earth (D2E) Player Booklet

Appendix D

380

Appendix D

381

Appendix D

382

Appendix D

383

Appendix D

384

Appendix D

385

Appendix D

386

Appendix D

387

Appendix D

388

Appendix D

389

390

Appendix E Maximal Voluntary Isometric

Contraction (MVIC) Procedures. Adapted from

Konrad.[263]

Muscles Tested MVIC Set Position Procedure

Rectus Femoris Vastus Medialis

Perform a single leg knee extension

against resistance between 70 and

90 of knee flexion.

Biceps Femoris Medial Hamstrings

Arrange a good fixation of the hip

and perform a unilateral knee

flexion against resistance at

approximately 20 and 30 of knee

flexion.

Gluteus Maximus

Perform a single leg hip extension

against resistance with the hip

positioned at approximately 20 of

hyperextension.

Gluteus Medius

Perform a single leg hip abduction

against resistance in a side laying

position.

Gastrocnemius

Perform a single leg plantar flexion

against very rigid resistance with

the knees in full extension and

ankle positioned at 90 .

391

Appendix F Approach Speed Data Analysis Results

One-dimensional statistical parametric mapping (SPM1D) canonical

correlations analysis and linear regression was used to examine whether

changes in approach speed influenced hip and knee joint rotations and

moments (see Section 3.5.2 within this thesis for a description of SPM1D

procedures). Approach speed was found to have a statistically significant

effect at certain points along the hip and knee kinematic (see Appendix Figure

F.1, Page 392 and Appendix Figure F.2, Page 393, respectively) and kinetic (see

Appendix Figure F.3, Page 394 and Appendix Figure F.4, Page 395,

respectively) waveforms.

Appendix F

392

Appendix Figure F.1 SPM{X2} and SPM{t} test statistic trajectories

from canonical correlations analysis and SPM1D

linear regression between approach speed and hip joint rotations from the netball-specific

landing task. Dashed lines indicate the critical

threshold for statistical significance.

Appendix F

393


from canonical correlations analysis and SPM1D linear regression between approach speed and

knee joint rotations from the netball-specific landing task. Dashed lines indicate the critical


Appendix F

394


from canonical correlations analysis and SPM1D linear regression between approach speed and

hip joint moments from the netball-specific landing task. Dashed lines indicate the critical


Appendix F

395


from canonical correlations analysis and SPM1D

linear regression between approach speed and knee joint moments from the netball-specific

landing task. Dashed lines indicate the critical


396

Appendix G Descriptions of the Principal Patterns

Extracted from the Netball-Specific Landing Task used

in Chapter Six

Within Chapter Six, principal component analysis (PCA) was applied to the

extracted kinematic and kinetic data from the netball-specific landing task to

identify and calculate the principal patterns (PP) and PP-Scores, respectively.

This appendix provides a comprehensive description and interpretation of

these PPs. The figures included in this appendix provide a graphical display

of the patterns associated with high (‘+’ symbols) and low (‘–’ symbols) PP-

Scores for each PP.

Appendix G

397

G.1 Kinematic Data

G.1.1 Hip Flexion (FLEX) / Extension (EXT) Principal Patterns

Appendix Figure G.1 Mean hip FLEX/EXT kinematic data (solid line)

from the netball-specific landing task with the

effect of adding and subtracting a multiple of hip FLEX/EXT PP1 and PP2 to the mean curve shown

by the plus and minus symbols, respectively.

Hip FLEX/EXT PP1 explained 82.4% of the variance in hip FLEX/EXT

kinematic data, and was representative of the general shape and overall

magnitude. Higher PP-Scores indicated greater overall hip flexion, while

lower PP-Scores indicated reduced overall hip flexion. Hip FLEX/EXT PP2

explained 10.6% of the variance in hip FLEX/EXT kinematic data, and was

representative of the magnitude of hip flexion from initial contact (IC) to 50

milliseconds post IC. Higher PP-Scores indicated greater hip flexion, while

lower PP-Scores indicated reduced hip flexion.

Appendix G

398

G.1.2 Hip Adduction (ADD) / Abduction (ABD) Principal Patterns

Appendix Figure G.2 Mean hip ADD/ABD kinematic data (solid line)

from the netball-specific landing task with the effect of adding and subtracting a multiple of hip

ADD/ABD PP1 and PP2 to the mean curve shown by the plus and minus symbols,

respectively.

Hip ADD/ABD PP1 explained 84.6% of the variance in hip ADD/ABD


magnitude. Higher PP-Scores indicated greater overall hip adduction, while

lower PP-Scores indicated reduced overall hip adduction. Hip ADD/ABD PP2

explained 10.0% of the variance in hip ADD/ABD kinematic data, and was

representative of the magnitude of frontal plane hip angular displacement.

Higher PP-Scores indicated reduced frontal plane hip angular displacement,

while lower PP-Scores indicated greater frontal plane hip angular

displacement.

399

Appendix G

G.1

.3 H

ip In

tern

al (I

R) /

Ext

erna

l (ER

) Rot

atio

n Pr

inci

pal P

atte

rns

App

endi

x Fi

gure

G.3

M

ean

hip

IR/E

R k

inem

atic

dat

a (s

olid

lin

e) f

rom

the

netb

all-

spec

ific

land

ing

task

with

the

effe

ct o

f

addi

ng a

nd su

btra

ctin

g a

mul

tiple

of h

ip IR

/ER

PP1

, PP2

and

PP3

to th

e m

ean

curv

e sh

own

by th

e pl

us

and

min

us s

ymbo

ls, r

espe

ctiv

ely.

Appendix G

400

Hip IR/ER PP1 explained 73.0% of the variance in hip IR/ER kinematic data,

and was representative of the general shape and overall magnitude. Higher

PP-Scores indicated greater overall hip internal rotation, while lower PP-

Scores indicated greater overall hip external rotation. Hip IR/ER PP2 explained

16.1% of the variance in hip IR/ER kinematic data, and was representative of

the magnitude of hip external rotation from initial contact (IC) to 30

milliseconds post IC. Higher PP-Scores indicated reduced hip external

rotation, while lower PP-Scores indicated greater hip external rotation. Hip

IR/ER PP3 explained 7.1% of the variance in hip IR/ER kinematic data, and

was representative of the magnitude of peak hip internal rotation around 60 –

90 ms after IC. Higher PP-Scores indicated a greater hip internal rotation peak,

while lower PP-Scores indicated a reduced hip internal rotation peak.

