Available online at www.sciencedirect.com

Journal of Electromyography and Kinesiology 18 (2008) 390–396

www.elsevier.com/locate/jelekin

Non-uniform electromyographic activity during fatigue and recoveryof the vastus medialis and lateralis muscles

Nosratollah Hedayatpour, Lars Arendt-Nielsen, Dario Farina *

Center for Sensory-Motor Interaction (SMI), Department of Health Science and Technology, Aalborg University,

Fredrik Bajers Vej 7D-3, 9220 Aalborg East, Denmark

Received 11 October 2006; received in revised form 13 December 2006; accepted 13 December 2006

Abstract

The aim of the study was to investigate EMG signal features during fatigue and recovery at three locations of the vastus medialis andlateralis muscles. Surface EMG signals were detected from 10 healthy male subjects with six 8-electrode arrays located at 10%, 20%, and30% of the distance from the medial (for vastus medialis) and lateral (vastus lateralis) border of the patella to the anterior superior spineof the pelvic. Subjects performed contractions at 40% and 80% of the maximal force (MVC) until failure to maintain the target force,followed by 20 2-s contractions at the same force levels every minute for 20 min (recovery). Average rectified value, mean power spectralfrequency, and muscle fiber conduction velocity were estimated from the EMG signals in 10 epochs from the beginning of the contractionto task failure (time to task failure, mean ± SD, 70.7 ± 25.8 s for 40% MVC; 27.4 ± 16.8 s for 80% MVC) and from the 20 2 s time inter-vals during recovery. During the fatiguing contraction, the trend over time of EMG average rectified value depended on location for bothmuscles (P < 0.05). After 20-min recovery, mean frequency and conduction velocity of both muscles were larger than in the beginning ofthe fatigue task (P < 0.05) (supernormal values). Moreover, the trend over time of mean frequency during recovery was affected by loca-tion and conduction velocity values depended on location for both muscles (P < 0.05). The results indicate spatial dependency of EMGvariables during fatigue and recovery and thus the necessity of EMG spatial sampling for global muscle assessment.� 2006 Elsevier Ltd. All rights reserved.

Keywords: Multi-channel EMG; Conduction velocity; Vasti muscles; Recovery

1. Introduction

Muscle fatigue is defined as an exercise-induced decreasein maximal force-generating capacity of a muscle whichmay result from the metabolic accumulation, fuel reduction(Saltin and Karlsson, 1975), neuromuscular dysfunction(Bigland-Ritchie, 1984) and impairment of voluntary acti-vation (Bigland-Ritchie et al., 1978). Recovery after fatigueplays an important role in sport success and in preventingmuscle fiber damage during exercise training (Parra et al.,2000; Clarkson and Tremblay, 1988). The recovery processimplies the return of all mentioned parameters from abnor-

1050-6411/$ - see front matter � 2006 Elsevier Ltd. All rights reserved.

doi:10.1016/j.jelekin.2006.12.004

* Corresponding author. Tel.: +45 96358821; fax: +45 98154008.E-mail address: df@hst.aau.dk (D. Farina).

mal to normal condition (Fletcher, 1907; Ivy et al., 2002;Lannergren et al., 1989).

In large muscles, such as quadriceps, muscle fibers mayhave different pinnation angles and this allows a widedistribution of tensions. Muscle tension primarily dependson morphological and architectural features of musclefibers (Coyle et al., 1979; Ichinose et al., 1998). Specifictasks may be performed by preferential activation of differ-ent muscle parts. Accordingly, previous studies havereported a non-uniform distribution of electromyographic(EMG) activity over muscles during sustained contraction(Li and Sakamoto, 1996; Holtermann et al., 2005). A longlasting muscle hyperactivity with non-uniform motor unitrecruitment has also been correlated to the distribution ofmuscle soreness symptoms 48 h after eccentric exercise(Friden et al., 1986).

