Movement Disorders Vol. 7, No. 1, 1992, pp. 14-22 0 1992 Movement Disorder Society
The Coordination of Posture and Voluntary Movement in Patients with Cerebellar Dysfunction H.-C. Diener, J . Dichgans, B. Guschlbauer, M. Bacher, H. Rapp, and T. Klockgether Department of Neurology, University of Tuebingen, KEiniken Schnarrenberg, Tuebingen, F.R.G.
Summary: Postural adjustments associated with the task of rising on tiptoes were investigated in a reaction time paradigm in 10 normal subjects and 18 patients with cerebellar disorders. Cerebellar dysfunction was due to either degenerative cerebellar disease, tumor, or ischemia. Displacements of the center of foot pressure (CFP) were recorded. The task, accomplished by the triceps surae muscle (executional activity, mean latency of 41 1 ms),is mechanically effective only if the center of gravity has been shifted forward in advance. To this effect, a phasic burst of preparatory EMG activity in the tibialis anterior normally occurs at a mean latency of 163 ms, shifting the center of gravity forward. Shortly thereafter, activity of the quadriceps fernoris (175 rns) extends the knee and aids the forward shift of the center of gravity. Different aspects of this motor sequence were disturbed in individual patients: Latencies of preparatory and executional activity were uncorrelated in 15 of the 18 patients. Executional ( n = 16) or preparatory (n = 13) EMG activity was tonic instead of phasic. Latencies of either preparatory or executional EMG activities or both were prolonged (n = 10). The time interval between motor preparation and execution was increased ( n = 9). The trial-to-trial variability of biomechanical parameters and EMG latency was increased. Preparatory EMG activity in the quadriceps was entirely missing ( n = 9), resulting in knee bending at the unsuccessful attempt to rise on tiptoes. Patients who were most severely affected had no preparatory activity at all (n = 2 ) , and therefore were unable to perform the task. In conclusion, the cerebellum contributes to the scaling of size and duration of preparatory and executional motor activity and controls their temporal relationships. Key Words: Human stance-Postural adjustments-Motor preparation-Cerebellar disorders.
as a unit and processed in parallel pathways or whether they reflect a hierarchically organized motor pattern (3). In the latter case, one could assume that a movement signal generated in motor cortex can only be released if subcortical structures provide the additional postural activity necessary to stabilize the body above its support. The hierarchically organized pattern allows proprioceptive feedback to participate. Feedback could indicate whether preparation was biomechanically effective or not. Consequently, the study presented here was performed to obtain more insight into the structure of these postural synergies and the possible centers
Movements of the trunk or the arms in standing human subjects are preceded by preparatory muscular activity in postural trunk and leg muscles (1,2). This preparatory activity seems to be an integral part of a particular motor program in that it prevents destabilization of the center of gravity due to limb or trunk movements. An open question is whether executional and preparatory motor activity are centrally programmed
~
~~
~
Address correspondence and reprint requests to Dr. H.-C. Diener at Neurologische Universitatsklinik, Hufelandstr. 55, D4300 Essen, Federal Republic of Germany.
14
15
COORDINATION OF POSTURE AND MOVEMENT
timing and force that determines performance and equilibrium.
and pathways involved. We chose to investigate patients with cerebellar lesions or atrophy, because they exhibit overt difficulties of trunk stabilization when asked to perform leg or arm movements (4-6). We investigated the possibility of a specific pattern of alterations of motor preparation and its coordination with voluntary movement in patients with cerebellar diseases. For this purpose, the temporal sequence, i.e., timing, duration, and the size of electromyographic (EMG) responses in muscles involved in the task of motor preparation and execution, had to be studied in normals and patients. EMG activity in postural muscles not involved in the main motor task was called preparatory if it began prior to the onset of voluntary activity. By executional motor activation, we mean voluntary activity directly related to the execution of the task. In our case, the task of suddenly rising on tiptoes (executional movement) in normal subjects was performed by means of activating the triceps surae. It was preceded by preparatory activity of the tibialis anterior, quadriceps, and biceps femoris (among other muscles). If the subject stands erect, activity of the triceps surae (a biarticular muscle) alone would lead to a backward shift of the center of gravity and to flexion of the knee. The subject would fall backwards. Evidently, the task can properly be achieved only if for compensation the center of gravity is shifted forward (e.g., by the anterior tibialis) and if the knee is stabilized (e.g., by appropriate activity of the quadriceps and biceps femoris). It is the precision of coordination in terms of
METHODS Normal Subjects and Patients Ten normal subjects matched in age and sex to the population of patients were investigated. They were normal upon neurological examination. Among the 18 patients with cerebellar lesions investigated (Table l), 12 suffered from a degenerative cerebellar atrophy without involvement of the basal ganglia, central motor pathways, or cerebral cortex according to neurological, neurophysiological, and computed tomography (CT) criteria. Four patients possibly suffered from olivopontocerebellar atrophy in its very early stages. Although they presented with cerebellar symptoms only, invs701vement of structures outside the cerebellum was assumed on the basis of their pathological blink reflexes or brainstem acoustic evoked responses. Furthermore, nuclear magnetic resonance or CT imaging revealed pontine and cerebellar atrophy. One patient was investigated more than 2 years after the successful operative removal of a unilateral cerebellar tumor and another patient after ischemic lesion of one cerebellar hemisphere. All patients had clinical signs of a lesion of the cerebellar vermis (eye movement disturbances, ataxia of gait). Muscle tone was normal in all patients. Patients with multiple sclerosis, inflammatory disorders, or toxic degeneration (alcohol, drugs) were excluded. All
TABLE 1. Clinical features and diagnosis of the patients with cerebellar disorders Name
Sex
BA BJ BR CP GA
M M
HE HR KK MA NB NW PA RT
RJ RA
ss
TC WP
F M M M M F F M M M M F F M F M
Age (years)
Diagnosis
48 53 65 19 63 55 35 54 48 53 52 53 20 17 60 49 64 44
Ischemia OPCA CA Tumor CA CA OPCA CA OPCA OPCA CA CA CA CA CA CA CA CA
OPCA = olivopontocerebellar atrophy, CA
=
Oculomotor disturbances 1
1 2 3 4 4 2 3 3 5 2 3 4 4 1
4 3 4
Dysmetria 1 3 1 3 3 2 3 4 4 3 2 2 2 0 3 4 3 2
Intentional tremor
Postural ataxia
0 2 1 3
0 3 4 1 5 2 4 4 3 5 4 3 3 3 2 4 4 4
1
0 3 3 1 3 0 0 0 0 2 2 0 0
cerebellar atrophy.