Appendix G

401

G.1.4 Knee Extension (EXT) / Flexion (FLEX) Principal Patterns

Appendix Figure G.4 Mean knee EXT/FLEX kinematic data (solid line)

from the netball-specific landing task with the effect of adding and subtracting a multiple of

knee EXT/FLEX PP1 and PP2 to the mean curve shown by the plus and minus symbols,

respectively.

Knee EXT/FLEX PP1 explained 83.3% of the variance in knee EXT/FLEX


magnitude. Higher PP-Scores indicated reduced overall knee flexion, while

lower PP-Scores indicated greater overall knee flexion. Knee EXT/FLEX PP2

explained 8.9% of the variance in knee EXT/FLEX kinematic data, and was

representative of the magnitude of sagittal plane knee angular displacement.

Higher PP-Scores indicated reduced sagittal plane knee angular displacement,

while lower PP-Scores indicated greater sagittal plane knee angular

displacement.

Appendix G

402

G.1.5 Knee Adduction (ADD) / Abduction (ABD) Principal Patterns

Appendix Figure G.5 Mean knee ADD/ABD kinematic data (solid line)


knee ADD/ABD PP1 and PP2 to the mean curve shown by the plus and minus symbols,

respectively.

Knee ADD/ABD PP1 explained 79.2% of the variance in knee ADD/ABD


magnitude. Higher PP-Scores indicated reduced knee abduction but greater

knee adduction, while lower PP-Scores indicated greater overall knee

abduction. Knee ADD/ABD PP2 explained 11.0% of the variance in knee

ADD/ABD kinematic data, and was representative of the timing of the knee

adduction peak. Higher PP-Scores indicated an earlier knee adduction peak,

while lower PP-Scores indicated a delayed knee adduction peak.

403

Appendix G

G.1

.6 K

nee

Inte

rnal

(IR

) / E

xter

nal (

ER) R

otat

ion

Prin

cipa

l Pat

tern

s

App

endi

x Fi

gure

G.6

M

ean

knee

IR

/ER

kin

emat

ic d

ata

(sol

id li

ne) f

rom

the

netb

all-

spec

ific

land

ing

task

wit

h th

e ef

fect

of

addi

ng a

nd s

ubtr

actin

g a

mul

tiple

of

knee

IR

/ER

PP1

, PP2

and

PP3

to th

e m

ean

curv

e sh

own

by th

e

plus

and

min

us s

ymbo

ls, r

espe

ctiv

ely.

Appendix G

404

Knee IR/ER PP1 explained 58.0% of the variance in knee IR/ER kinematic data,


PP-Scores indicated greater overall knee internal rotation, while lower PP-

Scores indicated reduced overall knee internal rotation. Knee IR/ER PP2

explained 22.6% of the variance in knee IR/ER kinematic data, and was

representative of the magnitude of knee external rotation from initial contact

(IC) to 30 milliseconds post IC. Higher PP-Scores indicated reduced knee

external rotation, while lower PP-Scores indicated greater knee external

rotation. Knee IR/ER PP3 explained 13.4% of the variance in knee IR/ER

kinematic data, and was representative of the magnitude of peak knee internal

rotation. Higher PP-Scores indicated greater peak knee internal rotation, while

lower PP-Scores indicated reduced peak knee internal rotation.

Appendix G

405

G.2 Kinetic Data

G.2.1 Hip Extension (EXT) / Flexion (FLEX) Moment Principal Patterns

Appendix Figure G.7 Mean hip EXT/FLEX kinetic data (solid line) from

the netball-specific landing task with the effect of adding and subtracting a multiple of hip

EXT/FLEX moment PP1, PP2, PP3 and PP4 to the

mean curve shown by the plus and minus symbols, respectively.

Appendix G

406

Hip EXT/FLEX moment PP1 explained 61.9% of the variance in hip FLEX/EXT

kinetic data, and was representative of the general shape and overall

magnitude. Higher PP-Scores indicated reduced overall hip flexion moment,

while lower PP-Scores indicated greater overall hip flexion moment. Hip

EXT/FLEX moment PP2 explained 19.1% of the variance in hip FLEX/EXT

kinetic data, and was representative of the magnitude of the peak hip flexion

moment. Higher PP-Scores indicated a reduced peak hip flexion moment,

while lower PP-Scores indicated a greater peak hip flexion moment. Hip


kinetic data, and was representative of the magnitude of the hip flexion

moment from 30 – 60 ms post initial contact (IC). Higher PP-Scores indicated

reduced hip flexion moment, while lower PP-Scores indicated greater hip

flexion moment from 30 – 60 ms post IC. Hip EXT/FLEX moment PP4

explained 5.9% of the variance in hip FLEX/EXT kinetic data, and was

representative of the magnitude of the hip flexion moment from 60 – 90 ms

post IC. Higher PP-Scores indicated reduced hip flexion moment, while lower

PP-Scores indicated greater hip flexion moment from 60 – 90 ms post IC.

407

Appendix G

G.2

.2 H

ip A

bduc

tion

(ABD

) / A

dduc

tion

(AD

D) M

omen

t Pri

ncip

al P

atte

rns

App

endi

x Fi

gure

G.8

M

ean

hip

ABD

/AD

D k

inet

ic d

ata

(sol

id li

ne) f

rom

the

netb

all-

spec

ific

land

ing

task

with

the

effe

ct o

f

addi

ng a

nd s

ubtr

actin

g a

mul

tiple

of

hip

ABD

/AD

D m

omen

t PP

1, P

P2 a

nd P

P3 t

o th

e m

ean

curv

e

show

n by

the

plus

and

min

us s

ymbo

ls, r

espe

ctiv

ely.