VastusMedialis

VastusLateralis

ASSP

ASSP/PatellaLateral Border

LineASSP/PatellaMedial Border

Line

N. Hedayatpour et al. / Journal of Electromyography and Kinesiology 18 (2008) 390–396 391

Surface EMG signals are often used to investigate fati-gue-induced changes at the muscle fiber membrane level(Merletti et al., 1990). However, there have been only afew reports on EMG features during recovery from fatigue(e.g., van der Hoeven et al., 1993; van der Hoeven andLange, 1994). Moreover, no study investigated spatialdependence of EMG features with recovery. Therefore,the aim of the study was to investigate EMG signal featuresduring fatigue and recovery at three locations of the vastusmedialis and lateralis muscles.

Distal Array

Middle Array

Proximal Array

Patella Lateral Border

PatellaMedial Border

10%20

%30%

Fig. 1. Schematic representation of the locations of adhesive electrodearrays over the vastus medialis and lateralis muscles. The three locationscorrespond to distances from the patella of 10% (distal), 20% (middle) and30% (proximal) of the distance between the anterior superior spine of thepelvic (ASSP) and medial (vastus medialis) or lateral (vastus lateralis)border of the patella.

2. Materials and methods

2.1. Subjects

Ten healthy, male subjects (age, mean ± SD, 25.6 ± 3.6 yr,body mass 70.4 ± 12.9 kg, height 1.77 ± 0.09 m) participated tothe study. The study was conducted in accordance with theDeclaration of Helsinki, approved by the Local Ethics Commit-tee, and written informed consent was obtained from all subjectsprior to inclusion.

2.2. General procedures

The subject sat comfortably on a chair fixed with a belt at thehip with the right knee 90� flexed. A strap connected by a chain toa load cell was attached to the ankle to measure knee extensionisometric force. Force was provided to the subject as visualfeedback on an oscilloscope. The subject performed three maxi-mal voluntary contractions (MVC) separated by 2-min rest.During each MVC contraction, verbal encouragement was pro-vided. The highest force was considered the reference MVC forsubmaximal contraction levels. After the MVCs, surface EMGelectrodes were placed on the vastus medialis and lateralis mus-cles, as described below. The subject trained with the visualfeedback on force and, 10 min later, performed two contractionsat 40% and 80% MVC (random order) until task failure, with aresting period of 40 min in between. After the sustained con-traction, EMG signals were recorded at intervals of 1 min for20 min, during 2-s contractions at the same force level as duringthe fatiguing contraction. An additional 2-s contraction wasperformed before the second contraction at the same force level asthe first contraction. Skin temperature was measured at the bellyof both muscles using skin thermometers (Ellab Ltd., Copenha-gen, Denmark).

2.3. EMG recordings

Surface EMG signals were recorded from thee sites over thevastus medialis and lateralis muscles with linear electrode arrays.The lengths from the anterior superior spine of the pelvic (ASSP)to the medial and lateral border of the patella were measured asanatomical references for vastus medialis and lateralis, respec-tively (Zipp, 1982). Three adhesive arrays (ELSCH008, SPESMedica, Salerno, Italy) of eight equi-spaced electrodes (inter-electrode distance 5 mm, electrodes 5 mm · 1 mm) (Masuda et al.,1985; Merletti et al., 2003) were placed at a distance from thepatella of 10%, 20% and 30% (distal, middle and proximal site) ofthe measured anatomical lengths (Fig. 1). At each site, the ori-entation of the array was selected during test contractions bymoving a dry array at different angles until a clear propagation of

the action potentials without evident shape changes was observed(Masuda et al., 1985).

Before electrode placement, the skin was lightly abraded. Toassure proper electrode–skin contact, 20–30 lL of conductive gelwere inserted into the cavities of the adhesive electrode array.Surface EMG signals were amplified bipolarly (EMG amplifier,EMG-16, LISiN – OT Bioelettronica, Torino, Italy; bandwidth10–500 Hz), sampled at 2048 Hz, and stored after 12 bit A/Dconversion.

