Movement Disorder5 Vol. 12, No. 2, 1997, pp. 206-215 0 1997 Movement Disorder Society

Step Initiation in Parkinson’s Disease: Influence of Levodopa and External Sensory Triggers

*Anne Burleigh-Jacobs, *?+Fay B. Horak, $John G. Nutt, and §Jose A. Obeso

Departments of *Physiology and $Neurology, Oregon Health Sciences University, and fR. S. Dow Neurological Sciences Institute, Legacy Good Samaritan Hospitul and Medical Center, Portland, Oregon, U.S.A.; and .C;Neurrrrest Restainracion

Neurologica, Clinica Vintersol, Los Christianos, Tenerife, Spain

Summary: We studied anticipatory postural adjustments con- tributing to gait initiation deficits in patients with Parkinson’s disease (PD) to determine if these deficits could be improved by administration of levodopa or by external stimuli. Ground reaction forces and body kinematics were recorded for self- generated and cutaneous cue-triggered step initiation in normal subjects and in PD subjects when OFF and when ON. The effects of assisting anticipatory postural sway with a surface translation coupled with a cutaneous cue were also examined. Decreased force production, decreased velocity of movement, and slowed execution of the anticipatory postural adjustments for self-generated step characterized step initiation in PD sub- jects when OFF. These impairments were significantly less evident when the PD subjects were ON. Both PD and normal subjects increased force and velocity of movement when a

cutaneous cue was used as a go signal. When subjects volun- tarily initiated a step in response to the surface translation, both PD and normal subjects executed the anticipatory postural ad- justments for step more rapidly, but the PD subjects, both ON and OFF, failed to increase force to execute push-off more rapidly. In conclusion, dopaminergic therapy and an external stimulus similarly improve the deficient force production for the anticipatory postural adjustments associated with step ini- tiation in PD. The findings also suggest that force production during the postural adjustment phase of self-generated, but not externally triggered, step initiation is influenced by dopa- minergic pathways. Key Words: Gait initiation-Parkinson’s disease-Levodopa-Anticipatory postural adjustments-Sen- sory cues-Stepping.

Difficulties with gait initiation are characteristic of Parkinson’s disease (PD). When attempting to voluntar- ily initiate the first step to begin walking, patients often exhibit start hesitation and freezing. Their feet appear to be “stuck” to the floor. When they do begin walking, their movements are slow, and the steps are small. Clini- cal observations suggest that these abnormalities of gait initiation can be responsive to levodopa replacement, al- though start hesitation and freezing can also be exacer- bated by levodopa therapy (1). Clinical observations also suggest that gait initiation in PD patients is sensitive to external stimuli. Auditory stimuli, such as marching mu- sic and visual stimuli, such as lines on the floor, often

Received February 2, 1996, and in revised form May 22, 1996. Accepted May 23, 1996.

Address correspondence and reprint requests to Dr. F. B. H o d at R. S. Dow Neurological Sciences Institutc, 1120 N W 20th Avenue, Portland, OR 97209, U.S.A.

assist patients with difficulty initiating gait. How exter- nal stimuli improve gait initiation in PD is a subject of controversy, however (2,3).

The process of gait initiation requires coordination of anticipatory postural adjustments to move the body mass forward and over the stance limb in preparation for single-limb support during the first step. These postural adjustments involve the production of both vertical and horizontal forces that move the center of pressure (COP) backward and toward the swing limb (4-9). As a result, the body center of mass (COM) moves forward and lat- erally over the stance limb in preparation for foot-off (10). Abnormalities of gait initiation associated with PD include reduced lateral shift of the body mass over the stance limb (1 l), decreased propulsive forces (1 2), and prolonged anticipatory postural adjustments (1 2). It is not known how administration of levodopa or the pre- sentation of external sensory stimuli influences these ab- normal anticipatory postural adjustments associated with gait initiation in persons with PD.

206

STEP INlTlATION IN PARKINSON’S DISEASE 207

The purpose of this study was to identify deficits in the anticipatory postural adjustments that contribute to ab- normal gait initiation in PD subjects with prominent freezing and to determine if these deficits could be im- proved either by administration of levodopa or by the presentation of external stimuli. The influence of dopa- mine deficiency on gait initiation was determined by studying PD subjects both OFF (i.e., with the usual dos- age of levodopa withheld so that parkinsonian features and gait difficulties become prominent) and ON (i.e., with the usual dosage of levodopa administered so that parkinsonian features and gait difficulties are dimin- ished).

The effects of external stimuli were determined by instructing subjects to initiate a step as soon as possible in response to two types of somatosensory stimuli. The first was a small current pulse delivered to either the hand or the earlobe, producing a cutaneous cue. The second was a small backward surface translation produc- ing a slight forward sway of the body, paired with the cutaneous cue. We have previously shown that anticipa- tory postural adjustments are more rapid and forceful when healthy young subjects initiate a step in response to a backward surface translation (13,14). Because the backward surface translation causes forward displace- ment of the body, which promotes a more forceful and longer step, we predicted that the translation would fur- ther promote gait initiation in the subjects with PD.