Appendix G

408

Hip ABD/ADD moment PP1 explained 66.9% of the variance in hip ABD/ADD


magnitude. Higher PP-Scores indicated reduced overall hip adduction

moment, while lower PP-Scores indicated greater overall hip adduction

moment. Hip ABD/ADD moment PP2 explained 20.1% of the variance in hip

ABD/ADD kinetic data, and was representative of the magnitude of the hip

abduction moment from initial contact (IC) to 30 ms post IC. Higher PP-Scores

indicated a greater hip abduction moment, while lower PP-Scores indicated a

reduced hip abduction moment from IC to 30 ms post IC. Hip ABD/ADD

moment PP3 explained 6.1% of the variance in hip ABD/ADD kinetic data, and

was representative of the presence of bimodal peaks in frontal plane hip

moments. Higher PP-Scores indicated the presence of hip abduction and

adduction moment peaks, while lower PP-Scores indicated the presence of a

singular hip adduction moment peak.

Appendix G

409

G.2.3 Hip External (ER) / Internal (IR) Rotation Moment Principal Patterns

Appendix Figure G.9 Mean hip ER/IR kinetic data (solid line) from the

netball-specific landing task with the effect of adding and subtracting a multiple of hip ER/IR

moment PP1 and PP2 to the mean curve shown


Hip ER/IR moment PP1 explained 77.1% of the variance in hip ER/IR kinetic

data, and was representative of the general shape and overall magnitude.

Higher PP-Scores indicated greater overall hip internal rotation moment,

while lower PP-Scores indicated reduced overall hip internal rotation

moment. Hip ER/IR moment PP2 explained 14.2% of the variance in hip ER/IR

kinetic data, and was representative of the magnitude of the peak hip external

rotation moment. Higher PP-Scores indicated a reduced peak hip external

rotation moment, while lower PP-Scores indicated a greater peak hip external

rotation moment.

410

Appendix G

G.2

.4 K

nee

Flex

ion

(FLE

X) /

Ext

ensi

on (E

XT)

Mom

ent P

rinc

ipal

Pat

tern

s

App

endi

x Fi

gure

G.1

0 M

ean

knee

FLE

X/EX

T ki

netic

dat

a (s

olid

line

) fro

m th

e ne

tbal

l-sp

ecif

ic la

ndin

g ta

sk w

ith

the

effe

ct o

f

addi

ng a

nd s

ubtr

actin

g a

mul

tiple

of

knee

FLE

X/EX

T m

omen

t PP

1, P

P2 a

nd P

P3 to

the

mea

n cu

rve

show

n by

the

plus

and

min

us s

ymbo

ls, r

espe

ctiv

ely.

Appendix G

411

Knee FLEX/EXT moment PP1 explained 49.3% of the variance in knee

FLEX/EXT kinetic data, and was representative of the general shape and

overall magnitude. Higher PP-Scores indicated greater overall knee flexion

moment, while lower PP-Scores indicated reduced overall knee flexion

moment. Knee FLEX/EXT moment PP2 explained 36.4% of the variance in

knee FLEX/EXT kinetic data, and was representative of the timing of peak knee

flexion moment. Higher PP-Scores indicated delayed peak knee flexion

moment, while lower PP-Scores indicated earlier peak knee flexion moment.


FLEX/EXT kinetic data, and was representative of magnitude of sagittal plane

knee moment peaks. Higher PP-Scores indicated reduced knee flexion and

extension peak moments, while lower PP-Scores indicated greater knee flexion

and extension peak moments.

412

Appendix G

G.2

.5 K

nee

Abd

ucti

on (A

BD) /

Add

ucti

on (A

DD

) Mom

ent P

rinc

ipal

Pat

tern

s

App

endi

x Fi

gure

G.1

1 M

ean

knee

ABD

/AD

D k

inet

ic d

ata

(sol

id li

ne) f

rom

the

netb

all-s

peci

fic

land

ing

task

with

the

effe

ct o

f

addi

ng a

nd s

ubtr

actin

g a

mul

tiple

of

knee

ABD

/AD

D m

omen

t PP1

, PP2

and

PP3

to th

e m

ean

curv

e

show

n by

the

plus

and

min

us s

ymbo

ls, r

espe

ctiv

ely.

Appendix G

413

Knee ABD/ADD moment PP1 explained 61.2% of the variance in knee

ABD/ADD kinetic data, and was representative of magnitude of knee

adduction moment. Higher PP-Scores indicated reduced knee adduction

moment, while lower PP-Scores indicated greater knee adduction moment.


ABD/ADD kinetic data, and was representative of magnitude of peak knee

abduction moment. Higher PP-Scores indicated greater peak knee abduction

moment, while lower PP-Scores indicated reduced peak knee abduction

moment. Knee ABD/ADD moment PP3 explained 9.0% of the variance in knee

ABD/ADD kinetic data, and was representative of the timing of peak knee

abduction moment. Higher PP-Scores indicated earlier and greater peak knee

abduction moment, while lower PP-Scores indicated delayed and reduced

peak knee abduction moment.

Appendix G

414

G.2.6 Knee External (ER) / Internal (IR) Rotation Moment Principal Patterns

Appendix Figure G.12 Mean knee ER/IR kinetic data (solid line) from

the netball-specific landing task with the effect of adding and subtracting a multiple of knee

ER/IR moment PP1 and PP2 to the mean curve shown by the plus and minus symbols,

respectively.

Knee ER/IR moment PP1 explained 81.4% of the variance in knee ER/IR kinetic


Higher PP-Scores indicated greater overall knee internal rotation moment,

while lower PP-Scores indicated reduced overall knee internal rotation

moment. Knee ER/IR moment PP2 explained 12.8% of the variance in knee

ER/IR kinetic data, and was representative of magnitude of peak knee internal

rotation moment. Higher PP-Scores indicated greater peak knee internal

rotation moment, while lower PP-Scores indicated reduced peak knee internal

rotation moment.