Movement Disorders, Vol. 7, N o . I , 1992
16
H.-C. DIENER ET AL.
patients had ataxia of stance and gait, but were able to stand freely without support. Informed consent was obtained from all normal subjects and patients. Methods and Evaluation The subjects stood upright and relaxed on a force-measuring platform. This platform measured the displacements of the center of foot pressure (CFP) in the anterior-posterior direction. The subject’s weight was equally distributed on the two feet. The feet were 5 cm apart. Subjects were instructed to rise as fast as possible on tiptoes after an acoustic trigger signal and to stay in this position for at least 5 s. Five training trials were given to each subject prior to the eight recorded trials. The initial position of the CFP was measured online prior to the trigger signal, and the patient instructed verbally to shift the CFP either forward or backward to reproduce the same initial position in all trials. [The initial posture has a considerable influence on the timing of preparation and execution (7).] EMG surface electrodes 5 cm apart were attached bilaterally over the rectus femoris (QUA), anterior tibialis (TA), and medial gastrocnemius (TS) muscles. The electrodes were connected to small preamplifiers that were attached to the skin adjacent to the electrodes. In order to relate EMG activity and movement, we recorded trunk position and changes in hip, knee, and ankle joint angle in five normal subjects and five cerebellar patients. Small light-reflecting dots were attached to the tip and base of the small toe and the lateral aspects of the ankle, knee, and hip joints. The position of these markers in space was traced by a video camera under infrared stroboscopic illumination. Position was extracted from the digitized video frames. Relative angles between earth vertical and thigh, lower leg, and foot were than calculated (ELITE system). In these normals and patients, we recorded EMG signals from the biceps femoris in addition.
The latencies of individual EMGs from single trials were evaluated on an interactive computer terminal. Mean values and SDs of latencies were calculated for single muscles within a particular experimental condition for each individual subject. Intermuscle latency differences (e.g., the time interval between tibialis anterior and triceps surae onset) were derived from these means. The overall variability of latencies of EMG responses contained the variability of the reaction times within and across subjects. To obtain estimates of the intraand interindividual variability, the SDs of the individual means ( = interindividual SD) and the mean SD of individual subjects ( = intraindividual SD) were also determined. In addition, durations of EMG activity from each single trial were measured. The beginning and end of a particular EMG response were visually identified and the integral calculated between the two. For graphical representation of the data, we extracted the variability induced by different reaction times of the TS in single trials by aligning the onset of the TS. The comparison of data from patients with normal subjects was performed in terms of absolute latency, interindividual variability ( = SD of mean individual latencies), intraindividual variability in a single subject across the trials, frequency of occurrence of a particular EMG response, latency, intermuscular latency differences, and the shape of the EMG (phasic or tonic). The distribution of these parameters in patients was skewed, which made a comparison of mean values difficult. We therefore chose to use the upper range of normal values (e.g., the highest value encountered in one of the normal subjects) as the upper limit of normality. Regular measurements of EMG response duration and integral were only possible in anterior tibia1 muscle. The quadriceps femoris and gastrocnemius sometimes showed tonic activity, which made it impossible to define the end of these responses.
Data Evaluation EMG signals were amplified, full wave rectified, high-pass filtered (I5 Hz), and stored on a PDP 111 44 computer together with the output of the platform, measuring the anterior-posterior displacement of the CFP. Data from the ELITE system were handled by a PDP 11/73. The sampling period was 1,00&2,000 ms. The sampling rate was 1 kHz for EMGs, 200 Hz for the recording of the CFP, and 50 Hz for the ELITE system.
Statistical Analysis The comparison of latencies of different muscles within subjects was done by a parametric analysis of variance for related samples. The comparison of latencies and EMG durations across patients and normals was performed by a Friedman nonparametric analysis of variance for unrelated samples and multiple comparisons, Comparisons between the two sides were performed by Wilcoxon tests, and comparisons for a particular muscle between nor-
Movement Disorders, Vol. 7, N o . I , 1992
17
COORDINATION OF POSTURE AND MOVEMENT
mals and patients with t tests. Correlations were calculated with a linear regression model. The level for significant differences was set at p < 0.05.
RESULTS I
Normals The temporal relation between shifts in leg position and EMG recordings can be seen in Fig. 1. After the tone signal (start of recording), the subject phasically activates the anterior tibia1 muscle. This activity results in a forward shift of the lower leg and the center of gravity. With both feet in close contact with the platform, tibialis activity pulls the body forward, thereby exerting a backward momentum on the platform that is reflected by the downward deflection of the CFP in the later figures. After a mean interval of 10 ms, activity starts in the quadriceps femoris and in another 20 ms in the biceps femoris. Activity of these two muscles results in a partial stabilization of the knee and hip joint. A “perfect” stabilization of the knee joint would result in almost identical changes in lower leg and thigh angle. Absent or reduced stabilization results in traces that are displaced in opposite directions (see Fig. I vs. Fig. 2). Stabilization of the knee joint is a prerequisite for shifting the center of gravity forward. The elevation on tiptoes is performed by activity of the triceps surae muscle. This phasic activity (rise), which starts on average 250 ms after the tibialis anterior (Table 2), is followed by tonic activity (hold). The push by which the heels lift off the ground leads to an upward deflection of the recording of the CFP as shown in the following figures. The quadriceps activity is partly inhibited at the moment when the triceps is activated. This sequence is appropriate to the task. During the inhibition phase of the quadriceps, the triceps surae leads to a slight flexion in the knee joint. The new joint angle is then maintained by the second increase in quadriceps activity. The initial inhibition of the triceps surae resting activity after the tone signal, as described by Lipshits et al. (l), could not be observed in our experiments. Normal subjects had no significant side differences in the latencies of the EMG responses (Wilcoxon test). Latencies of right and left tibialis anterior or triceps surae were significantly correlated (TA, r = 0.76, df = 72, p S 0.001; TS, r = 0.88, df = 77, p S 0.001). The latencies of the tibialis and triceps surae were correlated in six nor-
II
Ill
IV
THIGH
7
LOWER LEO
UVl
L
uv,
7
1
I.
lA
Ts
I
QUA
1 4
BICEPS
t
.