MATERIALS AND METHODS The experimental protocol, in accordance with the

1964 Helsinki Declaration, was approved by the Internal Review Boards of Legacy Good Samaritan Hospital and Medical Center and of the Oregon Health Sciences Uni- versity. Six subjects (aged 66.7 f 6.2 years; height 173.8 _+ 7.4 cm) with idiopathic PD (Hoehn and Yahr stages 111-IV when OFF) and six neurologically normal age- and height-matched controls (aged 66.3 & 6.4 years; height 179.8 f 8.7 cm) participated after having given informed consent. The duration of disease from the time of diagnosis ranged from 8 to 34 years. All PD subjects could stand and walk independently when ON but ex- hibited prominent gait initiation problems (freezing, start hesitation) when OFF. Three PD subjects reported to the laboratory in the morning, after levodopa/carbidopa (Sin- emet) had been withheld overnight, and were initially tested when OFF. They then took their normal dose of medication and were retested -1 h later when ON. The other three PD subjects reported to the laboratory for the initial testing while ON, and then they were retested - 3 4 later when OFF. All subjects stood on a platform with two force plates that could be translated under the

control of a hydraulic servomotor. The test conditions were as follows.

Condition 1: Self-generated Step Subjects were instructed to take a forward step with

the left foot and follow through with the right. They were to self-initiate and execute the step at any time following notification that the recording device was ready.

Condition 2: Step to Cue Subjects were instructed to take a forward step with

the left foot as soon as they perceived a brief, 4-ms current pulse delivered to the hand or the earlobe and to continue the step through with the right foot. Subjects had no advance notice before onset of the cutaneous cue. No subjects reported the cutaneous stimulus to be pain- ful. No differences in latency of response were noted between the three subjects who had the cue to the hand and the three subjects who had the cue to the earlobe; data were therefore pooled for analysis.

Condition 3: Step to Perturbation Subjects were instructed to take a forward step with

the left foot as soon as they perceived the cutaneous cue and to continue to step through with the right foot. In this condition, the onset of the cutaneous cue was simulta- neous with the onset of a backward surface translation. Subjects had no advance notice before onset of the cue. Both plates of the platform surface were translated hori- zontally 3.6 cm at 5 cm/s in the backward direction, resulting in forward body sway. This relatively small- amplitude, low-velocity translation was chosen to cause a small forward sway that might promote step initiation in the PD subjects. The slow translation did not consis- tently elicit automatic postural responses in the gastroc- nemius, nor did it result in loss of balance or force a compensatory step. Subjects stepped from the moving surface to a fixed surface. All subjects were familiarized with the perturbation before recording by having five trials of the surface translation, during which they re- mained standing on the platform.

For all conditions, subjects stood comfortably with arms folded across their waists, with one foot in the center of each plate and with their weight equally dis- tributed between both feet. To ensure that weight distri- bution and the initial stance position were the same across all trials, the distribution of weight between right and left limbs was viewed on-line. For each trial, sub- jects stood in tracings of their feet made on the platform during quiet stance.

Forces were measured with one horizontal and four vertical strain gauges mounted in each force plate,

Movement Disordes, Vol. 12, No. 2 , 1997

208 A. BURLEIGH-JACOBS ET AL

sampled at 500 Hz and stored for later analysis. The relative position of the total body anterior-posterior and mediolateral COP was determined using the vertical forces from both force plates. The anterior-posterior COP was the difference of anterior and posterior vertical forces divided by the sum of anterior and posterior forces. The result was then multiplied by the distance, in centimeters, from the center of the plates to the location of the vertical force transducers. The mediolateral COP was the difference of total vertical forces between the left and right plates divided by the sum of the vertical forces in both plates. The result was then multiplied by the distance, in centimeters, from the center between the force plates to the center of a single plate. The magnitude and direction of the COP vector were calculated from the forces in the two plates using the peak change in the lateral and posterior components of the COP before heel- off.

Two components of the ground reaction forces of the initial swing limb were also analyzed: the initial increase in the vertical force (Fz) and the horizontal (fore/aft) force (Fx). Before heel-off, an increase in Fz of the initial swing limb contributes to the lateral COP excursion to- ward the swing limb. Forward displacement of Fz pro- duced by ankle dorsiflexion contributes to the posterior excursion of the COP. Changes in Fx are indicative of anterior-posterior horizontal acceleration. Together, these ground reaction forces have the effect of propelling the body COM forward toward the stance limb. Forces were compared among subjects by normalizing changes in Fz and Fx as a percentage of total body weight. A pressure-sensitive resistor, taped to the platform under the heel of the initial swing limb, was used to determine the precise time of heel-off during step initiation. Foot- off was determined as the precise moment there were no vertical forces exerted through the initial swing limb plate. Three phases of the anticipatory postural adjust- ments for step initiation were studied: reaction time phase (cue to initial increase of swing limb Fz), antici- pation phase (initial increase of swing limb Fz to heel- off), and push-off phase (heel-off to foot-off of the initial swing limb).