415

App

endi

x H

M

etho

dolo

gica

l Qua

lity

Ass

essm

ent f

or S

tudi

es In

clud

ed in

Cha

pter

Sev

en

App

endi

x Ta

ble

H.1

R

esul

ts o

f met

hodo

logi

cal q

ualit

y as

sess

men

t und

erta

ken

on in

clud

ed s

tudi

es.

Stud

y 1

2 3

4 5

6 7

8 9

10

11

Tota

l ( /

11)

Cor

tes

& O

nate

[409

] 1

1 0

0 1

N/A

1

1 1

1 1

8 (/1

0)

Dud

ley

et a

l.[414

] 1

1 0

1 1

N/A

1

1 1

1 1

9 (/1

0)

Ekeg

ren

et a

l.[85]

1 1

0 1

1 N

/A

1 1

1 1

1 9

(/10)

G

oets

chiu

s et a

l.[412

] 1

1 1

1 1

1 0

1 1

1 1

10

Her

ring

ton

et a

l.[256

] 1

1 0

1 1

N/A

1

1 1

1 1

9 (/1

0)

McL

ean

et a

l.[92]

1 1

0 1

1 N

/A

0 1

0 1

1 7

(/10)

M

yer e

t al.[8

7]

N/A

1

N/A

N

/A

1 N

/A

N/A

N

/A

0 N

/A

1 3

(/4)

Mye

r et a

l.[86]

N/A

1

0 N

/A

1 N

/A

N/A

1

1 N

/A

1 5

(/5)

Mye

r et a

l.[88]

0 1

0 1

1 N

/A

0 1

0 1

1 6

(/10)

M

yer e

t al.[9

3]

1 1

0 1

1 N

/A

0 1

1 1

1 8

(/10)

M

yer e

t al.[8

9]

1 1

0 1

1 N

/A

0 1

0 1

1 7

(/10)

M

yer e

t al.[9

0]

1 1

0 1

1 N

/A

0 1

0 1

1 7

(/10)

M

yer e

t al.[4

11]

N/A

N

/A

N/A

N

/A

N/A

N

/A

N/A

N

/A

N/A

N

/A

N/A

N

/A

Mye

r et a

l.[415

] N

/A

1 N

/A

N/A

1

N/A

N

/A

N/A

0

N/A

1

3 (/4

) O

nate

et a

l.[94]

1 1

0 1

0 N

/A

1 1

1 1

1 8

(/10)

Pa

dua

et a

l.[91]

1 1

0 1

1 N

/A

0 1

1 1

1 8

(/10)

416

Appendix H

App

endi

x Ta

ble

H.1

R

esul

ts o

f met

hodo

logi

cal q

ualit

y as

sess

men

t und

erta

ken

on in

clud

ed s

tudi

es (c

ontin

ued)

.

Stud

y 1

2 3

4 5

6 7

8 9

10

11

Tota

l ( /

11)

Pa

dua

et a

l.[104

] 1

1 1

1 1

1 1

1 0

1 1

10

Padu

a et

al.[4

10]

1 1

0 1

1 N

/A

1 1

1 1

1 9

(/10)

Sm

ith e

t al.[1

05]

1 1

1 1

1 1

0 1

1 1

1 10

St

roub

e et

al.[4

13]

1 1

0 1

1 1

1 1

0 1

1 9

1 =

Subj

ect i

nclu

sion/

excl

usio

n cr

iteria

clea

rly d

escr

ibed

(1 p

oint

) 2

= En

ough

info

rmat

ion

to id

entif

y se

tting

(1 p

oint

) 3

= St

udy

desig

n –

pros

pect

ive (

1 po

int)

or re

tros

pect

ive (

0 po

ints

) col

lectio

n of

dat

a 4

= A

ge (m

ean/

med

ian

and

stan

dard

dev

iatio

n/ra

nge)

and

gen

der r

epor

ted

(1 p

oint

) 5

= D

escr

iptio

n of

scre

enin

g m

etho

d ha

d su

ffici

ent d

etai

l to

perm

it re

plic

atio

n of

the t

est.

Test

dev

ice o

r ins

trum

ents

, pro

toco

l of s

cree

ning

met

hod

repo

rted

(1 p

oint

) 6

= G

roup

s wer

e com

para

ble a

t bas

eline

, inc

ludi

ng a

ll m

ajor

conf

ound

ing

and

prog

nost

ic fa

ctor

s (1

poin

t)

7 =

Inve

stig

ator

s wer

e kep

t ‘bl

ind’

to im

port

ant c

onfo

undi

ng a

nd p

rogn

ostic

fact

ors (

1 po

int)

8

= Fo

r var

iabl

e/s o

f int

eres

t, de

tails

giv

en o

n m

ean/

med

ian,

stan

dard

dev

iatio

n/co

nfid

ence

inte

rval

s and

pre

dict

ive v

alue

(1 p

oint

) 9

= Re

liabi

lity

repo

rted

(1 p

oint

) 10

= A

ll in

clud

ed su

bjec

ts m

easu

red

and,

if a

ppro

pria

te, m

issin

g da

ta o

r with

draw

als f

rom

stud

y re

port

ed o

r exp

lain

ed (1

poi

nt)

11 =

Out

com

e clea

rly d

efine

d an

d m

etho

d of

exam

inat

ion

of o

utco

me a

dequ

ate (

1 po

int)

A

scor

e of z

ero

was

giv

en fo

r crit

eria

if in

suffi

cien

t inf

orm

atio

n w

as a

vaila

ble f

or a

ccur

ate a

sses

smen

t N

/A –

cri

teri

a no

t app

licab

le

417

Appendix I Full Results of Intra- and Inter-Rater

Reliability Analyses for the Landing Error Scoring

System (LESS) and Tuck Jump Assessment

I.1 Landing Error Scoring System (LESS) Reliability

Appendix Table I.1 Percentage of exact agreement (PEA) and Kappa (κ) statistics for intra- and inter-rater reliability of

the Landing Error Scoring System (LESS) scoring

criteria.