2.4. Signal analysis

In the fatiguing contraction, EMG signals were divided intoepochs of duration 10% of the time to task failure. For eachepoch, average rectified value and mean power spectral frequencywere estimated from the central single differential channel of thearray while muscle fiber conduction velocity was computed(Farina et al., 2001) from the maximum number of channelsshowing propagation of the action potentials with minimal shapechanges without the presence of the innervations zone (visualselection of the channels). The same channels were used to com-pute the same EMG variables during each of the 2-s contractionsin the recovery phase. Average rectified value is reported as thevalue at the skin surface before amplification.

2.5. Statistical analysis

Three-way repeated measures analysis of variance (ANOVA)was used to assess the dependency of EMG variables on con-traction force (40% and 80% MVC), location of the array on themuscle (distal, middle, proximal), and time interval (10 timeintervals for the fatigue phase and 20 for the recovery phase).Three-way ANOVA was also used to compare the EMG variablevalues in the beginning of the fatiguing contraction and after 20-min recovery (with factors contraction force, location on themuscle, and two time intervals in the beginning and end of thetask). Two-way ANOVA (factors location and two time intervals

392 N. Hedayatpour et al. / Journal of Electromyography and Kinesiology 18 (2008) 390–396

in the beginning of the task and before the second contraction)was used to verify that EMG variables returned to initial valuesbefore the second contraction. Paired t-test was applied to com-pare time to task failure at the two contraction levels and skintemperatures. P-values less than 0.05 were considered significant.Results are reported as mean and standard deviation (SD) in thetext and table and standard error (SE) in the figures.

3. Results

Time to task failure was 70.7 ± 25.8 s (40% MVC) and27.4 ± 16.8 s (80% MVC) (significantly different, paired t-test P < 0.001). Temperature at the beginning of the fatigu-ing contraction was not different from the temperature attask failure and lower than temperature after the 20-minrecovery (for all muscles and forces P < 0.05) (Table 1).EMG variables before the beginning of the second contrac-tion were not different with respect to the beginning of thefirst contraction (recovery of initial values). Fig. 2 showsexample of recorded EMG signals.

Table 1Skin temperature (mean ± SD, �C) at the beginning of the fatiguingcontraction, at the failure point, and after the 20-min recovery

Vastus medialis Vastus lateralis

40% MVC 80% MVC 40% MVC 80% MVC

Beginning 31.6 ± 0.6 32.2 ± 0.7 31.6 ± 0.6 32.1 ± 0.7Failure point 31.8 ± 0.7 32.3 ± 0.7 31.8 ± 0.7 32.3 ± 0.7After 20-min

recovery32.7 ± 0.8 33.1 ± 0.7 32.8 ± 1.1 33.4 ± 0.7

Distal

Middle

Proximal

50 msnu

Vastus Medialis

Fig. 2. Example of signals recorded with the six electrode arrays from vastunormalized units.

3.1. Sustained contractions

EMG average rectified value of the vastus medialisincreased with relative force (F = 11.3, P < 0.01; 40%MVC: 18.0 ± 8.7 lV; 80% MVC: 28.1 ± 15.0 lV),depended on muscle location (F = 3.7, P < 0.05; fromdistal to proximal: 29.9 ± 17.1 lV, 25.1 ± 15.3 lV, 15.8 ±10.9 lV, first different from last P < 0.05), and time interval(F = 2.1, P < 0.05; first larger than last three P < 0.05).Moreover, there was a significant interaction among thethree factors (F = 2.1, P < 0.01), indicating a location-dependent trend of EMG amplitude over time (Fig. 3).Average rectified value for the vastus lateralis muscledepended on contraction force (F = 14.3, P < 0.01; 40%MVC: 19.9 ± 8.2 lV; 80% MVC: 26.4 ± 10.7 lV) and ontime interval (F = 3.1, P < 0.01; first larger than the thirdand subsequent, P < 0.05). The decay over time dependedon the location over the muscle (interaction between timeand location, F = 1.9, P < 0.05).