Body COM position and step length in the sagittal plane were calculated from body kinematics collected at 100 Hz (Northern Digital's Watsmart motion analysis system). Two optoelectric cameras detected the two- dimensional position of infrared light-emitting markers attached to the platform and to the subjects' initial swing limb at the fifth metatarsal head, calcaneus, malleolus, femoral condyle, and greater trochanter and also to the acromian and auditory meatus. Length measurements of each individual subject's hand, forearm, upper arm,

trunk, leg, and foot as well as ankle height were used to determine the COM of each segment based on a frac- tional mass table specific to males and females (15). The fractional mass of each segment relative to the total body mass and the marker coordinates were then used to cal- culate the body COM in the sagittal plane according to the following formula: X,,, = Xfoot . Mfoot + &hank .

Mhead + neck, where X refers to sagittal plane coordinates and M is the fractional mass of each segment relative to total body mass. The forward displacement and velocity of the body COM were determined using the sagittal plane IRED coordinates. Step length was determined from the displacement of the heel IRED.

Planned comparisons were made between control and PD subjects using unpaired t tests and between PD sub- jects ON and OFF using paired t tests. The Bonferoni correction for multiple t tests required an adjusted level of p < 0.025 for statistical significance. Planned com- parisons were also made between the step-to-cue and step-to-translation conditions within each group using paired t tests.

Mshank + Xthigh ' Mthlgh + 'trunk ' Mtrunk + dImS 4- 'head +

RESULTS

General Observations When OFF, all PD subjects exhibited marked diffi-

culty walking, including episodes of freezing, when in the laboratory. However, during the test protocols, the PD subjects were consistently able to step when in- structed to do so. Only two PD subjects were unable to self-generate a step during two trials each when OFF. In these trials, no anticipatory postural adjustment could be discerned from the COP trajectory or from the ground reaction forces.

During successful step initiation, all subjects exhibited an appropriate posteriolateral COP trajectory for the an- ticipatory postural adjustment in all three test conditions. In order to propel the body COM forward and over the initial stance limb, the COP moved backward and toward the swing limb (Fig. 1A). During self-generated step ini- tiation, both the lateral and posterior COP excursions were smaller in PD subjects OFF compared with ON and smaller than in control subjects (Fig. IS). Since both the lateral and posterior COP were reduced similarly, COP excursion in anticipation of a step was in the same di- rection, but smaller in magnitude in PD subjects OFF compared with ON and compared with control subjects (Fig. 1C).

Self-Generated Step Initiation Figure 2 illustrates the increase in swing limb vertical

force anticipatory to step initiation for a representative

Mnwmcwt Dlsovdrrs. Vol. 12, No. 2, 1997

STEP INITIATION IN PARKINSON'S DISEASE 209

A. Schematic of COP excursion B. Net Lateral & Posterior COP vs. time

Initial Swing Limb I Initial Stance Limb I

Control

Posterior COP

PD-OFF \

Time

c. Posterior- lateral peak COP during self-generated step init iat ion

Control PD-OFF PD-ON

FIG. 1. The PD subjects exhibit the appropriate posterior-lateral direction of COP excursion, but the magnitude is reduced when PD subjects are OFF. A: Schematic demonstrates the backward and lateral motion of COP toward the initial swing limb for step initiation. B: The posterior and later- al components of the COP are plotted independently versus time, with peak posterior and peak lateral excursions identified from representative sub- jects. C: The magnitude and direc- tion of the COP vector for each trial is computed from peak posterior and lateral COP.

parkinsonian and normal subject. Although PD subjects were generally able to self-generate step initiation, the PD subjects OFF showed reduced vertical force produc- tion compared with PD ON and compared with the con- trol subjects. Corresponding to the diminished generation

of force in PD subjects OFF, both the anticipation phase and push-off phase of the anticipatory postural adjust- ments were prolonged.

The force production and duration of anticipatory and push-off phases did not differ between control and PD

FIG. 2. Examples of swing limb ver- tical forces (Fz) for a representative PD subject OFF and ON and a height- and weight-matched control subject during three trials of self- generated step. When OFF, the PD subject shows reduced force and pro- longed, variable duration of anticipa- tion phase and push-off phase com- pared with the same PD subject ON and with the control subject. Unla- beled arrows indicate the time of heel-off.

