Impaired scaling of long latency postural reflexes in patients with Parkinson's disease

22

Electroencephalography and clinical Neurophysiology , 89 (1993) 22-28 © 1993 Elsevier Scientific Publishers Ireland, Ltd. 0924-980X/93/$06.00

ELMOCO 92100

Impaired scaling of long latency postural reflexes in patients with Parkinson's disease D.J. Beckley a,,, B.R. B l o e m

a,b,**

and M.P. R e m l e r a

a Department of Neurology, Veteran's Administration Medical Center, Martinez, CA, and School of Medicine, UC Davis, Davis, CA (USA), and a Department of Neurology and Clinical Neurophysiology, University Hospital Leiden, Leiden (The Netherlands)

(Accepted for publication: 14 October 1992)

Summary Young normal subjects adapt the size of posturally stabilizing reflexes in the lower extremity to predictable and unpredictable perturbations through shifts in cognitive set. It is unlcnown whether limitations in this ability to shift cognitive set may contribute to impaired scaling of postural reflexes in patients with Parkinson's disease. In this study, we have addressed this issue in 12 posturally unstable Parkinson patients and 13 age- and sex-matched controls. Postural stability was disturbed by sudden toe-up rotations of a supporting Platform upon which subjects were standing. Subjects' cognitive set was altered by varying the perturbation amplitude either predictably (serial 4° versus serial 10°) or unpredictably (random mixture of 4 ° and 10°). Posturally stabilizing long latency (LL) reflexes were recorded from the shortened tibialis anterior muscle of both legs. We found that Parkinson patients, unlike some control subjects, were unable to scale the size of their LL reflex in response to variations in perturbation amplitude during predictable conditions. In addition, we observed that Parkinson patients could not modify the amplitude of the LL reflex through alterations in cognitive set during random conditions. We conclude that Parkinson patients have a fundamental difficulty in modifying the size of posturally stabilizing LL reflexes, as reflected by both problems with amplitude scaling and difficulties with changes in cognitive set. It is possible that this inability to modify LL reflexes may be a factor contributing to postural instability in Parkinson's disease. Key words: Parkinson's disease; Long latency reflex; EMG; Cognitive set; Posture; Balance

The mechanisms underlying postural instability in Parkinson's disease (PD) are insufficiently understood. Abnormal control of postural reflexes in leg muscles seems to be one factor which contributes to balance impairment in PD. Examples of abnormal postural reflexes in standing Parkinson patients include enhancement of destabilizing stretch reflexes (Scholz et al. 1987; Beckley et al. 1991a; Schieppati and Nardone 1991) and underscaling of anticipatory postural reflexes (Traub et al. 1980; Dick et al. 1986). Traditional paradigms for assessment of postural reflexes in standing subjects involve translations or rotations of a supporting platform which elicit a characteristic distal-to-proximal reflex activation in leg muscles (Nashner 1976, 1977). For example, in response to toe-up rotations, short latency (SL) and

Correspondence to: Dennis J. Beckley, M.D., Department of Neurology, 127, VAMC, 150 Muir Road, Martinez, CA 94553 (USA). Tel.: (510) 372 2577; Fax: (510) 229 4897.

* Supported by a Veteran's Administration Research Grant. ** Supported by Hoffman-La Roche Nederland and Schering Nederland.

medium latency (ML) responses are seen in the stretched gastrocnemius muscle, while long latency (LL) responses occur in the shortened tibialis anterior muscle. In this specific paradigm, SL and ML responses aggravate the backward-induced body sway, whereas LL responses contribute to postural stability (Allure and Bfidingen 1979; Allum 1983; Diener et al. 1983, 1984). It has been previously shown that the amplitude of posturally stabilizing LL responses in normal subjects has a linear relationship with perturbation amplitude (Diener et al. 1984). In addition, unlike a "pure" reflex the LL response can be modified by changes in postural set (Diener et al. 1987; Nardone et al. 1990) and cognitive set (Beckley et al. 1991c). We have recently used this toe-up rotational paradigm to study whether variations of perturbation amplitude influenced the stabilizing LL responses of young normal subjects (Beckley et al. 1991c). When exposed to predictable small or large perturbations, young controls consistently generated different LL scaling for the different test conditions. Moreover, in response to randomly mixed small and large perturbations, young controls selected as a default response the LL response that anticipated the larger of the two possible perturbations.