Scoring Criteria Intra-Rater PEA

Intra-Rater Cohen’s κ

Inter-Rater PEA


Knee flexion angle at IC 100% 1.000 80% 0.600

Hip flexion angle at IC 100% 1.000 100% 1.000

Trunk flexion angle at IC 80% 0.524 70% 0.400

Ankle plantar-flexion angle at IC 90% 0.615 90% 0.375

Knee valgus angle at IC 100% 1.000 90% 0.800

Lateral trunk flexion angle at IC 100% 1.000 100% 1.000

Stance width – Wide 100% 1.000 100% 1.000

Stance width – Narrow 90% 0.615 80% 0.545

Foot position – Toe in 100% 1.000 100% 1.000

Foot position – Toe out 100% 1.000 100% 1.000

Symmetrical initial foot contact 100% 1.000 100% 1.000

Knee flexion displacement 100% 1.000 80% 0.600

Hip flexion at maximum knee flexion 90% 0.783 80% 0.286

Trunk flexion at maximum knee flexion 100% 1.000 80% 0.600

Knee valgus displacement 100% 1.000 100% 1.000

Joint displacement 100% 1.000 80% 0.643

Overall impression 100% 1.000 100% 1.000

ALL CRITERIA 97% 0.933 90% 0.745

IC – Initial contact; PEA – Percentage of exact agreement; κ - Kappa

Appendix I

418

I.2 Tuck Jump Assessment Reliability

Appendix Table I.2 Percentage of exact agreement (PEA) and Kappa (κ) statistics for intra- and inter-rater reliability of

the Tuck Jump Assessment scoring criteria.

Scoring Criteria Intra-Rater

PEA Intra-Rater Cohen’s κ

Inter-Rater PEA


Knee valgus on landing 100% 1.000 80% 0.545

Thighs not reaching parallel 100% 1.000 70% 0.400

Thighs not equal side to side 90% 0.800 90% 0.800

Foot placement not shoulder width apart 90% 0.737 80% 0.524

Foot placement not parallel 100% 1.000 70% 0.400

Foot contact timing not equal 100% 1.000 90% 0.800

Does not land in same foot print 100% 1.000 100% 1.000

Excessive landing contact noise 100% 1.000 100% 1.000

Pause between jumps 100% 1.000 90% 0.800

Technique declines prior to 10 seconds 90% 0.783 90% 0.783

ALL CRITERIA 97% 0.940 84% 0.689

IC – Initial contact; PEA – Percentage of exact agreement; κ - Kappa

419

Appendix J Descriptions of the Principal Patterns

Extracted from the Single-Leg Squat Task used in

Chapter Nine

Within Chapter Nine, principal component analysis (PCA) was applied to the

extracted kinematic data from the single-leg squat task to identify and

calculate the principal patterns (PP) and PP-Scores, respectively. This

appendix provides a comprehensive description and interpretation of these

PPs. The figures included in this appendix provide a graphical display of the

patterns associated with high (‘+’ symbols) and low (‘–’ symbols) PP-Scores for

each PP.

Appendix J

420

J.1 Kinematic Data

J.1.1 Hip Flexion (FLEX) / Extension (EXT) Principal Patterns

Appendix Figure J.1 Mean hip FLEX/EXT kinematic data (solid line)

from the single-leg squat task with the effect of adding and subtracting a multiple of hip

FLEX/EXT PP1 and PP2 to the mean curve shown







representative of the timing of peak hip flexion. Higher PP-Scores indicated

delayed peak hip flexion, while lower PP-Scores indicated earlier peak hip

flexion.

421

Appendix J

J.1.2

Hip

Add

ucti

on (A

DD

) / A

bduc

tion

(ABD

) Pri

ncip

al P

atte

rns

App

endi

x Fi

gure

J.2

Mea

n hi

p A

DD

/ABD

kin

emat

ic d

ata

(sol

id li

ne) f

rom

the

sing

le-le

g sq

uat t

ask

with

the

effe

ct o

f add

ing

and

subt

ract

ing

a m

ultip

le o

f hip

AD

D/A

BD P

P1, P

P2 a

nd P

P3 to

the

mea

n cu

rve

show

n by

the

plus

an

d m

inus

sym

bols

, res

pect

ivel

y.

Appendix J

422






representative of the magnitude of hip adduction at ~0-25% and ~75-100% of

the squat. Higher PP-Scores indicated greater hip adduction, while lower PP-

Scores indicated reduced hip adduction at ~0-25% and ~75-100% of the squat.


kinematic data, and was representative of the timing of peak hip adduction.

Higher PP-Scores indicated earlier peak hip adduction, while lower PP-Scores

indicated delayed peak hip adduction.

Appendix J

423

J.1.3 Hip Internal (IR) / External (ER) Rotation Principal Patterns

Appendix Figure J.3 Mean hip IR/ER kinematic data (solid line) from the single-leg squat task with the effect of adding

and subtracting a multiple of hip IR/ER PP1 and

PP2 to the mean curve shown by the plus and minus symbols, respectively.



PP-Scores indicated reduced overall hip external rotation, while lower PP-



the magnitude of hip internal and external rotation peaks. Higher PP-Scores

indicated greater hip transverse plane peak motion, while lower PP-Scores

indicated reduced hip transverse plane peak motion.

Appendix J

424

J.1.4 Knee Extension (EXT) / Flexion (FLEX) Principal Patterns

Appendix Figure J.4 Mean knee EXT/FLEX kinematic data (solid line) from the single-leg squat task with the effect of

adding and subtracting a multiple of hip

EXT/FLEX PP1 and PP2 to the mean curve shown by the plus and minus symbols, respectively.




lower PP-Scores indicated greater overall knee flexion. Knee EXT/FLEX PP2

explained 17.2% of the variance in knee EXT/FLEX kinematic data, and was

representative of the timing of peak knee flexion. Higher PP-Scores indicated

delayed peak knee flexion, while lower PP-Scores indicated earlier peak knee

flexion.