Movement Disorders, Vol. 12, No. 2, 1997

210 A. BURLEIGH-JACOBS ET AL.

subjects ON, suggesting that dopamine deficiency sig- nificantly impaired gait initiation. Reduced anticipatory postural adjustments were characteristic of the entire PD subject group when OFF. The peak swing limb Fz was significantly diminished (p < 0.0005) in the PD subjects OFF compared with the controls (Fig. 3A). The forward velocity of the COM at heel-off was also significantly reduced 0, = 0.01 1) in the PD subjects OFF compared with the controls (Fig. 3B). The decreased velocity for the PD subjects OFF was not related to a smaller forward displacement of the COM, since the forward displace- ment of the COM at heel-off was not significantly dif- ferent between groups (control = 2.6 f 0.2 cm; OFF = 2.2 k 0.2 cm). In addition, the duration of the anticipation phase of postural adjustments (initial increase in the swing limb Fz to the time of heel-off) was prolonged for the PD subjects OFF compared with the controls (p = 0.025) (Fig. 3C). The push-off phase was also longer in the PD subjects than in controls, but the difference did not reach significance (Fig. 3D).

Administration of levodopa improved motor perfor-

FIG. 3. Group means and standard errors of peak percentage increase in Fz (A), COM velocity (B), anticipation (C), and push-off phase (D) for self-generated step are shown. The PD subjects OFF (light shaded bars) show decreased swing limb vertical forces (Fz), decreased veloc- ity of the COM at heel-off, prolonged duration of the anticipation phase, and prolonged duration of the push-off phase compared with the same subjects ON (dark shaded bars) and compared with controls (open bars). (The asterisk indicatesp < 1.01.)

mance in the PD subjects. The peak swing limb Fz was significantly increased by levodopa 0, = 0.006; Fig. 3A). Four PD subjects also showed an increase in Fx with levodopa; however, owing to variability, the in- crease was not significant for the group (data not shown). COM velocity also increased with levodopa but did not reach significance (Fig. 3B). The duration of the antici- pation phase of postural adjustments (Fig. 3C) was sig- nificantly shortened with levodopa 0, = 0.004). Like- wise, shortening of the duration of the push-off phase (heel-off to foot-off) when ON compared with OFF al- most reached significance (p = 0.03; Fig. 3D). The im- provement in these parameters with levodopa erased the significant differences between control subjects and PD subjects when ON (Fig. 3A-D).

Characteristics of Externally Triggered Step Initiation

Step to Cue The cutaneous cue improved the anticipatory postural

adjustments for step initiation in the PD subjects OFF, in a manner similar to levodopa administration. PD subjects when OFF generated greater swing limb vertical forces (Fz) in the step-to-cue condition than in the self- generated step condition (Fig. 4). In contrast, control subjects and PD subjects ON showed similar perfor- mance in the two conditions (Fig. 5A).

When a step was initiated in response to the cue, the force production was similar for PD subjects both OFF and ON and the controls; administration of levodopa did not further enhance motor performance. As the vertical forces and velocity of the COM were increased, the du- ration of the anticipation and push-off phases of postural adjustments were decreased for the step-to-cue condition compared with the self-generated-step condition (lightly shaded bars in background, Fig. 5) . These changes were most prominent in the PD subjects when OFF. The per- formance of the PD subjects when OFF was not signifi- cantly different from either the controls or the same sub- jects ON when the step was initiated in response to the external cue (foreground bars in Fig. 5) .

Step to Perturbation When the onset of a small backward surface transla-

tion was paired with the cutaneous cue, both the PD and control subjects had shortened reaction times for onset of postural adjustments. However, the control subjects, but not the PD subjects, responded to the perturbation with increased force, COM velocity, and step length. Admin- istration of levodopa did not influence the effects of the perturbation on motor performance since there were no

Movement Disorders, Vol. 12, No. 2, 1997

STEP INITIATION IN PARKINSON’S DISEASE 211

Control

Step-to-cue Self-generated

PD-OFF

Step-to-cue Self-generated

differences between the PD subjects ON or OFF for any variable measured.

Reaction times (Fig. 6A) for the step to cue were simi- lar in the controls and PD subjects OFF and ON (control = 260 25 ms; OFF = 287 f 54 ms; ON = 268 f 67 ms) and were similarly shortened (p < 0.01 ) in controls and patients in the step-to-perturbation condition (control = 189 f 20 ms; OFF = 208 k 49 ms; ON = 195 f 18 ms). The duration of the anticipation phase (onset to heel-off) was shortened (p < 0.01) for all subjects in the step-to perturbation condition (Fig. 6B). In contrast, the push-off phase was prolonged for the PD subjects OFF

FIG. 5. The means and standard errors of group data for the externally triggered step-to-cue condi- tion are shown in contrast to the self-generated step data (lightly shaded background bars) shown in Fig. 3 . During step to cue, there are no significant dif- ferences between the controls (open bars) and the PD subjects OFF (light shaded bars) and ON (dark shaded bars) for the peak percentage increase swing limb vertical forces (Fz) (A), velocity of the COM at heel-off (B), duration of the anticipation phase (C), and push-off phase of step initiation (D).