IMPAIRED SCALING OF LL REFLEXES IN PD

In the present study, we have extended this paradigm to a group of posturally unstable Parkinson patients in order to determine if amplitude scaling and cognitive set modification of LL responses are abnormal. Since Parkinson patients have problems with the ability to make an appropriate response to a given signal, based upon discrimination of the situational context of a task or "cognitive set" (Buchwald et al. 1975; Cools et al. 1984; Flowers and Robertson 1985; Prasher and Findley 1991), we predicted that Parkinson patients would be unable to consistently select an adequate size default LL response in reaction to unpredictable sized perturbations. Additionally, problems with generation of an initial EMG burst amplitude during reaction time movements (Hallett and Koshbin 1980; Berardelli et al. 1986) were expected to lead to underscaling of functionally corrective LL responses during predictable variations of perturbation amplitude. Preliminary portions of this paper have previously been published in abstract form (Beckley et al. 1990).

Subjects and methods

23 TABLE I Individual clinical characteristics of the 12 Parkinson patients. Patient

73/M 66/M

2.5 3.0

62 43

28 29

3

67/M

2.5

22

27

4

66/M

3.0

75

27

5

80/F

2.5

6

56/M

4.0

28 74

29 30

7

77/M

3.0

71

27

8 9

57/M 63/M

2.5 3.0

47 60

29 29

10

75/M

2.5

34

30

11

72/M

3.0

56

29

12

44/F

2.5

38

30

Mean± S.D. 66.3_+ 10.3 2.8_+0.4 Controls

Subjects Subjects included 12 Parkinson patients (10 males, 2 females) and 13 age- and sex-matched healthy volunteers (9 males, 4 females) (Table I). Disease duration ranged from 2 to 12 years (mean 6.1 years + 2.3). Disease severity was rated using the Hoehn and Yahr rating scale (Hoehn and Yahr 1967) and the Unified Parkinson's Disease Rating Scale (UPDRS) (Fahn et al. 1987). All patients had impaired righting reflexes in response to a sudden sternal push. Cognitive function was tested using the Mini-Mental State Examination (MMSE) (Folstein et al. 1975). There was no significant difference between patients and controls (Table I). Patients with a prominent postural tremor or other neurological disorders were excluded. Antiparkinsonian medication was continued during the study. Patients were tested 30 rain after taking their first morning dose of medication. All subjects gave oral and written consent approved by the Institutional Review Board of the Martinez Veteran's Administration Hospital.

Experimental paradigm The experimental paradigm is described in detail elsewhere (Beckley et al. 1991c) and will only briefly be outlined here. Subjects stood with their feet 10 cm apart on a movable forceplate platform (NeuroCom) and were told to look straight ahead. No pretest information regarding perturbation type, amplitude, velocity, direction or sequence was provided. Stimuli consisted of 3 different protocols, each of 20 "toe-up" forceplate rotations of pre-seleeted amplitude and pre-

Age/sex H and Y a UPDRS b MMSE c Meds

1 2

65.7_+5.8 -

LD/CD d LD/CD; A e ACh f LD/CD ACh LD/CD DA g LD/CD LD/CD ACh LD/CD DA LD/CD LD/CD ACh LD/CD ACh LD/CD DA LD/CD DA; ACh

50.8_+ 18.2 28.7_+ 1.2 29.4 _+0.7

a Hoehn and Yahr scale; b motor score of the Unified Parkinson Disability Rating scale; c Mini-mental State exam; d levodopa/ carbidopa; e amantadine; ~ anticholinergics; g dopamine agonist.