Appendix J

425

J.1.5 Knee Adduction (ADD) / Abduction (ABD) Principal Patterns

Appendix Figure J.5 Mean knee ADD/ABD kinematic data (solid line) from the single-leg squat task with the effect of

adding and subtracting a multiple of knee


respectively.



magnitude. Higher PP-Scores indicated greater overall knee adduction, while

lower PP-Scores indicated greater overall knee abduction. Knee ADD/ABD

PP2 explained 5.7% of the variance in knee ADD/ABD kinematic data, and was

representative of the magnitude of knee adduction and abduction peaks.

Higher PP-Scores indicated greater knee frontal plane peak motion, while

lower PP-Scores indicated reduced knee frontal plane peak motion.

Appendix J

426

J.1.6 Knee Internal (IR) / External (ER) Rotation Principal Patterns

Appendix Figure J.6 Mean knee IR/ER kinematic data (solid line) from the single-leg squat task with the effect of

adding and subtracting a multiple of knee IR/ER

PP1 and PP2 to the mean curve shown by the plus and minus symbols, respectively.






representative of the magnitude of transverse plane knee angular

displacement. Higher PP-Scores indicated greater transverse plane knee

angular displacement, while lower PP-Scores indicated reduced transverse

plane knee angular displacement.

427

Appendix J

J.1.7

Ank

le D

orsi

flexi

on (D

F) /

Plan

tarf

lexi

on (P

F) P

rinc

ipal

Pat

tern

s

App

endi

x Fi

gure

J.7

Mea

n an

kle

DF/

PF k

inem

atic

dat

a (s

olid

line

) fro

m th

e si

ngle

-leg

squa

t tas

k w

ith th

e ef

fect

of a

ddin

g

and

subt

ract

ing

a m

ultip

le o

f ank

le D

F/PF

PP1

, PP2

and

PP3

to th

e m

ean

curv

e sh

own

by th

e pl

us a

nd

min

us s

ymbo

ls, r

espe

ctiv

ely.

Appendix J

428

Ankle DF/PF PP1 explained 71.0% of the variance in ankle DF/PF kinematic


Higher PP-Scores indicated greater overall ankle dorsiflexion, while lower PP-

Scores indicated reduced overall ankle dorsiflexion. Ankle DF/PF PP2

explained 16.0% of the variance in ankle DF/PF kinematic data, and was

representative of the ankle dorsiflexion range of motion (ROM) and timing of

dorsiflexion peak. Higher PP-Scores indicated reduced ROM and delayed

peak dorsiflexion, while lower PP-Scores indicated greater ROM and earlier

peak dorsiflexion. Ankle DF/PF PP3 explained 8.7% of the variance in ankle

DF/PF kinematic data, and was representative of the ankle dorsiflexion range

of motion (ROM) and timing of dorsiflexion peak. Higher PP-Scores indicated

reduced ROM and earlier peak dorsiflexion, while lower PP-Scores indicated

greater ROM and delayed peak dorsiflexion.

Appendix J

429

J.1.8 Ankle Inversion (INV) / Eversion (EVE) Principal Patterns

Appendix Figure J.8 Mean ankle INV/EVE kinematic data (solid line) from the single-leg squat task with the effect of

adding and subtracting a multiple of ankle

INV/EVE PP1 to the mean curve shown by the plus and minus symbols, respectively.

Ankle INV/EVE PP1 explained 91.6% of the variance in ankle INV/EVE


magnitude. Higher PP-Scores indicated reduced overall ankle inversion, while

lower PP-Scores indicated greater overall ankle inversion.

Appendix J

430

J.1.9 Ankle Internal (IR) / External (ER Rotation Principal Patterns

Appendix Figure J.9 Mean ankle IR/ER kinematic data (solid line)

from the single-leg squat task with the effect of adding and subtracting a multiple of ankle IR/ER

PP1, PP2, PP3 and PP4 to the mean curve shown


Appendix J

431

Ankle IR/ER PP1 explained 61.3% of the variance in ankle IR/ER kinematic


Higher PP-Scores indicated reduced overall ankle external rotation, while

lower PP-Scores indicated greater overall ankle external rotation. Ankle IR/ER

PP2 explained 13.4% of the variance in ankle IR/ER kinematic data, and was

representative of the magnitude of ankle external rotation at ~0-20% of the

squat. Higher PP-Scores indicated reduced ankle external rotation, while

lower PP-Scores indicated greater ankle external rotation at ~0-20% of the

squat. Ankle IR/ER PP3 explained 9.1% of the variance in ankle IR/ER

kinematic data, and was representative of the magnitude of ankle external

rotation at ~75-100% of the squat. Higher PP-Scores indicated reduced ankle

external rotation, while lower PP-Scores indicated greater ankle external

rotation at ~75-100% of the squat. Ankle IR/ER PP4 explained 7.0% of the

variance in ankle IR/ER kinematic data, and was representative of the

magnitude of ankle external rotation at ~60-75% of the squat. Higher PP-Scores

indicated greater ankle external rotation, while lower PP-Scores indicated

reduced ankle external rotation at ~60-75% of the squat.

432

Appendix K Descriptions of the Principal Patterns

Extracted from the Netball-Specific Landing Task used

in Chapter Nine

Within Chapter Nine, principal component analysis (PCA) was applied to the

extracted kinematic and kinetic data from the netball-specific landing task to

identify and calculate the principal patterns (PP) and PP-Scores, respectively.

This appendix provides a comprehensive description and interpretation of

these PPs. The figures included in this appendix provide a graphical display

of the patterns associated with high (‘+’ symbols) and low (‘–’ symbols) PP-

Scores for each PP.

Appendix K

433

K.1 Kinematic Data

K.1.1 Hip Flexion (FLEX) / Extension (EXT) Principal Patterns

Appendix Figure K.1 Mean hip FLEX/EXT kinematic data (solid line)

from the netball-specific landing task with the

effect of adding and subtracting a multiple of hip FLEX/EXT PP1 and PP2 to the mean curve shown







representative of the duration of hip flexion. Higher PP-Scores indicated

prolonged hip flexion, while lower PP-Scores indicated a shortened duration

of hip flexion.