For both vastus medialis and lateralis mean powerspectral frequency decreased over time (F > 2.2, P < 0.05;in both cases, first time interval larger the last three, P <0.05) (Fig. 4).

Conduction velocity of the vastus medialis depended onlocation on the muscle (F = 6.5, P < 0.05, most distal differ-ent from the other two locations, P < 0.05; 4.9 ± 1.1 m/s,3.1 ± 0.8 m/s, 3.0 ± 0.9 m/s). Moreover, there was an inter-action between contraction force and time interval (F = 3.0,P < 0.01), with conduction velocity decreasing over time(different for each time interval, P < 0.05) at 80% MVC butnot at 40% MVC. Conduction velocity of the vastus lateralis

Vastus Lateralis

s medialis and lateralis muscles during a contraction at 80% MVC. nu:

10

15

20

25

30

35

40

45

50

55

10

15

20

25

30

35

40

45

50

55Endurance Endurance

Ave

rag

e re

ctif

ied

val

ue

(mea

n±S

E, μ

V)

DistalMiddleProximal

Vastus Medialis Vastus Lateralis

Percent time-to-

task failureRecovery (min)

10% 100% 1 20Percent time-to-

task failureRecovery (min)

10% 100% 1 20

Fig. 3. EMG average rectified value (mean ± SE over the 10 subjects) in the 10 time intervals during the sustained contraction (increments of 10% of thetime to task failure) and the 20 time intervals (spaced by 1 min) during recovery. Contraction level 80% MVC.

Percent time-to-

task failureRecovery (min)

65

70

75

80

85

90

95

100

65

70

75

80

85

90

95

100

Task Failure

Task Failure

Mea

n p

ow

er s

pec

tral

fre

qu

ency

(m

ean

±SE

, Hz)

DistalMiddleProximal

VastusMedialis VastusLateralis

10% 100%1 20 10% 100%1 20Percent time-to-

task failureRecovery (min)

Fig. 4. EMG mean power spectral frequency (mean ± SE over the 10 subjects) in the 10 time intervals during the sustained contraction (increments of 10%of the time to task failure) and the 20 time intervals (spaced by 1 min) during recovery. Contraction level 80% MVC.

N. Hedayatpour et al. / Journal of Electromyography and Kinesiology 18 (2008) 390–396 393

depended on time (F = 2.6, P < 0.05) but only the first andlast intervals resulted different (smaller in the last, P < 0.05).

3.2. Recovery

Average rectified value for vastus medialis increasedwith force level (F = 16.5, P < 0.01; 40% MVC:

14.8 ± 6.7 lV; 80% MVC: 24.7 ± 16.2 lV) and changedwith location (F = 3.8, P < 0.05; from distal to proximal:25.1 ± 15.1 lV, 22.9 ± 13.1 lV, 15.9 ± 11.4 lV, first differ-ent from last P < 0.05). Average rectified value for vastuslateralis depended only on force (F = 31.8, P < 0.001;40% MVC: 16.7 ± 5.3 lV; 80% MVC: 27.6 ± 13.1 lV)(Fig. 3).

394 N. Hedayatpour et al. / Journal of Electromyography and Kinesiology 18 (2008) 390–396

For both vastus medialis and lateralis, mean frequencydepended on time (F > 2.0, P < 0.01; first smaller than last,P < 0.05) and on the interaction between time interval andlocation (F > 2.1, P < 0.05), indicating a location-depen-dent trend over time (Fig. 4).

Conduction velocity of vastus medialis depended onlocation (F = 4.1, P < 0.05; most distal larger than mostproximal, 5.2 ± 1.0 m/s vs 4.3 ± 0.8 m/s).

Average rectified value for vastus medialis after 20-minrecovery was smaller than in the beginning of the endur-ance contraction (F = 8.4, P < 0.05; 21.2 ± 4.2 lV vs26.1 ± 5.1 lV). Mean frequency and conduction velocityfor both muscles were larger in the end of the 20-min recov-ery than in the beginning of the endurance contraction(F > 7.4, P < 0.05) (Fig. 4).