1s
0 1 FIG. 1. Changes of inclination of thigh, lower leg, and heel against vertical during lifting on tiptoes in a normal subject. Three consecutive runs are superimposed in the middle part of the figure. Upward deflections indicate a forward shift against vertical. In the upper part of the figure, lower leg angle is depicted as a black triangle. 1 = start of forward shift of center of gravity; I1 = lifting of heels; 111 = maximal knee flexion; and IV = subject reaches the final position with extended knees. Rectified EMGs from the preparatory tibialis anterior (TA), quadriceps femoris (QUA), biceps femoris (BICEPS), and executional triceps surae (TS) are shown in the lower part of the figure. All traces are shifted in time to align the onset of TS activity with 0 on the time scale.
Movement Disorders. Vol. 7, No. I , 1992
18
H.-C. DIENER ET AL. CEREBELLUM THIGH
1
TABLE 2. Mean latencies (AM, ms), and interindividual standard deviations (SD, ms), measured for preparatory (tibialis anterior, quadriceps femoris) and focal (triceps surae) EMG activity and the mean latency differences (ms) between normals and patients
i Muscle
LOWER LEG
HEEL
TA
TS
QUA
BICEPS
r
I
.s
1 FIG. 2. Rectified EMG recordings and movement parameters in a patient with cerebellar atrophy. Quadriceps activity is not preparatory, but follows activity in the triceps surae. The result is flexion of the knee, at the moment when the patient tries to rise on tiptoes. The patient drops back on his heels. Note the high amount of cocontraction of almost all muscles. 0
mals (p < 0.05, r > 0.70) and showed no correlation in the remaining four, indicating that the motor sequence is probably preprogrammed in two-thirds of normal subjects. However, we can not exclude the possibility that the onset of the TS results from a sensory feedback signal. Patients
Two of our patients were unable to perform the task. Figure 2 gives an example from one of them. There was a phasic preparatory EMG activity in the TA resulting in a forward shift of the body (trace “lower leg”). Quadriceps activity builds up slowly and is considerably delayed. It begins clearly after the onset of triceps activity. Inadequate fixation of the knee joint and insufficient forward shift of the
Movement Disorders, Vol. 7, N o . 1 , 1992
Tibialis anterior AM SD Quadriceps femoris AM SD Triceps surae AM SD
Normals
Cerebellum
Difference
163.5 26.9
190.3 72.8
26.8
174.6 29.4
295.0* 186.0
120.4
411.5 63.5
503.6* 104.0
92.1
* Significant difference compared to normals.
center of gravity are the result. Although lifting on tiptoes is possible through activity of the triceps, the same activity results in flexion of the knee (trace “thigh”). Consequently, the patient’s center of gravity does not shift forward to a sufficient degree and he is not able to maintain the position, i.e., he falls back on his heels. This example shows that adequate timing of preparatory activity is mandatory for the proper execution of the task. The data of these two patients were not considered for statistical analysis. The following two examples may serve to demonstrate the variety of pathological deviations from the physiological pattern seen in our patients. Patient A (Fig. 3) showed the normal sequence of preparation (TA, QUA) and execution (TS). The beginning of motor preparation was delayed beyond the longest latency seen in normals (210 ms for TA and 225 ms for QUA). The TS latency was just within the upper range of normality. The latencies between TA and QUA (upper limit of 65 ms) and TA and TS (upper limit of 348 ms) were normal in this case. The duration of the preparatory activity clearly exceeded that of the normal subject, resulting in cocontraction of the TA and TS. The overshoot of CFP displacement illustrates the resulting dysmetria. Patient F (Fig. 4) combines almost all of the pathological features described. Preparation in the QUA is absent but is preserved, although delayed, in the TA. The TA exhibits a much slower buildup of force and considerably prolonged activity compared to the normal subject (duration exceeds 360 ms). The time interval between preparation and ex-
19
COORDINATION OF POSTURE AND MOVEMENT CEREBELLUM
FIG. 3. Rectified and averaged EMGs (eight runs) in a patient with cerebellar disorder (left) and an age-matched normal subject (right). The upper limit of EMG onset latency in normal subjects is shown by a big vertical arrow. The vertical lines indicate the average onset and offset latency as calculated from single trials. The longest duration of quadriceps (QUA) and tibialis (TA) seen in a normal subject is shown as a hatched bar. The horizontal arrow shows the longest time interval between TA and TS in normals. The lowest trace depicts the displacement of the center of foot pressure (CFP). The forward shift of the body is reflected by a downward deflection of the recording of the CFP, and the lift to the toes by an upward deflection. Note the delay in preparatory activity in the tibialis anterior and quadriceps and the tonic nature of this part of the motor pattern.
500
250
NORMAL
A
TA
TA
T I
TS
750
rns
7
0
250
500
750
ms
ters in a particular patient correlated with the severity of cerebellar disease. Compared to normals, the latencies of all EMG responses were delayed in patients with cerebellar disorders (Table 2). The delay, however, was not uniform but was much longer for the second (QUA) and third (TS) part of the motor sequence and shortest for the TA (Table 3). The coordination of latencies between the two legs was preserved in patients with cerebellar disease as revealed by the significant correlation coefficients (Table 4). The intraindividual variation of latencies of a given muscle was increased in more than 50% of our patients. This increase was more pronounced in the preparatory than in the executional muscles (Table 5). One may conclude that only a smaller part of the temporal delay can be ascribed to the well-known prolongation of the reaction time in cerebellar patients. In
Analysis of Group Data The quantitative analysis across all 18 cerebellar patients revealed that 16 had a tonic response of TS, i.e., no clearly discernable transient increase in EMG activity during the rise phase. Thirteen patients showed a tonic preparatory EMG activity in the TA that overlapped the beginning of the TS. Ten patients each had delayed onset latencies of either execution or preparation or both. Ten patients had an absent preparatory response in the QUA. Nine patients showed an increased delay between TA and TS and nine an increased variability of TA and/ or TS latencies across trials. Two patients showed a repetition pattern and two were unable to perform the task. The number of pathological parameF
NORMAL
*
0
ecution is too long, and the variability in TS onset is increased.