FIG. 4. Examples of swing limb ver- tical force (Fz) for a PD subject OFF and a height- and weight-matched control subject during three trials of externally triggered steps to cue (bold lines) contrasted with three trials for the same subjects during the self- generated steps (light lines). In the step-to-cue condition, the PD subject generates approximately the same amount of force and duration of an- ticipation and push-off phases a5 the control subject. Unlabeled arrows in- dicate the time of heel-off. The data for the self-generated steps are the same as in Fig. 2.

(p = 0.0027) and ON (p = 0.025) compared with the controls (Fig. 6C). Visual observation confirmed that al- though PD subjects quickly initiated heel-off in response to the perturbation, the foot appeared to stick or hesitate before the toe was lifted from the platform.

Control subjects, but not PD subjects, experienced in- creased magnitude of ground reaction forces during an- ticipatory postural adjustments and were therefore able to more rapidly execute foot-off and increase step length when the perturbation imposed a small, passive forward sway of the body (Fig. 7). The peak Fz and Fx were increased in the control subjects when step was initiated

Movement Disorders, Vol. 12, No. 2 , 1997

212 A . BURLEIGH-JACOBS ET AL.

FIG. 6. The means and standard errors of group data are shown, demonstrating the shortened reaction time phase (A) and shortened anticipation phase (B) of the postural adjustments for all subjects in the step-to-perturbation compared with the step-to-cue condition. In contrast to control subjects, the duration of the push-off phase (C) during step to perturbation was increased in the PD subject5 both ON and OFF. (The asterisk indicates a significant difference compared with controls; p < 0.025.)

in response to perturbation (p < 0.03) but remained un- changed in the PD subjects (Fig. 7). Accordingly, the swing limb peak vertical force (Fz) was significantly less in the PD subjects OFF (p = 0.002) and ON (p = 0.025) compared with the controls in the step-to-perturbation condition (Fig. 7A). The swing limb peak horizontal force (Fx) also tended to be less in the PD subjects com- pared with the controls, but these differences were not significant (Fig. 7B).

The velocity of the COM at heel-off was increased during the step to perturbation in the control subjects (p = 0.018) but not in the PD subjects, corresponding to their decreased force production. The COM velocity was significantly less in the PD subjects both OFF (p = 0.002) and ON (p = 0.017) compared with the controls for the step-to-perturbation condition (Fig. 7C). Like- wise, when a step was initiated in response to the per- turbation, the control subjects increased their step length, whereas PD subjects did not (Fig. 7D). Only the control group showed a significant difference between condi- tions (p = 0.016) when step length was normalized to the step-to-cue condition.

DISCUSSION We have shown that subjects with PD have decreased

force production and delayed execution of anticipatory postural adjustments for self-generated step initiation. These impairments were most pronounced when the PD subjects were OFF and improved with the administration of levodopa, suggesting that they are, in part, related to dopamine deficiency. When an external cutaneous cue was used as a go signal, control subjects as well as PD subjects ON and OFF increased force and velocity of their anticipatory postural adjustments. When the cue

was combined with a perturbation that caused forward sway of the body, all subjects exhibited shortened reac- tion times and shortened durations of anticipatory pos- tural adjustments for step initiation. However, unlike the controls, the PD subjects, both ON and OFF, failed to further increase force to more rapidly execute foot-off in response to postural perturbations. These findings sug- gest either a nondopaminergic mechanism or a failure of levodopa administration to fully normalize dopaminergic mechanisms for externally triggered postural responses.

Influence of Levodopa Administration for Self-generated Step Initiation

Our results show that subjects with Parkinson’s dis- ease can execute a motor command for self-generated step initiation, but the execution is impaired owing to deficient force production. These findings are similar to those of Ingvarsson et al. (1 1 ) and Crenna et al. (12), but, in addition, we show that these impairments are im- proved by levodopa or by an external cue. Our findings are also consistent with the slowing of gait, shorter steps, and prolonged time in the stance or double-support phase of the step cycle (16,17). Ineffective force production results in a slower displacement of the body COM. If stable alignment of the body COM toward the stance limb is delayed, then unloading of the swing limb will also be delayed (14,18,19). We find that unloading of the swing limb is delayed in the PD subjects OFF, probably related to decreased propulsive forces during the postural adjustment phase of self-generated step initiation. A flexed posture cannot be the primary explanation for re- duced force production since performance is markedly improved with an external stimulus despite the persis- tence of a flexed posture.