Appendix K

434

K.1.2 Hip Adduction (ADD) / Abduction (ABD) Principal Patterns

Appendix Figure K.2 Mean hip ADD/ABD kinematic data (solid line)

from the netball-specific landing task with the effect of adding and subtracting a multiple of hip


respectively.






representative of the magnitude of frontal plane hip angular displacement.

Higher PP-Scores indicated greater frontal plane hip angular displacement,

while lower PP-Scores indicated reduced frontal plane hip angular

displacement.

Appendix K

435

K.1.3 Hip Internal (IR) / External (ER) Rotation Principal Patterns

Appendix Figure K.3 Mean hip IR/ER kinematic data (solid line) from

the netball-specific landing task with the effect of adding and subtracting a multiple of hip IR/ER

PP1 and PP2 to the mean curve shown by the plus and minus symbols, respectively.



PP-Scores indicated greater overall hip internal rotation, while lower PP-



the magnitude of hip external rotation from initial contact (IC) to 30

milliseconds post IC. Higher PP-Scores indicated reduced hip external

rotation, while lower PP-Scores indicated greater hip external rotation.

Appendix K

436

K.1.4 Knee Extension (EXT) / Flexion (FLEX) Principal Patterns

Appendix Figure K.4 Mean knee EXT/FLEX kinematic data (solid line)


knee EXT/FLEX PP1 and PP2 to the mean curve shown by the plus and minus symbols,

respectively.




lower PP-Scores indicated greater overall knee flexion.

437

Appendix K

K.1

.5 K

nee

Add

ucti

on (A

DD

) / A

bduc

tion

(ABD

) Pri

ncip

al P

atte

rns

App

endi

x Fi

gure

K.5

M

ean

knee

AD

D/A

BD k

inem

atic

dat

a (s

olid

line

) fro

m th

e ne

tbal

l-spe

cifi

c lan

ding

task

with

the

effe

ct

of a

ddin

g an

d su

btra

ctin

g a

mul

tiple

of

knee

AD

D/A

BD P

P1, P

P2 a

nd P

P3 to

the

mea

n cu

rve

show

n

by th

e pl

us a

nd m

inus

sym

bols

, res

pect

ivel

y.

Appendix K

438



magnitude. Higher PP-Scores indicated reduced knee abduction but greater

knee adduction, while lower PP-Scores indicated greater overall knee

abduction. Knee ADD/ABD PP2 explained 10.7% of the variance in knee

ADD/ABD kinematic data, and was representative of the timing of the knee

adduction peak. Higher PP-Scores indicated an earlier knee adduction peak,

while lower PP-Scores indicated a delayed knee adduction peak. Knee

ADD/ABD PP3 explained 7.8% of the variance in knee ADD/ABD kinematic

data, and was representative of the presence of bimodal peaks in frontal plane

knee motion. Higher PP-Scores indicated a singular knee abduction peak,

while lower PP-Scores indicated knee abduction and adduction peaks.

439

Appendix K

K.1

.6 K

nee

Inte

rnal

(IR

) / E

xter

nal (

ER) R

otat

ion

Prin

cipa

l Pat

tern

s

App

endi

x Fi

gure

K.6

M

ean

knee

IR

/ER

kin

emat

ic d

ata

(sol

id li

ne) f

rom

the

netb

all-s

peci

fic

land

ing

task

wit

h th

e ef

fect

of

addi

ng a

nd s

ubtr

actin

g a

mul

tiple

of

knee

IR

/ER

PP1

, PP2

and

PP3

to th

e m

ean

curv

e sh

own

by th

e

plus

and

min

us s

ymbo

ls, r

espe

ctiv

ely.

Appendix K

440






representative of the magnitude of knee external rotation from initial contact

(IC) to 30 milliseconds post IC. Higher PP-Scores indicated reduced knee

external rotation, while lower PP-Scores indicated greater knee external

rotation. Knee IR/ER PP3 explained 8.7% of the variance in knee IR/ER

kinematic data, and was representative of the magnitude of peak knee internal

rotation. Higher PP-Scores indicated greater peak knee internal rotation, while

lower PP-Scores indicated reduced peak knee internal rotation.

441

Appendix K

K.2

Kin

etic

Dat

a

K.2

.1 H

ip E

xten

sion

(EX

T) /

Flex

ion

(FLE

X) M

omen

t Pri

ncip

al P

atte

rns

App

endi

x Fi

gure

K.7

M

ean

hip

EXT/

FLEX

kin

etic

dat

a (s

olid

line

) fro

m th

e ne

tbal

l-sp

ecif

ic la

ndin

g ta

sk w

ith th

e ef

fect

of

addi

ng a

nd s

ubtr

acti

ng a

mul

tiple

of

hip

EXT/

FLEX

mom

ent

PP1,

PP2

and

PP3

to

the

mea

n cu

rve

show

n by

the

plus

and

min

us s

ymbo

ls, r

espe

ctiv

ely.

.

Appendix K

442

Hip EXT/FLEX moment PP1 explained 56.8% of the variance in hip FLEX/EXT


magnitude. Higher PP-Scores indicated reduced overall hip flexion moment,

while lower PP-Scores indicated greater overall hip flexion moment. Hip


kinetic data, and was representative of the magnitude of the peak hip flexion

moment. Higher PP-Scores indicated a reduced peak hip flexion moment,

while lower PP-Scores indicated a greater peak hip flexion moment. Hip


kinetic data, and was representative of the magnitude of the hip flexion

moment from 30 – 60 ms post initial contact (IC). Higher PP-Scores indicated

reduced hip flexion moment, while lower PP-Scores indicated greater hip

flexion moment from 30 – 60 ms post IC.

443

Appendix K

K.2

.2 H

ip A

bduc

tion

(ABD

) / A

dduc

tion

(AD

D) M

omen

t Pri

ncip

al P

atte

rns

App

endi

x Fi

gure

K.8

M

ean

hip

ABD

/AD

D k

inet

ic d

ata

(sol

id li

ne) f

rom

the

netb

all-s

peci

fic

land

ing

task

with

the

effe

ct o

f

addi

ng a

nd s

ubtr

actin

g a

mul

tiple

of

hip

ABD

/AD

D m

omen

t PP

1, P

P2 a

nd P

P3 t

o th

e m

ean

curv

e

show

n by

the

plus

and

min

us s

ymbo

ls, r

espe

ctiv

ely.