CEREBELLUM
A
F OUA
& TA
TA 4%]
TS
FIG. 4. Example from a patient with absent preparation in QUA, delayed EMG onset in TA and TS, increased intermuscle latency, prolonged TA activity, and increased variability of TS onset latency.
H
CFP
I
4 I
0
250
500
750
0
250
500
ms
750
Movement Disorders, Vol. 7, No. 1, 1992
20
H.-C. DIENER ET AL.
TABLE 3. Mean time differences ( A M ) and standard deviation (SD) in ms between preparatory (tibialis anterior) and executional (triceps surae) motor activity as well as time differences between the two components of Preparatory activity (tibialis anterior, quadriceps femoris) calculated f r o m single trials Time difference, tibialis anterior/ triceps surae
Time difference, tibialis anterior/ quadriceps femoris
Condition
AM
SD
AM
SD
Normals Cerebellum
24718 31 1.6
66.7 98.8
10.9 117.7*
33.5 186.2
TABLE 4. Linear correlation of EMG latencies between the two legs and TA and TS on the same side
TA - TA TS - TS TA - TS"
0.7645'' 0.8807" + correlation, n correlation, n
Cerebellum 0.7695" = =
12; 8
+ a
0.7300" + correlation, n - correlation, n
= significant positive correlation (p < 0.01); Calculated for individual subjects and legs.
Movement Disorders, Vol. 7. N o . 1 , 1992
=
= =
Tibialis anterior
Normals Cerebellum
36.2 (15.868) 43.9* (6.1-161)
Quadriceps femoris 31.4 (12.4-53)
46.0* (16.0-103)
Triceps surae 53.6 (32-98) 67.3 (19-152)
~
addition, the temporal coupling becomes less reproducible, not only between preparation and execution, but also and even more so within preparation, i.e., TA and QUA. The basic structure, with tibialis first followed by quadriceps and triceps surae, however, was preserved in 16 of 18 patients. This suggests that the basic structure of the motor sequence is programmed outside the cerebellum, or maybe the cerebellar disorders were not severe enough to produce timing errors large enough to change the order. The latencies of preparational and executional EMG activity were correlated in 12 of 20 measurements in normals, but only in 4 of 36 in patients (Table 4). The temporal structure of the motor pattern, therefore, seems at least partly to be preprogrammed in normals, which is not the case in patients. The duration of tibialis activity was significantly increased by 100 ms or more compared to normals (393.1 k 196 vs. 289.1 ? 93.3 ms). Integrals of tibialis activity, however, were normal. The single patient with a unilateral lesion of the right cerebellar hemisphere had normal tibialis latencies on both sides. Early quadriceps was missing on the affected side. The onset of EMG in the TS was bilaterally delayed.
Normals
Condition
* Significant difference compared to controls.
* Significant difference compared to normals.
Correlation
TABLE 5. Mean intraindividual standard deviations (ms) across trials f o r preparatory (tibialis anterior, quadriceps femoris) and executional (triceps surae) activity and range of standard deviations (in parentheses) of individual subjects and patients
4; 32
no correlation.
DISCUSSION Upright stance in humans is maintained by muscle tone and corrective muscular action based on sensory input from proprioceptive, visual, and vestibular receptors and also by feedforward control. The latter is part of any voluntary motor action, since the task of keeping the projection of the center of gravity inside the area of stability delineated by the surface of the feet must have priority. Postural adjustments, therefore, precede and accompany rapid voluntary body or arm movements (8-18). An example of preparatory postural activity is the early bilateral activity of the extensors of the foot and knee prior to ankle flexors when the task is to raise as quickly as possible to the toes (1). This study in patients with cerebellar disease showed that the sequence of preparatory and executional activity is basically preserved in terms of order, but is severely disordered in terms of its precise timing and the fast buildup of EMG activity and therefore the proper scaling of force of each of the participating muscles. Moderately affected patients often still exhibit preparation by activity of the extensors of the foot (TA) but show no or delayed EMG activity in muscles that stabilize the knee. Severely disabled patients may show no preparation at all and consequently are unable to perform the task. These observations indicate that the cerebellum not only participates in the execution of the voluntary movement, but also in the coordination of execution and the concomitant postural adjustments that have been shown to form an integral and necessary part of the volitional act. Furthermore, the preparatory response does not come as a package with a preprogrammed temporal sequence of properly scaled EMG responses, but is within itself coordinated by the cerebellum. The temporal coordination between preparation and execution is strictly fixed in two-thirds of normals. This temporal relationship is lost in cerebellar disease. This
COORDINATION OF POSTURE AND MOVEMENT
indicates a role for internal (corollary discharge) or external feedback about preparatory efficacy prior to the release of the final execution. The bilateral temporal coordination of the pattern between the two legs, however, seems not to be controlled by the cerebellum. The most critical parameters controlled by the cerebellum in this task are the absolute timing of onset latencies, durations of EMG responses, timing of the proper interval between consecutive EMG responses, and the recruitment of force. If the time interval between preparation and execution is too short (as observed in one of our patients), the shift of the body cannot occur prior to TS activity and remains partially ineffective. More frequently, the time interval between preparation and execution is too long, again causing preparation to be ineffective, or even leading to a forward fall of the patient (hypermetria). This observation makes an important role of the feedback transmitted to the cerebellum unlikely. Another important feature of cerebellar dysfunction is the altered shape of EMG responses. At least in young normal subjects who perform faster movements, these responses are phasic with a rapid but transient buildup of force performing the “rise” followed by tonic activity for the “hold.” EMG responses in cerebellar patients exhibit a much slower rise in activity, frequently no discernable phasic component, and are increased in duration. The pathological features observed seem to be specific for cerebellar disease. Ten patients with Parkinson’s disease had normal latencies of all EMG responses, normal interreponse intervals, and normal EMG durations and integrals (7). Patients with cerebellar atrophy exhibit delayed motor reaction times in terms of the latency of the triceps surae. A similar delay of the reaction time in cerebellar patients was observed earlier for arm and hand movements (5,6). We additionally confirmed the work of Pal’tsev and El’ner (19), who described delayed anticipatory activity in some patients with cerebellar damage. Their observation of its absence in others was repeated for quadriceps activity and rarely also for the TA with our paradigm. The buildup of EMG activity in the TA and QUA is slower and less effective in patients, which makes it useful to delay the execution of the task until the body is shifted far enough forward. Alternatively, the delays could be due to learned motor set through everyday experience changing the motor program according to the experiences made before with inadequate motor preparation.