Movemmt DisordnT, Vol. 12. No. 2 , 1997

STEP INITIATION IN PARKINSON'S DISEASE 213

A. Peak Swing Limb Fz B. Peak Swing Limb Fx

FIG. 7. The means and standard errors of group

conditions are shown In contra\t to the control sub- jects, the PD subjects fail to show increases in ver- Step-to-cue Step-to-perturbation Step-to-cue Step-to-perturbation tical force (Fz) (A), honzontal force (Fx) (B), ve- locity of the COM at heel-off (C), and step length dunng the step-to-perturbation condition (D) (The asterisk indicates a significant difference compared with controls; p < 0 025 )

'0 17 data for the step-to-cue versus step-to-perturbation Z

I - C. COM Velocity at Heel-off D. Step-Length

I Step-to-cue Step-to-perturbation Step-to-cue Step-to-perturbation

Many studies suggest that one essential component of bradykinesia exhibited by PD subjects during voluntary arm movements is an inability to effectively generate muscle activity (20-23). These findings for voluntary movement are consistent with the diminished force pro- duction for anticipatory postural adjustments associated with voluntary rise to toes (24) and voluntary step ini- tiation in the current study. Administration of levodopa is associated with increased velocity of upper-extremity voluntary movement (21,25,26). Similarly, we show that velocity of body movement is increased during self- generated step initiation in our study and during rise to toes (24) with the administration of levodopa.

The deficient force production during self-generated step initiation in the PD subjects OFF does not explain the phenomenon of freezing. Although our subjects with PD were selected because of problems with freezing, in our experimental paradigm, freezing was seldom ob- served. It is possible that the division between the force plates and the stationary platform, the heightened level of

attention, andfor the instruction to step rather than walk may have contributed to fewer episodes of freezing in the experimental paradigm. The occasional inability on the part of two subjects to initiate a step during the experi- ment was associated with no anticipatory postural adjust- ments at all, suggesting that freezing is related to lack of initiation of postural adjustments rather than to weak, ineffective patterns (27).

Influence of an External Sensory Trigger for Step Initiation

Presentation of an external cue resulted in increased force production and movement velocity, similar to the improvements seen with the administration of levodopa for self-generated step initiation. When a cutaneous cue was used to trigger step initiation, the forces, timing, and velocities associated with the anticipatory postural ad- justments were no longer different between the PD and control subjects. Administration of levodopa had no fur- ther effect on motor performance, which is not surprising

Movement Disorders, Vol. 12, No. 2, 1997

214 A. BURLEIGH-JACOBS ET AL.

since the PD subjects had near normal performance in this externally triggered behavior.

Several possible mechanisms may explain the im- proved postural performance from an external cue. The subject’s attention level may have been heightened in anticipation of the cue, resulting in a more rapid and forceful movement. Alternatively, the external cue may have substituted for a dopaminergic influence on the fa- cilitation of centrally initiated voluntary movement (28). The most attractive explanation is that different neural mechanisms mediate self-generated (internally triggered) and reaction-time (externally triggered) movements (24,

Electrophysiological correlates suggest that the supplementary motor area (SMA) is involved in the ini- tiation of self-generated movements, whereas other cor- tical regions, including the premotor cortex, are involved in the generation of movement in response to a visual (29,30,33) or auditory stimulus (34). Deficiencies of self- generated movement in PD agree with this hypothesis, since one major output of the basal ganglia is to the SMA (35). Our findings show that a somatosensory stimulus can provide an external trigger, demonstrating that the phenomenon is not specific to the visual and auditory pathways.

Our findings further indicate that the motor control mechanisms for step initiation are intact in PD and that normal anticipatory postural adjustments can be elicited using an external sensory trigger. These findings are similar to the improvement of parkinsonian gait with the presentation of visual cues (36). When subjects are pro- vided with external (visual) cues, both the temporal and spatial parameters of step initiation improve to near nor- mal values.

29-32).

Failure to Increase Force Production with Change in External Conditions

When a platform perturbation was paired with the cu- taneous cue, reaction times were faster for all subjects, suggesting that combined cutaneous and fast propriocep- tive stimuli can enhance reaction times for step initiation, even in subjects with PD. In addition, the small forward sway induced by the perturbation promoted a more rapid execution of heel-off for all subjects. However, unlike the control subjects, the perturbation did not further en- hance step initiation in the PD subjects because they could not increase the force needed for rapid foot-off and longer step length.

It is not clear why the PD subjects, even when ON after levodopa administration, failed to increase force, velocity of movement, and step length in response to forward sway perturbations. It is possible that the PD

subjects were using more caution or had already reached their maximal force production. It is also possible that the external trigger served only to initiate a motor com- mand, and a separate mechanism is involved in altering or modifying force production for the first step, after the command is initiated. However, since administration of levodopa also has no influence on automatic, externally triggered postural responses (32,37,38), we suggest that either nondopaminergic mechanisms or inactivated do- paminergic mechanisms play a part in the disturbed re- active postural responses found in PD subjects (24,32).