dictability: (a) serial predictable - 4 ° amplitude; (b) serial predictable - 1 0 ° amplitude; (c) random mixture of 4°/10 ° amplitude, all at a constant velocity of 50°/see. Although no pre-information was provided, serial perturbations became predictable through repetition. The sequence of the 3 sets was randomly varied between subjects to reduce a potential confounding effect of ordering. These 3 protocols produced 4 different testing conditions: 4° predictable (4P), 4° random (4R), 10° predictable (10P) and 10° random (10R). Surface EMG recordings were obtained from the tibialis anterior muscle of both legs. SL and ML responses in the medial gastrocnemius muscle were not analyzed since in normal man these responses are not modified by predictable and unpredictable variations of perturbation amplitude (Beckley et al. 1991c). Raw EMG data were pre-amplified, band-pass filtered (10 Hz-1 kHz), full-wave rectified, integrated over a time constant of 4.5 msec and stored to computer disk for off-line analysis (RC Electronics). Onset latencies of LL responses in the tibialis anterior muscle were measured relative to the falling edge of the trigger pulse, which coincided with the onset of forceplate movement. Onset latencies were determined by a reference window of 100-200 msec post-trigger. The mean amplitude of the LL response was calculated over the first 75 msec following onset. Pre-stimu-

24

lus background muscle activity (BGR) was defined as the period of 50 msec preceding the onset of platform movement. LL amplitudes were converted to standardized Z scores to permit comparison of EMG responses between subjects within each group over the range of testing conditions (Hansen et al. 1988; Beckley et al. 1991c). While Z scores permit within group comparisons of EMG responses, they cannot be used for direct comparisons of absolute EMG amplitudes between the two different groups. Therefore, to evaluate whether abnormal scaling of LL responses in PD could be ascribed to a consistent under- or overscaling of reflex amplitudes, LL amplitudes were normalized by computing a ratio of mean LL activity to the first 50 msec of BGR ( L L / B G R ratio) for each individual and for each condition (Cody et al. 1986; Rothwell et al. 1986).

Statistical analysis All calculations were performed using the data measured during individual trials. Statistical analysis consisted of 2-factor (amplitude × cognitive set) repeated measures ANOVA which independently compared the mean LL Z scores for each group across the 4 testing conditions (4P, 4R, 10P, and 10R) (SYSTAT v5.0). This was followed by post hoc multiple comparisons with linear contrast and appropriate Bonferroni adjustments if the univariate F statistic had a significant P value of < 0.05. Onset latencies and normalized LL amplitudes ( L L / B G R ratios) were each separately analyzed by 3-factor repeated measures ANOVA with one between subject variable (group) and two within subject variables (amplitude × cognitive set). Statistical analysis showed no significant difference in the results between the left and right leg. Therefore, only the results obtained from the left leg will be used to illustrate the results.

D.J. B E C K L E Y E T AL.

test conditions (10P > 4P; P < 0.01) (Fig. 2a). During the random conditions, controls selected a default LL response (significant interaction effect between amplitude and cognitive set, P < 0.05) which equalled the one needed for a small perturbation (10P > 10R, 4R, and 4P; P < 0.05). There was no overlap of the 95% confidence intervals between LL amplitudes in the 10P condition and LL amplitudes in the other conditions. In contrast to controls, Parkinson patients could not consistently scale the LL response during either predictable ( P = 0.14) or unpredictable conditions ( P = 0.52 for amplitude × set) (Fig. 2b). Furthermore, 95% confidence intervals revealed a considerable overlap in LL amplitudes across all 4 conditions. Additionally, individual analysis comparing LL amplitudes of the 4P and 10P conditions showed that 5/13 (42%) of controls could scale their LL responses during the predictable conditions in comparison to only 1/12 (8%) of PD patients. Follow-up data analysis from our previous study of younger normal subjects (Beckley et al. 1991c) indicated that 7/10 (70%) were able to scale their LL responses during the predictable conditions.

Comparison of absolute LL ampfitudes between groups Comparison of normalized responses ( L L / B G R ratios) between the two groups showed that LL responses in controls were significantly increased compared to Parkinson patients across all 4 conditions (P < 0.01) (Fig. 3) and within each individual condition (P < 0.01; LSD t test). BGR in the TA muscle was similar in both

TRIGGER

iovo,

CONTROL

Results

0/JVolt

TEL

Onset latencies Reliable LL responses could be discerned in all trials of both patients and controls. Representative LL responses of a control subject and a Parkinson patient are shown in Fig. 1. Onset latencies of LL responses did not differ across the 4 test conditions (4P, 4R, 10P and 10R) in either patients or controls. Comparison of LL responses between groups showed that LL latencies were significantly increased in Parkinson patients (148.9 + 12.7) compared to controls (133.4 + 8.6; P < 0.001,).