21
Patients with delayed preparatory activity show a disproportional increase in intermuscular latency of executional with respect to preparatory activity. This could be an argument against a fixed central program. A lesion within central structures generating fixed central programs could manifest itself in a delay of preparatory movement, but then the time interval to focal muscular activity should be equally delayed. Indeed, the delays observed are best explained by changes in biomechanical necessities as explained earlier. In summary, our results indicate that the cerebellum contributes to the scaling of size and duration of preparatory and executional motor activity and controls their temporal relationships. Acknowledgment: The research was supported by the Deutsche Forschungsgemeinschaft (SFB 307 A3). The authors thank S. Gandevia, J. Hore, and D. Tweed for critical comments. We acknowledge the technical help of K. Wessel and E. Scholz.
REFERENCES I . Lipshits MI, Mauritz K, Popov KE. Quantitative analysis of anticipatory postural components of a complex voluntary movement. Fiziol Chelov 1982;7:411-419. 2. Massion J. Postural changes accompanying voluntary movement. Normal and pathological aspects. Hum Neurobiol 1984;2:261-267. 3. Brown JE, Frank JS. Influence of event anticipation of postural actions accompanying voluntary movement. Exp Brain Res 1987;67:645-650. 4. Dichgans J, Diener HC. Clinical evidence for functional compartmentalization of the cerebellum. In: Bloedel JR, Dichgans J, Precht W, eds. Cerebellar functions. Heidelberg: Springer, 1985:126-147. 5 . Hallett M, Shahani BT, Young RR. EMG analysis of patients with cerebellar deficits. J Neurol Neurosurg Psychiatry 1975;38:1163-1 169. 6. Holmes G. The symptoms of acute cerebellar injuries due to gunshot injuries. Brain 1917;40:461-535. 7. Diener HC, Dichgans J, Guschlbauer B, Bacher M, Rapp H , Langenbach P. Associated postural adjustments with body movements in normal subjects and patients with parkinsonism and cerebellar disease. Rev Neurol (Paris) 1990;146: 555-563. 8. Bazalgette D, Zattara M, Bathien N , Bouisset S , Rondot P. Postural adjustments associated with rapid voluntary arm movements in patients with Parkinson’s disease. In: Yahr MD, Bergmann KJ, eds. Advances in Neurology, Vol. 45. New York: Raven Press, 1986:371-374. 9. Belen’kii VY, Gurfinkel VS, Pal’tsev YI.Elements of control of voluntary movements. Biofizika 1967;12:134-141. 10. Bouisset S, Zattara M. A sequence of postural movements precedes voluntary movement. Neurosci Lett 1981;22:263270. 1 1 . Bouisset S, Zattara M. Anticipatory postural movements related to a voluntary movement. In: Centre National D’ktudes Spatiales, ed. Space physiology. Toulouse: Cepadues, 1983:137-141. 12. Cordo PJ, Nashner LM. Properties of postural adjustments
Movement Disorders, Vol. 7, N o . 1. 1992
22
H.-C. DIENER ET AL.
associated with rapid arm movement. J Neurophysiol 1982; 47~187-302. 13. Dick JPR, Rothwell JC, Berardelli A, et al. Associated postural adjustments in Parkinson’s disease. J Neurol Neurosurg Psychiatry 1986;49:1378-1385. 14. Friedli WG, Hallett M, Simon SR. Postural adjustments associated with rapid voluntary arm movements. I: Electromyographic data. J Neurol Neurosurg Psychiatry 1984;47: 61 1422. 15. Horak FB, Esselman PE, Anderson ME, Lynch MK. The effects of movement velocity, mass displaced and task certainty on associated postural adjustments made by normal and hemiplegic individuals. J Neurol Neurosurg Psychiatry 1984;47:1020-1 028.
Movement Disorders, Vol. 7 , No. 1, 1992
16. Lee WA. Anticipatory control of postural and task muscles during rapid arm flexion. J Motor Behav 1980;12:185-196. 17. Lee WA, Buchanan TS, Rogers MW. Effects of arm acceleration and behavioural conditions on the organization of postural adjustments during arm flexion. Exp Brain Res 1987;66:257-270. 18. Zattara M, Bouisset S. I h d e chronometrique du programme posturo-cinktique liC au mouvement volontaire. J Physiol (Paris) 1986;81:14-16. 19. Pal’tsev YI, El’ner AM. Preparatory and compensatory period during voluntary movement in patients with involvement of the brain of different localization. Biofizika 1967;12: 142-147.
The Coordination of Posture and Voluntary Movement in Patients with Cerebellar Dysfunction H.-C. Diener, J . Dichgans, B. Guschlbauer, M. Bacher, H. Rapp, and T. Klockgether Department of Neurology, University of Tuebingen, KEiniken Schnarrenberg, Tuebingen, F.R.G.