Although Parkinson’s disease provides a model to study disruption of the basal ganglia dopaminergic sys- tem, the disease process involves multiple other systems that may also contribute to motor problems (39-41). With progression of the disease, the emerging symptoms are often less responsive to levodopa therapy, suggesting that increased disability results from disruption of non- dopaminergic systems (42) or failure of levodopa admin- istration to fully replicate normal dopaminergic mecha- nisms. The responsiveness to levodopa also appears to depend on the extent to which voluntary effort is made (43), suggesting a mechanism by which the administra- tion of levodopa has a more prominent effect on the cortical (i.e., self-generated) control of movement and less effect on movements associated with automatic or externally triggered responses.

In conclusion, both dopdminergic replacement therapy and the presentation of an external stimulus improve force production for the anticipatory postural adjust- ments associated with step initiation in subjects with Par- kinson’s disease. The findings also suggest that force production during the postural adjustment phase of self- generated, but not externally triggered step initiation is influenced by dopaminergic pathways.

Acknowledgment: The authors extend their appreciation to the subjects for their participation in this study and to Dr. Ron Jacobs for his suggestions for data analysis and contributions to the discussion of the findings. This research was supported by a fellowship froin the Foundation for Physical Therapy awarded to Anne Burleigh and by a grant from the National Institute on Aging (AG-06457) awarded to Fay Horak.

REFERENCES

1. Ambani LM, VanWoert MH. Start hesitation: a side-effect of long- term levodopa therapy. N Engl J Med 1973;288:1113-1115.

2. Martin JP. The basal ganglia and posture. London: Pitman Medi- cal Publishers, 1967.

3. Mestre D, Blin 0, Serratrice G. Contrast sensitivity is increased in a case of nonparkinsonian freezing gait. Neurology 1992;42: 189- 194.

4. Mann RA, Hagy JL, White V, Liddell D. The initiation of gait. J Bone Joint Surg 1979;61 a(2):232-239.

Movement Disorders, Vol. 12, No. 2, 1997

STEP INITIATION IN PARKINSON’S DISEASE 215

5 . Breniere Y, Do MC, Sanchez J. A bioniechdnical study of the gait initiation process. J Biophys Med Nucl 1981;5(4): 197-205.

6. Breniere Y, Do MC, Bouisset S. Are dynamic phenomena prior to stepping essential to walking? J Motor Behav 1987; 19( 1):62-76.

7. Nissan M, Whittle MW. Initiation of gait in normal subjects: a preliminary study. J Biomed Eng 1990;12:165-171,

8. Breniere Y, Do MC. Control of gait initiation. .I Motor Behav I991;23(4):235-240.

9. Crenna P, Frigo C. A motor programme for the initiation of for- ward-oriented movements in humans. J Physiol 1991;437:635- 653.

10. Jian Y, Winter DA. Ishac M, Gilchrist L. Trajectory of the body COG and COP during initiation and termination of gait. Gait Pos- ture 1993;1(1):9-22.

11. Ingvarsson PE, Johnels B, Steg G. Dissolution of motor program coordination in Parkinson’s disease. In: Grillner S , Stein PG, Stuart DG, Forssberg H, Herman RM, eds. Neurobiology of vertebrate locomution. New York: MacMillan Publishing, 1986:105-118.

12. Crenna P, Frigo C, Giovannini P, Piccolo I. The initiation of gait in Parkinson’s disease. In: Berardelli A, Benecke R, Manfredi M, Marsden CD, eds. Motor disturbances, vol 2. London: Academic Press, 1990:161-173.

13. Burleigh AL, Horak FB, Malouin F. Modification of postural re- sponses and step initiation: evidence for goal-directed postural interactions. J Neurophysiol 1994;72(6):2892-2902.

14. Burleigh AL, Horak FB. Influence of instruction, prediction and afferent sensory infotmation on the postural organization of step initiation. ./ Neurophysiol 1996;75: 1619-1628.

15. Diffrient N, Tilley AR, Bardagly JC. Huniunscale 1013. Boston: MIT Press, 1974.

16. Murray MP, Sepic SB, Gardner GM, Downs WJ. Walking patterns of men with parkinsonism. Am J Phys Med 1978;57:278-294.

17. Knutsson E. An analysis of parkinsonian gait. Brain 1972;95:475- 486.

18. Rogers MW, Pai YC. Patterns of muscle activation accompanying transitions in stance during rapid leg flexion. J Electromyogr Ki- nesiol 1993;3(3):149-156.