Scaling of LL amplitudes In the control group, LL amplitude Z scores varied directly with perturbation size during the predictable

500 pVolt

PARKINSON 0 #Volt

LL I

-2 0

-100

100

0

i

1

2 0

300

400

100 m s e c / d l v . Fig. 1. Average of 20 trials in the 4 ° predictable condition for a control subject and a patient with Parkinson's disease. Vertical arrows indicate onset of the LL responses in the tibialis anterior (TA). The indentation in the top trace represents the trigger pulse, the falling edge of which marks the onset of forceplate movement (dotted vertical reference line) relative to which response latencies are measured.

I M P A I R E D S C A L I N G O F LL R E F L E X E S IN PD

25 PARKINSON

OLDER CONTROLS ( N = 1 3 )

SUBJECTS (N=12)

~J O £P

os

[/3 [-q

0.4

r~7

0

06P

r'h

C2 -

T j_~ 0.2

~

-

-

IOR

j

:

F

D

10Pmr

[--

10P

/

o.o ,,1 <:~

o2 .

--1 ,.-2 z .< 7.q

oa

l 4P

4R

o6

i SERIAL

(a)

•

<

4° o 10

o~p

-a

TOE-UP

Z

FORCEPLATE AMPLITUDE

bJ

4P J c 6 :~

t

•

10°

t

FORCEPLATE AMPLITUDE L

SERIAL

TEST CONDITIONS

4°

TOE UP J

RANDOM

-

TEST CONDITIONS

J

RANDOM

(b)

Fig. 2. a: comparison of the grand m e a n and standard error of the m e a n for tibialis anterior (TA) LL Z scores for all control subjects for each of the two levels of both independent variables - forceplate amplitude (4 ° and 10°) and cognitive set (predictable or "serial" and unpredictable or "random"). There is a significant interaction effect ( F (1, 12) = 8.31, P < 0.05) for the 10° random condition (10R) indicating that this response is strongly influenced by the degree of perturbation predictability. The difference in m e a n amplitude between the predictable and unpredictable 4 ° perturbation is not significant, b: comparison of the grand m e a n and standard error of the m e a n for tibialis anterior (TA) LL Z scores for all Parkinson subjects for each of the two levels of both independent variables - forceplate amplitude (4 ° and 10°) and cognitive set (predictable or "serial" and unpredictable or "random"). There is no significant main effect or interaction effect, indicating that there is no significant difference in the m e a n amplitudes between any of the 4 conditions (4P, 4R, 10P, 10R).

groups, indicating that the increased L L / B G R ratio was attributable to a larger corrective LL response in the older control subjects in comparison to patients.

Discussion

The question addressed by the present study was whether Parkinson patients can modify their LL postural reflexes in response to changes in perturbation amplitude or cognitive set. We found that Parkinson patients, unlike age-matched controls, are unable to scale the amplitude of their LL postural reflexes to

CONFIDENCE INTERVALS ABOUT THE MEAN

80

z

60

r~ -I a

40

,? ±

T

zo

99% CII.

E2395%

0

i

i

CONTROLS

•

C.L

MERN

PD'

GROUP

Fig. 3. Confidence intervals (CI) of the grand m e a n normalized LL response, averaged over all 4 conditions for the control subjects and the Parkinson patients. Control's m e a n LL normalized response = 60.29 (99%, CI 44.10-76.48); Parkinson patient's m e a n LL normalized response = 28.02 (99%, CI 22.99-33.05).

predictable and unpredictable changes in perturbation amplitude. Since scaling of LL responses in the older control subjects differed slightly from the pattern described earlier for the young healthy subjects (Beckley et al. 1991c), we will first discuss the influence of normal aging on LL responses. Subsequently, we will discuss the abnormalities observed in the Parkinson patients.