4. Discussion

EMG variables and their trends over time during sus-tained contraction and recovery of the vastus medialisand lateralis muscles depended on location over the mus-cles. Moreover, conduction velocity and mean frequencyhad supernormal values after 20-min recovery with respectto the beginning of the task.

4.1. Sustained contraction

EMG amplitude or its trend over time depended onlocation over the two muscles analyzed. Mean frequencywas the same in the three locations and conduction velocitydepended on location only for the vastus medialis muscle.Mean frequency and conduction velocity decreased overtime during the sustained contraction. However, theirtrends over time did not depend on location (no interactionbetween time interval and location). On the contrary, forboth muscles, EMG amplitude changes during the sus-tained contraction depended on location.

Dependence of EMG variables on location is in agree-ment with previous findings on other muscles (Li andSakamoto, 1996; Holtermann et al., 2005). Non-uniformEMG amplitude can be explained by non-uniform fibermembrane properties or non-uniform motor unit recruit-ment. In broad muscles with distributed mechanicalactions, muscle fibers are not all mechanically equivalentwith respect to their direction of force. In the distal portionof the vasti, fibers are more obliquely distributed than inthe proximal portions (Weinstabl et al., 1989; Peeleret al., 2005). This pattern of fiber orientation enables thedifferent parts of these muscles to contribute in varioustypes of activities, such as stabilization of the patella, exter-nal and internal rotation of the tibia and extension of theknee (Goodfellow and O’Connor, 1978). Variations inmorphological and architectural characteristics of musclefiber with location indicates that different parts of the vastimuscles are differently activated during a specific task. Inagreement with the present results, Morrish et al. (2003)observed a greater value of EMG amplitude in the oblique

portion of the vastus medialis than in the other parts of themuscle.

4.2. Recovery

After 20-min recovery, in both muscles, mean frequencyand conduction velocity showed supernormal values whileEMG amplitude in vastus medialis was smaller than in thebeginning of the task. Moreover, the trend of mean fre-quency during recovery was affected by recording location.

Simultaneous increase in mean frequency and conduc-tion velocity indicated that the overshooting of mean fre-quency was partly due to augmented conduction velocity.A long lasting overshoot of conduction velocity was earlierreported on elbow flexors and adductor pollicis muscleafter fatiguing isometric contraction (van der Hoevenet al., 1993; van der Hoeven and Lange, 1994; Milleret al., 1987). Numerous mechanisms have been suggestedfor the increase in conduction velocity over normal values,including changes in muscle temperature and muscle fiberswelling.

An increase in muscle temperature in previous studiesresulted in lower EMG amplitude and higher conductionvelocity (Stewart et al., 2003; Winkel and Jørgensen,1991), as a consequence of faster opening–closing of theNa+ channels in which the diffusion time of Na+ ion isdecreased. In this study, skin temperature increased by�1 �C from the onset of contraction to the end of recoveryin both muscles. This small change is probably not suffi-cient to explain the observed supernormal values (Merlettiet al., 1984).

Muscle fiber swelling is due to osmosis gradient differ-ence between the interstitium and intracellular space ofworking fibers (Lundvall et al., 1972) and is likely to playa role in the observed results, as discussed by van der Hoe-ven et al. (1993). An increase in water content of musclefibers has been reported following maximal fatiguing con-traction (Sahlin et al., 1978; Sjogaard et al., 1985). The pri-mary reason for increased intracellular water is theproduction of lactate during anaerobic exercise (Lundvallet al., 1972).