Summary: Postural adjustments associated with the task of rising on tiptoes were investigated in a reaction time paradigm in 10 normal subjects and 18 patients with cerebellar disorders. Cerebellar dysfunction was due to either degenerative cerebellar disease, tumor, or ischemia. Displacements of the center of foot pressure (CFP) were recorded. The task, accomplished by the triceps surae muscle (executional activity, mean latency of 41 1 ms),is mechanically effective only if the center of gravity has been shifted forward in advance. To this effect, a phasic burst of preparatory EMG activity in the tibialis anterior normally occurs at a mean latency of 163 ms, shifting the center of gravity forward. Shortly thereafter, activity of the quadriceps fernoris (175 rns) extends the knee and aids the forward shift of the center of gravity. Different aspects of this motor sequence were disturbed in individual patients: Latencies of preparatory and executional activity were uncorrelated in 15 of the 18 patients. Executional ( n = 16) or preparatory (n = 13) EMG activity was tonic instead of phasic. Latencies of either preparatory or executional EMG activities or both were prolonged (n = 10). The time interval between motor preparation and execution was increased ( n = 9). The trial-to-trial variability of biomechanical parameters and EMG latency was increased. Preparatory EMG activity in the quadriceps was entirely missing ( n = 9), resulting in knee bending at the unsuccessful attempt to rise on tiptoes. Patients who were most severely affected had no preparatory activity at all (n = 2 ) , and therefore were unable to perform the task. In conclusion, the cerebellum contributes to the scaling of size and duration of preparatory and executional motor activity and controls their temporal relationships. Key Words: Human stance-Postural adjustments-Motor preparation-Cerebellar disorders.
as a unit and processed in parallel pathways or whether they reflect a hierarchically organized motor pattern (3). In the latter case, one could assume that a movement signal generated in motor cortex can only be released if subcortical structures provide the additional postural activity necessary to stabilize the body above its support. The hierarchically organized pattern allows proprioceptive feedback to participate. Feedback could indicate whether preparation was biomechanically effective or not. Consequently, the study presented here was performed to obtain more insight into the structure of these postural synergies and the possible centers
Movements of the trunk or the arms in standing human subjects are preceded by preparatory muscular activity in postural trunk and leg muscles (1,2). This preparatory activity seems to be an integral part of a particular motor program in that it prevents destabilization of the center of gravity due to limb or trunk movements. An open question is whether executional and preparatory motor activity are centrally programmed
~
~~
~
Address correspondence and reprint requests to Dr. H.-C. Diener at Neurologische Universitatsklinik, Hufelandstr. 55, D4300 Essen, Federal Republic of Germany.
14
15
COORDINATION OF POSTURE AND MOVEMENT
timing and force that determines performance and equilibrium.
and pathways involved. We chose to investigate patients with cerebellar lesions or atrophy, because they exhibit overt difficulties of trunk stabilization when asked to perform leg or arm movements (4-6). We investigated the possibility of a specific pattern of alterations of motor preparation and its coordination with voluntary movement in patients with cerebellar diseases. For this purpose, the temporal sequence, i.e., timing, duration, and the size of electromyographic (EMG) responses in muscles involved in the task of motor preparation and execution, had to be studied in normals and patients. EMG activity in postural muscles not involved in the main motor task was called preparatory if it began prior to the onset of voluntary activity. By executional motor activation, we mean voluntary activity directly related to the execution of the task. In our case, the task of suddenly rising on tiptoes (executional movement) in normal subjects was performed by means of activating the triceps surae. It was preceded by preparatory activity of the tibialis anterior, quadriceps, and biceps femoris (among other muscles). If the subject stands erect, activity of the triceps surae (a biarticular muscle) alone would lead to a backward shift of the center of gravity and to flexion of the knee. The subject would fall backwards. Evidently, the task can properly be achieved only if for compensation the center of gravity is shifted forward (e.g., by the anterior tibialis) and if the knee is stabilized (e.g., by appropriate activity of the quadriceps and biceps femoris). It is the precision of coordination in terms of
METHODS Normal Subjects and Patients Ten normal subjects matched in age and sex to the population of patients were investigated. They were normal upon neurological examination. Among the 18 patients with cerebellar lesions investigated (Table l), 12 suffered from a degenerative cerebellar atrophy without involvement of the basal ganglia, central motor pathways, or cerebral cortex according to neurological, neurophysiological, and computed tomography (CT) criteria. Four patients possibly suffered from olivopontocerebellar atrophy in its very early stages. Although they presented with cerebellar symptoms only, invs701vement of structures outside the cerebellum was assumed on the basis of their pathological blink reflexes or brainstem acoustic evoked responses. Furthermore, nuclear magnetic resonance or CT imaging revealed pontine and cerebellar atrophy. One patient was investigated more than 2 years after the successful operative removal of a unilateral cerebellar tumor and another patient after ischemic lesion of one cerebellar hemisphere. All patients had clinical signs of a lesion of the cerebellar vermis (eye movement disturbances, ataxia of gait). Muscle tone was normal in all patients. Patients with multiple sclerosis, inflammatory disorders, or toxic degeneration (alcohol, drugs) were excluded. All
TABLE 1. Clinical features and diagnosis of the patients with cerebellar disorders Name
Sex
BA BJ BR CP GA
M M
HE HR KK MA NB NW PA RT
RJ RA
ss
TC WP
F M M M M F F M M M M F F M F M
Age (years)
Diagnosis
48 53 65 19 63 55 35 54 48 53 52 53 20 17 60 49 64 44
Ischemia OPCA CA Tumor CA CA OPCA CA OPCA OPCA CA CA CA CA CA CA CA CA
OPCA = olivopontocerebellar atrophy, CA
=
Oculomotor disturbances 1
1 2 3 4 4 2 3 3 5 2 3 4 4 1
4 3 4
Dysmetria 1 3 1 3 3 2 3 4 4 3 2 2 2 0 3 4 3 2
Intentional tremor
Postural ataxia
0 2 1 3
0 3 4 1 5 2 4 4 3 5 4 3 3 3 2 4 4 4
1
0 3 3 1 3 0 0 0 0 2 2 0 0
cerebellar atrophy.