Scaling of LL responses in older controls Using a paradigm identical to the present study, we have recently shown that young normal subjects are capable of scaling their LL responses if the perturbation amplitude is predictably varied (Becldey et al. 1991c). Furthermore, young normal subjects select a large default LL response during the unpredictable conditions. The current study indicates that older healthy subjects are also able to scale their LL responses under predictable and unpredictable conditions. However, unlike young controls, older subjects selected a default response which anticipated the smaller of the two perturbation amplitudes. This smaller default LL response could be less protective than a maximal one because subjects may fall if they select a small corrective reflex in response to an unexpectedly large perturbation. Selection of a small default response in healthy elderly subjects may suggest an aging effect on balance control and might be one of the factors which could contribute to the increased risk for falls in the elderly (Prudham and Evans 1981). This may suggest that the effect of normal aging on postural control is manifest by selection of an undersized de-

26

D.J. BECKLEYET AL.

fault response, while Parkinson's disease, as will further be discussed, results in the inability to appropriately scale the size of the LL response even under predictable conditions.

gators (Berardelli et al. 1986; Dick et al. 1986; Pullman et al. 1990; Nutt et al. 1992).

Scaling of LL responses in Parkinson patients

Abnormal scaling of LL reflexes under predictable conditions may be caused by impairment of the "putaminal loop" in the basal ganglia which is associated with motor functions. Movement amplitude specification is thought to be regulated by the supplementary motor area (SMA) which in turn is connected in a feedback circ.uit involving the pallidum and the thalamus (Riehle and Requin 1989; Alexander and Crutcher 1990; DeLong 1990). In PD there appears to be abnormal modulation of the SMA, which could manifest itself as defective scaling of movement amplitude. An alternative explanation for abnormal scaling of LL responses is increased muscle stiffness in PD (Watts et al. 1986; Dietz et al. 1988). Although we focused on E M G responses and did not measure joint angle movements or ankle torques, we doubt that increased muscle stiffness was a significant factor in our study because Dietz et al. (1988) did not observe increased stiffness in the tibialis anterior muscle of their Parkinson patients. Moreover, Allum et al. (1988) found no difference in initial ankle torque in the first 65 msec after rotational perturbation onset, when intrinsic viscoelastic properties are maximally influential, when comparing PD patients to controls. Finally, contractile properties of the muscles of Parkinson patients appear to be normal (Huffschmidt et al. 1991). Another possible reason as to why Parkinson patients cannot scale their postural reflexes is that their reflex amplitudes may saturate at a low level. Evidence in favor of this assumption is our observation that the Parkinson patients consistently underscaled their reflex amplitudes during all testing conditions. On the other hand, evidence somewhat against a saturation effect is that during individual trials Parkinson patients were at times capable of generating LL responses as large as those seen in controls. If present, such a saturation effect at small perturbation amplitudes would put Parkinson patients at serious functional disadvantage when exposed to large perturbations and could result in an increased risk of falling, even if the size of the perturbation is known in advance.

As opposed to young and old controls, Parkinson patients have a difficulty in scaling the amplitude of their LL postural reflexes when the perturbation amplitude is predictable. In addition, Parkinson patients are unable to modify their LL response through changes in cognitive set when the perturbation amplitude is uncertain. Furthermore, comparison of normalized LL amplitudes indicates that the inability to scale is caused by the fact that Parkinson patients consistently generate undersized LL responses during all test conditions. Scaling difficulties were not only evident in the complete patient group, but also on an individual basis since patients were 5 times less likely to scale their LL response than controls. It therefore seems that LL responses in Parkinson patients are impaired because of two reasons. In the first place, patients are unable to scale the corrective postural reflexes to the demands of the destabilizing stimulus. Secondly, patients consistently underscale the size of the LL response and, therefore, select protective reflexes which may be too small to prevent them from falling. It could be that these LL response abnormalities may contribute to the postural dysfunction commonly seen in patients with Parkinson's disease. It is of interest to note that our paradigm revealed abnormalities of the LL response which were not evident in studies where patients were exposed to a constant perturbation amplitude. Thus, normal or slightly increased LL amplitudes have been reported in standing Parkinson patients who merely received 3 - 4 ° toe-up perturbations (Diener et al. 1987; Scholz et al. 1987; Allum 1988; Busch et al. 1990; Schieppati and Nardone 1991) comparable to our 4P condition. Therefore, these studies did not address the issue of LL amplitude scaling. On the other hand, Nutt et al. (1992) also observed abnormal amplitudes of postural reflexes in Parkinson patients subjected to predictable variations in the amplitude of translations of a supporting platform. It therefore seems that scaling of postural reflexes is more sensitive in detecting reflex abnormalities in Parkinson's disease than the use of a single perturbation amplitude. It is possible that antiparkinsonian medication masked even more dramatic abnormalities in the Parkinson patients since our subjects were all tested while on medication. The fact that LL response scaling was impaired in patients who used therapeutic doses of levodopa suggests that amplitude specification may not be exclusively under dopaminergic control. This notion is consistent with the findings of several other investi-