It was also observed that trends of EMG mean powerfrequency during recovery depended on location. Thismay be explained by a non-uniform metabolic accumula-tion and lactate production. Accumulation of metabolitesdepends indeed on the number of active motor units underanaerobic condition which may be different in differentmuscle regions. During muscle contraction, the metabolicdemands of the different regions of active quadricepsincrease muscle fiber diameter (Nielsen et al., 1990). Clearyet al. (2006) reported larger increase in muscle size in thedistal portion than in other regions of the thigh, 30 minafter downhill running. This might be related to high per-centage of fast twitch fibers in this region of the quadriceps(Elder et al., 1982). Removal of metabolites may alsodepend on location into the muscle due to regional capil-lary and oxidative enzyme supply to muscle fibers (Tesch

N. Hedayatpour et al. / Journal of Electromyography and Kinesiology 18 (2008) 390–396 395

and Wright, 1983). Fiber type sensitivity to low pH andtemperature (Metzger and Moss, 1987) may be anotherreason for the non-uniform recovery.

4.3. Assessment of EMG variables

The variability in EMG variables with location mayhave also been due to factors not related to physiologicalmechanisms, such as non-uniform subcutaneous layerthickness, effect of fiber orientation on the bipolarly filteredEMG, or contact impedance. Variability of these factorswith location cannot be ruled out. The main conclusionis that EMG variables may substantially vary with elec-trode location, in agreement with previous reports (e.g.,Li and Sakamoto, 1996). Results obtained from a singlerecording point may thus not represent the behavior ofthe entire muscle and multiple recording points aresuggested, in particular in the case of large muscles wheredifferent parts may be activated differently depending onthe task.

5. Conclusion

Supernormal values of conduction velocity and meanfrequency were observed in the vasti muscles after recov-ery from sustained contraction. The initial value of EMGamplitude depended on electrode location. The trendsover time of mean frequency during recovery dependedon the location over the muscle, indicating non-uniformrecovery of electrophysiological membrane properties.The results highlight the spatial dependency of electro-physiological mechanisms within the same muscle andthus the necessity of EMG spatial sampling for globalmuscle assessment.

References

Bigland-Ritchie B. Muscle fatigue and the influence of changing neuraldrive. Clin Chest Med 1984;5:21–34.

Bigland-Ritchie B, Jones DA, Hosking GP, Edwards RH. Centraland peripheral fatigue in sustained maximum voluntary contrac-tions of human quadriceps muscle. Clin Sci Mol Med 1978;54:609–14.

Clarkson PM, Tremblay I. Exercise-induced muscle damage, repair, andadaptation in humans. J Appl Physiol 1988;65:1–6.

Cleary MA, Sitler MR, Kendrick ZV. Dehydration and symptoms ofdelayed-onset muscle soreness in normothermic men. J Athl Train2006;41:36–45.

Coyle EF, Costill DL, Lesmes GR. Leg extension power and muscle fibercomposition. Med Sci Sports 1979;11:12–5.

Elder GC, Bradbury K, Roberts R. Variability of fiber type distributionswithin human muscles. J Appl Physiol 1982;53:1473–80.

Farina D, Muhammad W, Fortunato E, Meste O, Merletti R, Rix H.Estimation of single motor unit conduction velocity from surfaceelectromyogram signals detected with linear electrode arrays. Med BiolEng Comput 2001;39:225–36.

Fletcher WM. Lactic acid in amphibian muscle. J Physiol 1907;35:247–309.

Friden J, Sfakianos PN, Hargens AR. Muscle soreness and intramus-cular fluid pressure: comparison between eccentric and concentricload. J Appl Physiol 1986;61:2175–9.

Goodfellow J, O’Connor J. The mechanics of the knee and prosthesisdesign. J Bone Joint Surg Br 1978;60:358–69.

Holtermann A, Roeleveld K, Karlsson JS. Inhomogeneities in muscleactivation reveal motor unit recruitment. J Electromyogr Kinesiol2005;15:131–7.

Ichinose Y, Kanehisa H, Ito M, Kawakami Y, Fukunaga T. Morpho-logical and functional differences in the elbow extensor muscle betweenhighly trained male and female athletes. Eur J Appl Physiol OccupPhysiol 1998;78:109–14.