Movement Disorders, Vol. 7, N o . I , 1992
16
H.-C. DIENER ET AL.
patients had ataxia of stance and gait, but were able to stand freely without support. Informed consent was obtained from all normal subjects and patients. Methods and Evaluation The subjects stood upright and relaxed on a force-measuring platform. This platform measured the displacements of the center of foot pressure (CFP) in the anterior-posterior direction. The subject’s weight was equally distributed on the two feet. The feet were 5 cm apart. Subjects were instructed to rise as fast as possible on tiptoes after an acoustic trigger signal and to stay in this position for at least 5 s. Five training trials were given to each subject prior to the eight recorded trials. The initial position of the CFP was measured online prior to the trigger signal, and the patient instructed verbally to shift the CFP either forward or backward to reproduce the same initial position in all trials. [The initial posture has a considerable influence on the timing of preparation and execution (7).] EMG surface electrodes 5 cm apart were attached bilaterally over the rectus femoris (QUA), anterior tibialis (TA), and medial gastrocnemius (TS) muscles. The electrodes were connected to small preamplifiers that were attached to the skin adjacent to the electrodes. In order to relate EMG activity and movement, we recorded trunk position and changes in hip, knee, and ankle joint angle in five normal subjects and five cerebellar patients. Small light-reflecting dots were attached to the tip and base of the small toe and the lateral aspects of the ankle, knee, and hip joints. The position of these markers in space was traced by a video camera under infrared stroboscopic illumination. Position was extracted from the digitized video frames. Relative angles between earth vertical and thigh, lower leg, and foot were than calculated (ELITE system). In these normals and patients, we recorded EMG signals from the biceps femoris in addition.
The latencies of individual EMGs from single trials were evaluated on an interactive computer terminal. Mean values and SDs of latencies were calculated for single muscles within a particular experimental condition for each individual subject. Intermuscle latency differences (e.g., the time interval between tibialis anterior and triceps surae onset) were derived from these means. The overall variability of latencies of EMG responses contained the variability of the reaction times within and across subjects. To obtain estimates of the intraand interindividual variability, the SDs of the individual means ( = interindividual SD) and the mean SD of individual subjects ( = intraindividual SD) were also determined. In addition, durations of EMG activity from each single trial were measured. The beginning and end of a particular EMG response were visually identified and the integral calculated between the two. For graphical representation of the data, we extracted the variability induced by different reaction times of the TS in single trials by aligning the onset of the TS. The comparison of data from patients with normal subjects was performed in terms of absolute latency, interindividual variability ( = SD of mean individual latencies), intraindividual variability in a single subject across the trials, frequency of occurrence of a particular EMG response, latency, intermuscular latency differences, and the shape of the EMG (phasic or tonic). The distribution of these parameters in patients was skewed, which made a comparison of mean values difficult. We therefore chose to use the upper range of normal values (e.g., the highest value encountered in one of the normal subjects) as the upper limit of normality. Regular measurements of EMG response duration and integral were only possible in anterior tibia1 muscle. The quadriceps femoris and gastrocnemius sometimes showed tonic activity, which made it impossible to define the end of these responses.
Data Evaluation EMG signals were amplified, full wave rectified, high-pass filtered (I5 Hz), and stored on a PDP 111 44 computer together with the output of the platform, measuring the anterior-posterior displacement of the CFP. Data from the ELITE system were handled by a PDP 11/73. The sampling period was 1,00&2,000 ms. The sampling rate was 1 kHz for EMGs, 200 Hz for the recording of the CFP, and 50 Hz for the ELITE system.
Statistical Analysis The comparison of latencies of different muscles within subjects was done by a parametric analysis of variance for related samples. The comparison of latencies and EMG durations across patients and normals was performed by a Friedman nonparametric analysis of variance for unrelated samples and multiple comparisons, Comparisons between the two sides were performed by Wilcoxon tests, and comparisons for a particular muscle between nor-
Movement Disorders, Vol. 7, N o . I , 1992
17
COORDINATION OF POSTURE AND MOVEMENT
mals and patients with t tests. Correlations were calculated with a linear regression model. The level for significant differences was set at p < 0.05.
RESULTS I
Normals The temporal relation between shifts in leg position and EMG recordings can be seen in Fig. 1. After the tone signal (start of recording), the subject phasically activates the anterior tibia1 muscle. This activity results in a forward shift of the lower leg and the center of gravity. With both feet in close contact with the platform, tibialis activity pulls the body forward, thereby exerting a backward momentum on the platform that is reflected by the downward deflection of the CFP in the later figures. After a mean interval of 10 ms, activity starts in the quadriceps femoris and in another 20 ms in the biceps femoris. Activity of these two muscles results in a partial stabilization of the knee and hip joint. A “perfect” stabilization of the knee joint would result in almost identical changes in lower leg and thigh angle. Absent or reduced stabilization results in traces that are displaced in opposite directions (see Fig. I vs. Fig. 2). Stabilization of the knee joint is a prerequisite for shifting the center of gravity forward. The elevation on tiptoes is performed by activity of the triceps surae muscle. This phasic activity (rise), which starts on average 250 ms after the tibialis anterior (Table 2), is followed by tonic activity (hold). The push by which the heels lift off the ground leads to an upward deflection of the recording of the CFP as shown in the following figures. The quadriceps activity is partly inhibited at the moment when the triceps is activated. This sequence is appropriate to the task. During the inhibition phase of the quadriceps, the triceps surae leads to a slight flexion in the knee joint. The new joint angle is then maintained by the second increase in quadriceps activity. The initial inhibition of the triceps surae resting activity after the tone signal, as described by Lipshits et al. (l), could not be observed in our experiments. Normal subjects had no significant side differences in the latencies of the EMG responses (Wilcoxon test). Latencies of right and left tibialis anterior or triceps surae were significantly correlated (TA, r = 0.76, df = 72, p S 0.001; TS, r = 0.88, df = 77, p S 0.001). The latencies of the tibialis and triceps surae were correlated in six nor-
II
Ill
IV
THIGH
7
LOWER LEO
UVl
L
uv,
7
1
I.
lA
Ts
I
QUA
1 4
BICEPS
t
.
1s
0 1 FIG. 1. Changes of inclination of thigh, lower leg, and heel against vertical during lifting on tiptoes in a normal subject. Three consecutive runs are superimposed in the middle part of the figure. Upward deflections indicate a forward shift against vertical. In the upper part of the figure, lower leg angle is depicted as a black triangle. 1 = start of forward shift of center of gravity; I1 = lifting of heels; 111 = maximal knee flexion; and IV = subject reaches the final position with extended knees. Rectified EMGs from the preparatory tibialis anterior (TA), quadriceps femoris (QUA), biceps femoris (BICEPS), and executional triceps surae (TS) are shown in the lower part of the figure. All traces are shifted in time to align the onset of TS activity with 0 on the time scale.