.

Appendix K

444

Hip ABD/ADD moment PP1 explained 74.0% of the variance in hip ABD/ADD


magnitude. Higher PP-Scores indicated reduced overall hip adduction

moment, while lower PP-Scores indicated greater overall hip adduction

moment. Hip ABD/ADD moment PP2 explained 15.4% of the variance in hip

ABD/ADD kinetic data, and was representative of the magnitude of the hip

abduction moment from initial contact (IC) to 30 ms post IC. Higher PP-Scores

indicated a greater hip abduction moment, while lower PP-Scores indicated a

reduced hip abduction moment from IC to 30 ms post IC. Hip ABD/ADD

moment PP3 explained 5.2% of the variance in hip ABD/ADD kinetic data, and

was representative of the presence of bimodal peaks in frontal plane hip

moments. Higher PP-Scores indicated the presence of hip abduction and

adduction moment peaks, while lower PP-Scores indicated the presence of a

singular hip adduction moment peak.

Appendix K

445

K.2.3 Hip External (ER) / Internal (IR) Rotation Moment Principal Patterns

Appendix Figure K.9 Mean hip ER/IR kinetic data (solid line) from the

netball-specific landing task with the effect of adding and subtracting a multiple of hip ER/IR

moment PP1 and PP2 to the mean curve shown


Hip ER/IR moment PP1 explained 83.0% of the variance in hip ER/IR kinetic


Higher PP-Scores indicated greater overall hip internal rotation moment,

while lower PP-Scores indicated reduced overall hip internal rotation

moment. Hip ER/IR moment PP2 explained 10.2% of the variance in hip ER/IR

kinetic data, and was representative of the magnitude of the peak hip external

rotation moment. Higher PP-Scores indicated a reduced peak hip external

rotation moment, while lower PP-Scores indicated a greater peak hip external

rotation moment.

446

Appendix K

K.2

.4 K

nee

Flex

ion

(FLE

X) /

Ext

ensi

on (E

XT)

Mom

ent P

rinc

ipal

Pat

tern

s

App

endi

x Fi

gure

K.1

0 M

ean

knee

FLE

X/EX

T ki

netic

dat

a (s

olid

line

) fro

m th

e ne

tbal

l-sp

ecif

ic la

ndin

g ta

sk w

ith

the

effe

ct o

f

addi

ng a

nd s

ubtr

actin

g a

mul

tiple

of

knee

FLE

X/EX

T m

omen

t PP1

, PP2

and

PP3

to th

e m

ean

curv

e

show

n by

the

plus

and

min

us s

ymbo

ls, r

espe

ctiv

ely.

Appendix K

447


FLEX/EXT kinetic data, and was representative of the general shape and

overall magnitude. Higher PP-Scores indicated greater overall knee flexion

moment, while lower PP-Scores indicated reduced overall knee flexion

moment. Knee FLEX/EXT moment PP2 explained 29.4% of the variance in

knee FLEX/EXT kinetic data, and was representative of the timing of peak knee

flexion moment. Higher PP-Scores indicated delayed peak knee flexion

moment, while lower PP-Scores indicated earlier peak knee flexion moment.


FLEX/EXT kinetic data, and was representative of magnitude of sagittal plane

knee moment peaks. Higher PP-Scores indicated reduced knee flexion and

extension peak moments, while lower PP-Scores indicated greater knee flexion

and extension peak moments.

448

Appendix K

K.2

.5 K

nee

Abd

ucti

on (A

BD) /

Add

ucti

on (A

DD

) Mom

ent P

rinc

ipal

Pat

tern

s

App

endi

x Fi

gure

K.1

1 M

ean

knee

ABD

/AD

D k

inet

ic d

ata

(sol

id li

ne) f

rom

the

netb

all-s

peci

fic

land

ing

task

with

the

effe

ct o

f

addi

ng a

nd s

ubtr

actin

g a

mul

tiple

of

knee

ABD

/AD

D m

omen

t PP1

, PP2

and

PP3

to th

e m

ean

curv

e

show

n by

the

plus

and

min

us s

ymbo

ls, r

espe

ctiv

ely.

Appendix K

449


ABD/ADD kinetic data, and was representative of magnitude of knee

adduction moment. Higher PP-Scores indicated reduced knee adduction

moment, while lower PP-Scores indicated greater knee adduction moment.


ABD/ADD kinetic data, and was representative of magnitude of peak knee

abduction moment. Higher PP-Scores indicated greater peak knee abduction

moment, while lower PP-Scores indicated reduced peak knee abduction

moment. Knee ABD/ADD moment PP3 explained 8.0% of the variance in knee

ABD/ADD kinetic data, and was representative of the timing and magnitude

of peak knee abduction moment. Higher PP-Scores indicated earlier and

greater peak knee abduction moment, while lower PP-Scores indicated

delayed and reduced peak knee abduction moment.

Appendix K

450

K.2.6 Knee External (ER) / Internal (IR) Rotation Moment Principal Patterns

Appendix Figure K.12 Mean knee ER/IR kinetic data (solid line) from

the netball-specific landing task with the effect of adding and subtracting a multiple of knee

ER/IR moment PP1 and PP2 to the mean curve

shown by the plus and minus symbols, respectively.

Knee ER/IR moment PP1 explained 80.6% of the variance in knee ER/IR kinetic


Higher PP-Scores indicated greater overall knee internal rotation moment,

while lower PP-Scores indicated reduced overall knee internal rotation

moment. Knee ER/IR moment PP2 explained 14.9% of the variance in knee

ER/IR kinetic data, and was representative of magnitude of peak knee internal

rotation moment. Higher PP-Scores indicated greater peak knee internal

rotation moment, while lower PP-Scores indicated reduced peak knee internal

rotation moment.

Neuromechanics of the lower limb - Implications for ACL injury ...

Documents

Transcript of Neuromechanics of the lower limb - Implications for ACL injury ...