Pathophysiology of abnormal scaling during predictable conditions

Pathophysiology of abnormal scaling during unpredictable conditions Defective modulation of LL responses during unpredictable conditions might be caused by abnormal frontal cortical influence on the basal ganglia through a "caudate loop" (Alexander and Crutcher 1990; DeLong 1990). Many recent studies have shown that patients with PD experience difficulties with complex motor task which involve shifts in cognitive set, possibly

IMPAIRED SCALING OF LL REFLEXES IN PD

due to dysfunction of this "caudate loop" (Buchwald et al. 1975; Cools et al. 1984; Flowers and Robertson 1985; Benecke et al. 1986, 1987a,b; Gotham et al. 1988; Prasher and Findley 1991). It could be argued that defective LL scaling under predictable conditions might restrict the ability of Parkinson patients to consistently select a default response during random conditions. In order to determine if Parkinson patients truly have problems with cognitive set, it would be instructive to study LL responses in patients with milder disease. It is in this respect illustrative to note that 10 psychiatric patients on chronic neuroleptic medication (seven with mild parkinsonism) could scale their LL responses during predictable conditions, but showed no ability to consistently select a default LL response during random conditions (Beckley et al. 1991b). This observation supports the assumption that lack of a default response in PD is due to a true cognitive defect. Furthermore, this might suggest that cognitive processing ("caudate loop function") is impaired earlier in the course of the disease than motor processing difficulties ("putaminal loop function"). Onset latencies Predictable and unpredictable variations of perturbation amplitude did not affect the onset latencies of LL responses. This is consistent with the results of earlier studies (Diener et al. 1984; Horak et al. 1989) and is in keeping with the notion that latencies are programmed from minimal sensory information and independently from response amplitudes (Diener et al. 1988). The finding of increased onset latencies of LL responses in the PD group was unexpected, since we had previously found onset latencies comparable to controls (Beckley et al. 1991a). However, other investigators have also shown moderately increased latencies for some postural responses (Scholz et al. 1987; Allum et al. 1988; Schieppati and Nardone 1991). The pathophysiology of delayed onset latencies is unknown, although it has been attributed to a peripheral neuropathy (Scholz et al. 1987). Our patients had no clinical evidence of peripheral neuropathy but, in the absence of nerve conduction studies, we cannot exclude the possibility of a subclinical neuropathy. We express our thanks to Lourdes Ilog for her assistance in preparation of this manuscript. We also express our appreciation to Dr. Neal Willits, Statistical Laboratory, University of California, Davis, for providing statistical assistance.

References Alexander, G.E. and C.rutcher, M.D. Functional architecture of basal ganglia circuits: neural substrates of parallel processing. Trends Neurosci., 1990, 13: 266-271.