Ivy JL, Goforth Jr HW, Damon BM, McCauley TR, Parsons EC, PriceTB. Early postexercise muscle glycogen recovery is enhanced with acarbohydrate–protein supplement. J Appl Physiol 2002;93:1337–44.

Lannergren J, Larsson L, Westerblad H. A novel type of delayed tensionreduction observed in rat motor units after intense activity. J Physiol1989;412:267–76.

Li W, Sakamoto K. The influence of location of electrode on muscle fiberconduction velocity and EMG power spectrum during voluntaryisometric contraction measured with surface array electrodes. ApplHuman Sci 1996;15:25–32.

Lundvall J, Mellander S, Westling H, White T. Fluid transfer betweenblood and tissues during exercise. Acta Physiol Scand1972;85:258–69.

Masuda T, Miyano H, Sadoyama T. The position of innervation zones inthe biceps brachii investigated by surface electromyography. IEEETrans Biomed Eng 1985;32:36–42.

Merletti R, Sabbahi MA, De Luca CJ. Median frequency of themyoelectric signal. Effects of muscle ischemia and cooling. Eur J ApplPhysiol Occup Physiol 1984;52:258–65.

Merletti R, Knaflitz M, De Luca CJ. Myoelectric manifestations offatigue in voluntary and electrically elicited contractions. J ApplPhysiol 1990;69:1810–20.

Merletti R, Farina D, Gazzoni M. The linear electrode array: a usefultool with many applications. J Electromyogr Kinesiol 2003;13:37–47.

Metzger JM, Moss RL. Greater hydrogen ion-induced depression oftension and velocity in skinned single fibers of rat fast than slowmuscles. J Physiol 1987;393:727–42.

Miller RG, Giannini D, Milner-Brown HS, Layzer RB, Koretsky AP,Hooper D, et al.. Effects of fatiguing exercise on high-energyphosphates, force, and EMG: evidence for three phases of recovery.Muscle Nerve 1987;10:810–21.

Morrish GM, Woledge RC, Haddad FS. Activity in three parts of thequadriceps recorded isometrically at two different knee angles andduring a functional exercise. Electromyogr Clin Neurophysiol 2003;43:259–65.

Nielsen B, Savard G, Richter EA, Hargreaves M, Saltin B. Muscle bloodflow and muscle metabolism during exercise and heat stress. J ApplPhysiol 1990;69:1040–6.

Parra J, Cadefau JA, Rodas G, Amigo N, Cusso R. The distribution ofrest periods affects performance and adaptations of energy metabolisminduced by high-intensity training in human muscle. Acta PhysiolScand 2000;169:157–65.

Peeler J, Cooper J, Porter MM, Thliveris JA, Anderson JE.Structural parameters of the vastus medialis muscle. Clin Anat2005;18:281–9.

Sahlin K, Alvestrand A, Brandt R, Hultman E. Intracellular pH andbicarbonate concentration in human muscle during recovery fromexercise. J Appl Physiol 1978;45:474–80.

Saltin B, Karlsson J. Muscle glycogen utilization during work of differentintensities. Adv Exp Med Biol 1975.

Sjogaard G, Adams RP, Saltin B. Water and ion shifts in skeletal muscleof humans with intense dynamic knee extension. Am J Physiol1985;248:190–6.

Stewart D, Macaluso A, De Vito G. The effect of an active warm-up onsurface EMG and muscle performance in healthy humans. Eur J ApplPhysiol 2003;89:509–13.

Tesch PA, Wright JE. Recovery from short term exercise: its relation tocapillary supply and blood lactate concentration. Eur J Appl PhysiolOccup Physiol 1983;52:98–103.