19. Elble RJ, Moody C, Leffler K, Sinha R. The initiation of normal walking. Mov Disord 1994;9(2): 139-146.

20. Hallett M, Khoshbin S . A physiological mechanism of bradykine- sia. Bruin 1980;103:301-314.

21. Berardclli A, Dick JPR, Rothwell JC, Day BL, Marsden CD. Scal- ing of the size of the first agonist EMG burst during rapid wrist movements in patients with Parkinson’s disease. J Neurof Neuro- surg Psychiutry 1986;49: 1273-1279.

22. Marsden CD. Slowness of movement in Parkinson’s disease. Mov Disord 1989;4(suppl):2&37.

23. Godaux E, Koulischer D, Jacquy J. Parkinsonian bradykinesk is due to depression in the rate of rise of muscle activity. Ann Neurol 1992;31:93-100.

24. Horak F, Frank J . Three separate postural systems affected in parkinsonism. In: Stuart DG, Gurfinkel VS, Wiesendanger M, eds. Motor control, vol. 7. Tucson: Motor Control Press 1996: 343-346.

25. Weinrich M, Koch K, Garcia F, Angel RW. Axial versus distal

motor impairment in Parkinson’s disease. Neurology I988;38:540- 545.

26. Johnson MTV, Mendez A, Kipnis AN, Silverstein P, Zwiebel F, Ebner TJ. Acute effects of levodopa on wrist movement in Par- kinson’s disease: kinematics, volitional EMG modulation and re- flex amplitude modulation. Brain 1994; 117: 1409-1422.

27. Winter D. ABC (anatomy, biomechanics and control) of balance during standing and walking. Waterloo, Ontario: Waterloo Bio- mechanics, 1995.

28. Marsden CD, Obeso JA. The functions of the basal ganglia and the paradox of stereotaxic surgery in Parkinson’s disease. Brain 1994; 117:877-897.

29. Passingham RE. Two cortical systems for directing movement. In: Motor areas ofthe cerebral cortex. Chester, England: Wiley, 1987: 151-164. (Ciba Foundation Symposium 132).

30. Kurata K, Wise SP. Premotor and supplementary motor cortex in rhesus monkeys: neuronal activity during externally- and internally instructed motor tasks. Exp Brain Res 1988;72:237-248.

31. Papa SM, Artieda J, Obeso JA. Cortical activity preceding self- initiated and externally-triggered voluntary movement. Mov Dis- urd 1991;6(3):217-224.

32. Horak F, Frank J. Nutt J. Effects of dopamine on postural control in parkinsonian subjects: scaling, set and tone. .I Neurophysiol 1996;75:2380-2396.

33. Glickstein M, Stein J. Paradoxical movement in Parkinson’s dis- ease. Trends Neurosci 1991; 14( 1 1):480482.

34. Deiber MP, Passingham RE, Colebatch JG, Friston KJ, Nixon PD, Frackowiak RSJ. Cortical areas and selection of movement: a study with positron emission tomography. Exp Brain Res 1991;84: 393402.

35. Parent A. Extrinsic connections of the basal ganglia. Trends Neu- rosci 1990;13(7):254258.

36. Morris ME, Iansek R, Matyas TA, Summers JJ. The pathogenesis of gait hypokinesias in Parkinson’s disease. Brain 1994;117:1169- 1181.

37. Traub MM, Rothwell JC, Marsden CD. Anticipatory postural re- flexes in Parkinson’s disease and other akinetic-rigid syndromes and in cerebellar ataxia. Brain 1980;103:393412.

38. Horak FB, Nutt JG, Nashner LM. Postural inflexibility in parkin- sonian subjects. J Neurol Sci 1992;111:4&58.

39. Greenfield JG, Ragsdale CW Jr. The brainstem lesions in parkin- sonism. J Neurol Neurosurg Psychiatry 1953;16:213-226.

40. Hornykiewicz 0, Kish SJ. Biochemical pathophysiology of Par- kinson’s disease. In: Yahr MD, Bergmann KJ, eds. Advances in neurology, vol. 45. New York: Raven Press, 1986:19-34.

41. Carlsson M, Carlsson A. Interactions between glutaminergic and monoaminergic systems within the basal ganglia: implications for schizophrenia and Parkinson’s disease. Trends Neurosci 1990; 13 (7):272-276.

42. Bonnet AM, Loria Y, Saint-Hilaire MH, Lhermitte F, Agid Y. Does long term aggravation of Parkinson’s disease result from non-dopaminergic lesions’? Neurology 1987;37: 1539-1542.

43. Knutsson E, Martensson A. Quantitative effects of L-dopa on dif- ferent types of movements and muscle tone in parkinsonian pa- tients. Scand J Rehab Med 1971;3:121-130.

Movement Disorders, Vol 12, No 2, 1997