27 Allum, J.H.L. Organization of stabilizing reflex responses in tibialis anterior muscles following ankle flexion perturbations in standing man. Brain Res., 1983, 264: 297-301. Allum, J.H.L. and Biidingen, H.J. Coupled stretch reflexes in ankle muscles: an evaluation of the contributions of active muscle mechanisms to human posture stability. In: R. Granit and O. Pompeiano (Eds.), Reflex Control of Posture and Movement. Progress in Brain Research, Vol. 50. Elsevier/North-Holland, Amsterdam, 1979: 185-195. Allum, J.H.J., Keshner, E.A., Honegger, F. and Wuetrich, R. Disturbance of posture in Parkinson's disease. In: B. Amblard, A. Berthoz and F. Clarac (Eds.), Posture and Gait: Development, Adaptation and Modulation. Elsevier, Amsterdam, 1988: 245-257. Beckley, D.J., Bloem, B.R., Remler, M.P., Tetrud, J.W. and Langston, J.W. Impaired scaling of long latency reflexes in the lower extremities of Parkinson's patients. Ann Neurol., 1990, 28 (Suppl. 2): 237. Beckley, D.J., Bloem, B.R., Van Dijk, J.G., Roos, R.A.C. and Remler, M.P. Electrophysiological correlates of postural instability in Parkinson's disease. Electroenceph. clin. Neurophysiol., 1991a, 81: 263-268. Beckley, D.J., Bloem, B.R., Singh, J., Remler, M.P. and Roos, R.A.C. Scaling of the long latency response in patients on chronic neuroleptic medication. Neurology, 1991b, 41 (Suppl. 1): 192. Beckley, D.J., Bloem, B.R., Remler, M.P., Roos, R.A.C. and Van Dijk, J.G. Long latency postural responses are functionally modified by cognitive set. Electroenceph. clin. Neurophysiol., 1991c, 81: 353-358. Benecke, R., Rothwell, J.C., Dick, J.P.R., Day, B.L. and Marsden, C.D. Performance of simultaneous movements in patients with Parkinson's disease. Brain, 1986, 109: 739-757. Benecke, R., Rothwell, J.C., Dick, J.P.R., Day, B.L. and Marsden, C.D. Disturbance of sequential movements in patients with Parkinson's disease. Brain, 1987a, 110: 361-379. Benecke, R., Rothwell, J.C., Dick, J.P.R., Day, B.L. and Marsden, C.D. Simple and complex movements off and on treatment in patients with Parkinson's disease. J. Neurol. Neurosurg. Psychiat., 1987b, 50: 296-303. Berardelli, A., Dick, J.P.R., Rothwell, J.C., Day, B.L. and Marsden, C.D. Scaling of the size of the first agonist EMG burst during rapid wrist movements in patients with Parkinson's disease. J. Neurol. Neurosurg. Psychiat., 1986, 49: 1273-1279. Buchwald, N.A., Hull, C.D., Levine, M.S. and Villablanca, J. The basal ganglia and the regulation of response and cognitive sets, In: M.A.B. Brazier (Ed.), Growth and Development of the Brain. Raven Press, New York, 1975: 171-189. Busch, C., Krafczyk, S., Straube, S., Carl, I., Gasser, T., Ortell, W. and Paulus, W. Long-loop reflexes induced by platform tilt: clinical relevance in Parkinson's disease. In: T. Brandt, W. Paulus, W. Bles, M. Dieterich, S. Krafczyk and A. Straube (Eds.), Disorders of Posture and Gait. Georg Thieme, Stuttgart, 1990: 415-418. Cody, F.W.J., MacDermott, N., Matthews, P.B.C. and Richardson, H.C. Observations on the genesis of the stretch reflex in Parkinson's disease. Brain, 1986, 109: 229-249. Cools, A.R., Van den Berken, J.H.L., Horstink, M.W.I., Van Spaendonk, K.P.M. and Berger, H.I.C. Cognitive and motor shifting amplitude disorder in Parkinson's disease, J. Neurol. Neurosurg. Psychiat., 1984, 47: 443-453. DeLong, M.R. Primate models of movement disorders of basal ganglia origin. Trends Neurosci., 1990, 13: 281-285. Dick, J.P.R., Rothwell, J.C., Berardelli, A., Thompson, P.D., Gioux, M., Benecke, R. et al. Associated postural adjustment in Parkinson's disease. J. Neurol. Neurosurg. Psychiat., 1986, 49: 13781385. Diener, H.C., Bootz, F., Dichgans, J. and Bruzek, W. Variability of postural "reflexes" in humans. Exp. Brain Res., 1983, 52: 423-428.