396 N. Hedayatpour et al. / Journal of Electromyography and Kinesiology 18 (2008) 390–396

van der Hoeven JH, Lange F. Supernormal muscle fiber conductionvelocity during intermittent isometric exercise in human muscle. J ApplPhysiol 1994;77:802–6.

van der Hoeven JH, van Weerden TW, Zwarts MJ. Long-lastingsupernormal conduction velocity after sustained maximal isometriccontraction in human muscle. Muscle Nerve 1993;16:312–20.

Weinstabl R, Scharf W, Firbas W. The extensor apparatus of the kneejoint and its peripheral vasti: anatomic investigation and clinicalrelevance. Surg Radiol Anat 1989;11:17–22.

Winkel J, Jørgensen K. Significance of skin temperature changes insurface electromyography. Eur J Appl Physiol Occup Physiol1991;63:345–8.

Zipp P. Recommendations for the standardization of lead positions insurface electromyography. Eur J Appl Physiol 1982;50:41–5.

Nosratollah Hedayatapour was born in Shirvan,Iran, in 1972. He graduated in exercise physi-ology from Tehran university, Iran, in 1997.Since 2005, he is enrolled as a Ph.D. candidatein biomedical science and engineering, sup-ported by the Science Ministry of Iran, at theCenter for Sensory-Motor Interaction (SMI),Aalborg, Denmark. He is currently involved inprojects in the field of electromyography andmuscle physiology.

Lars Arendt-Nielsen, born in 1958, received theM.Sc.E.E. degree from Aalborg University,Denmark, in 1983, with specialisation in bio-medical engineering, and the Ph.D. degree in1992. In 1994 he received his Dr.Sci. degree inMedicine from the Medical Faculty, AarhusUniversity, Denmark.From 1983 to 1984 he was a Research fellow,Department of Clinical Neurophysiology, TheNational Hospital for Nervous Diseases, Lon-don. Since 1988 he has been with the Depart-ment of Medical Informatics and Image

Analysis, Aalborg University as an Associated Professor. In 1993 he was

appointed Professor in Biomedical Engineering and Principal investigatorat Center for Sensory-Motor Interaction, which was established in 1993 atAalborg University and in 1997 Head of the International DoctoralSchool in Biomedical Science and Engineering, Aalborg University, with55 Ph.D. students enrolled. During his career he has worked as guestprofessor in Japan and Australia. He is member of the Danish ResearchCouncil and the Danish Research Education Council. He has publishedapprox. 510 scientific papers within neuroscience with focus on motorcontrol and pain research and given more than 110 key-note lectures atinternational conferences.

Dario Farina graduated summa cum laude inElectronics Engineering (equivalent to M.Sc.)from Politecnico di Torino, Torino, Italy, inFebruary 1998. During 1998 he was a Fellowof the Laboratory for Neuromuscular SystemEngineering in Torino. In 2001 and 2002 heobtained the PhD degrees in Automatic Con-trol and Computer Science and in Electronicsand Communications Engineering from theEcole Centrale de Nantes, Nantes, France,and Politecnico di Torino, respectively. In1999–2004 he taught courses in Electronics

and Mathematics at Politecnico di Torino and in 2002–2004 he wasResearch Assistant Professor at the same University. Since 2004, he is

Associate Professor in Biomedical Engineering at the Department ofHealth Science and Technology of Aalborg University, Aalborg, Den-mark, where he teaches courses on biomedical signal processing, mod-eling, and neuromuscular physiology. He regularly acts as referee forapproximately 20 scientific International Journals, is an Associate Editorof IEEE Transactions on Biomedical Engineering, is on the EditorialBoards of the Journal of Neuroscience Methods, the Journal of Elec-tromyography and Kinesiology, and Medical and Biological Engineeringand Computing, and member of the Council ISEK (International Societyof Electrophysiology and Kinesiology). His main research interests are inthe areas of signal processing applied to biomedical signals, modeling ofbiological systems, basic and applied physiology of the neuromuscularsystem, and brain–computer interfaces. Within these fields he hasauthored or co-authored more than 100 papers in peer-reviewed Jour-nals. Dr. Farina is a Registered Professional Engineer in Italy.