28 Diener, H.C., Dichgans, J., Bootz, F. and Bacher, M. Early stabilization of human posture after a sudden disturbance: influence of rate and amplitude of displacement. Exp. Brain Res., 1984, 56: 126-134. Diener, H.C., Scholz, E., Guschlbauer, B. and Dichgans, J. Increased shortening reaction in Parkinson's disease reflects a difficulty in modulating long loop reflexes. Movem. Dis., 1987, 2: 31-36. Diener, H.C., Horak, F.B. and Nashner, L.M. Influence of stimulus parameters on human postural reflexes. J. Neurophysiol., 1988, 59: 1888-1905. Dietz, V., Berger, W. and Horstmnn, G.A., Posture in Parkinson's disease: impairment of reflexes and programming. Ann. Neurol., 1988, 24: 660-669. Fahn, S. and Elton, R.L. Members of the UPDRS Development Committee: Unified Parkinson's disease rating scale. In: S. Fahn, C.D. Marsden, D.B. Calne and A. Lieberman (Eds.), Recent Developments in Parkinson's Disease, Vol. II. MacMillan Health Care Information, Florham Park, NJ, 1987: 153-163. Flowers, K.A. and Robertson, C. The effect of Parkinson's disease on the ability to maintain a mental set. J. Neurol. Neurosurg. Psychiat., 1985, 48: 517-529. Folstein, M., Folstein, S. and McHugh, P.R. Mini-mental state: a practical method for grading the cognitive state of patients for the clinician. J. Psychiat. Res., 1975, 12: 189-198. Gotham, A.M., Brown, R.G. and Marsden, C.D. 'Frontal' cognitive function in patients with Parkinson's disease 'on' and 'off' levodopa. Brain, 1988, 111: 299-321. Hallett, M. and Koshbin, S. A physiological mechanism of bradykinesia. Brain, 1980, 103: 301-314. Hansen, P.D., Woollacott, M.H. and Debu, B. Postural responses to changing task conditions. Exp. Brain Res., 1988, 73: 627-636. Hoehn, M.M. and Yahr, M.D. Parkinsonism: onset, progression and mortality. Neurology, 1967, 17: 427-442. Horak, F.B., Diener, H.C. and Nashner, L.M. Influence of central set on human postural responses. J. Neurophysiol., 1989, 62: 841-853. Huffschmidt, A., Stark, K. and Liicking, C.H. Contractile properties of lower leg muscles are normal in Parkinson's disease. J. Neurol. Neurosurg. Psychiat., 1991, 54: 457-460. Nardone, A., Giordano, A., Corr?~, T. and Schieppati, M. Responses

D.J. BECKLEY ET AL. of leg muscles in humans displaced while standing. Brain, 1990, 113: 65-84. Nashner, L.M. Adapting reflexes controlling the human posture. Exp. Brain Res., 1976, 26: 59-72. Nashner, L.M. Fixed patterns of postural responses among leg muscles during stance. Exp. Brain Res., 1977, 30: 13-24. Nutt, J., Horak, F. and Frank, J. Scaling of Postural responses in Parkinson's disease. In: F. Horak and M. Woollacott (Eds.), Posture and Gait: Control Mechanisms (Vol. II). University of Oregon Books, Portland, OR, 1992: 4-7. Prasher, D. and Findley, L. Dopaminergic induced changes in cognitive and motor processing in Parkinson's disease: an electrophysiological investigation. J. Neurol. Neurosurg. Psychiat., 1991, 54: 603 -609. Prudham, D. and Evans, J.G. Factors associated with falls in the elderly: a community study. Age Ageing, 1981, 10: 141-146. Pullman, S.L., Watts, R.L., Juncos, J.L and Sanes, J.N. Movement amplitude choice reaction-time performance in Parkinson's disease may he independent of dopaminergic status. J. Neurol. Neurosurg. Psychiat., 1990, 53: 279-283. Riehle, A. and Requin, J. Monkey primary motor and premotor cortex: single-cell activity related to prior information about direction and extent of an intended movement. J. Neurophysiol., 1989, 61: 534-351. Rothwell, J.C., Day, B.L., Berardelli, A. and Marsden, C.D. Habituation and conditioning of the human long latency stretch reflex. Exp. Brain Res., 1986, 63: 197-204. Schieppati, M. and Nardone, A. Free and supported stance in Parkinson's disease. Brain, 1991, 114: 1227-1244. Scholz, E., Diener, H.C., Noth, J., Friedemann, H., Dichgans, J. and Bacher, M. Long latency EMG responses in leg muscles. II. Parkinson's disease. J. Neurol. Neurosurg. Psychiat., 1987, 50: 66-70. Traub, M.M., Rothwell, J.C. and Marsden, C.D. Anticipatory postural reflexes in Parkinson's disease and other akinetic-rigid syndromes and in cerebellar ataxia. Brain, 1980, 103: 393-412. Watts, R.L., Wiegner, A.W. and Young, R.R. Elastic properties of muscles measured at the elbow in man. II. Patients with parkinsonian rigidity. J. Neurol. Neurosurg. Psychiat., 1986, 49: 11771181.