Movement Disorders. Vol. 7, No. I , 1992
18
H.-C. DIENER ET AL. CEREBELLUM THIGH
1
TABLE 2. Mean latencies (AM, ms), and interindividual standard deviations (SD, ms), measured for preparatory (tibialis anterior, quadriceps femoris) and focal (triceps surae) EMG activity and the mean latency differences (ms) between normals and patients
i Muscle
LOWER LEG
HEEL
TA
TS
QUA
BICEPS
r
I
.s
1 FIG. 2. Rectified EMG recordings and movement parameters in a patient with cerebellar atrophy. Quadriceps activity is not preparatory, but follows activity in the triceps surae. The result is flexion of the knee, at the moment when the patient tries to rise on tiptoes. The patient drops back on his heels. Note the high amount of cocontraction of almost all muscles. 0
mals (p < 0.05, r > 0.70) and showed no correlation in the remaining four, indicating that the motor sequence is probably preprogrammed in two-thirds of normal subjects. However, we can not exclude the possibility that the onset of the TS results from a sensory feedback signal. Patients
Two of our patients were unable to perform the task. Figure 2 gives an example from one of them. There was a phasic preparatory EMG activity in the TA resulting in a forward shift of the body (trace “lower leg”). Quadriceps activity builds up slowly and is considerably delayed. It begins clearly after the onset of triceps activity. Inadequate fixation of the knee joint and insufficient forward shift of the
Movement Disorders, Vol. 7, N o . 1 , 1992
Tibialis anterior AM SD Quadriceps femoris AM SD Triceps surae AM SD
Normals
Cerebellum
Difference
163.5 26.9
190.3 72.8
26.8
174.6 29.4
295.0* 186.0
120.4
411.5 63.5
503.6* 104.0
92.1
* Significant difference compared to normals.
center of gravity are the result. Although lifting on tiptoes is possible through activity of the triceps, the same activity results in flexion of the knee (trace “thigh”). Consequently, the patient’s center of gravity does not shift forward to a sufficient degree and he is not able to maintain the position, i.e., he falls back on his heels. This example shows that adequate timing of preparatory activity is mandatory for the proper execution of the task. The data of these two patients were not considered for statistical analysis. The following two examples may serve to demonstrate the variety of pathological deviations from the physiological pattern seen in our patients. Patient A (Fig. 3) showed the normal sequence of preparation (TA, QUA) and execution (TS). The beginning of motor preparation was delayed beyond the longest latency seen in normals (210 ms for TA and 225 ms for QUA). The TS latency was just within the upper range of normality. The latencies between TA and QUA (upper limit of 65 ms) and TA and TS (upper limit of 348 ms) were normal in this case. The duration of the preparatory activity clearly exceeded that of the normal subject, resulting in cocontraction of the TA and TS. The overshoot of CFP displacement illustrates the resulting dysmetria. Patient F (Fig. 4) combines almost all of the pathological features described. Preparation in the QUA is absent but is preserved, although delayed, in the TA. The TA exhibits a much slower buildup of force and considerably prolonged activity compared to the normal subject (duration exceeds 360 ms). The time interval between preparation and ex-
19
COORDINATION OF POSTURE AND MOVEMENT CEREBELLUM
FIG. 3. Rectified and averaged EMGs (eight runs) in a patient with cerebellar disorder (left) and an age-matched normal subject (right). The upper limit of EMG onset latency in normal subjects is shown by a big vertical arrow. The vertical lines indicate the average onset and offset latency as calculated from single trials. The longest duration of quadriceps (QUA) and tibialis (TA) seen in a normal subject is shown as a hatched bar. The horizontal arrow shows the longest time interval between TA and TS in normals. The lowest trace depicts the displacement of the center of foot pressure (CFP). The forward shift of the body is reflected by a downward deflection of the recording of the CFP, and the lift to the toes by an upward deflection. Note the delay in preparatory activity in the tibialis anterior and quadriceps and the tonic nature of this part of the motor pattern.
500
250
NORMAL
A
TA
TA
T I
TS
750
rns
7
0
250
500
750
ms
ters in a particular patient correlated with the severity of cerebellar disease. Compared to normals, the latencies of all EMG responses were delayed in patients with cerebellar disorders (Table 2). The delay, however, was not uniform but was much longer for the second (QUA) and third (TS) part of the motor sequence and shortest for the TA (Table 3). The coordination of latencies between the two legs was preserved in patients with cerebellar disease as revealed by the significant correlation coefficients (Table 4). The intraindividual variation of latencies of a given muscle was increased in more than 50% of our patients. This increase was more pronounced in the preparatory than in the executional muscles (Table 5). One may conclude that only a smaller part of the temporal delay can be ascribed to the well-known prolongation of the reaction time in cerebellar patients. In
Analysis of Group Data The quantitative analysis across all 18 cerebellar patients revealed that 16 had a tonic response of TS, i.e., no clearly discernable transient increase in EMG activity during the rise phase. Thirteen patients showed a tonic preparatory EMG activity in the TA that overlapped the beginning of the TS. Ten patients each had delayed onset latencies of either execution or preparation or both. Ten patients had an absent preparatory response in the QUA. Nine patients showed an increased delay between TA and TS and nine an increased variability of TA and/ or TS latencies across trials. Two patients showed a repetition pattern and two were unable to perform the task. The number of pathological parameF
NORMAL
*
0
ecution is too long, and the variability in TS onset is increased.
CEREBELLUM
A
F OUA
& TA
TA 4%]
TS
FIG. 4. Example from a patient with absent preparation in QUA, delayed EMG onset in TA and TS, increased intermuscle latency, prolonged TA activity, and increased variability of TS onset latency.
H
CFP
I
4 I
0
250
500
750
0
250
500
ms
750
Movement Disorders, Vol. 7, No. 1, 1992
20
H.-C. DIENER ET AL.
TABLE 3. Mean time differences ( A M ) and standard deviation (SD) in ms between preparatory (tibialis anterior) and executional (triceps surae) motor activity as well as time differences between the two components of Preparatory activity (tibialis anterior, quadriceps femoris) calculated f r o m single trials Time difference, tibialis anterior/ triceps surae
Time difference, tibialis anterior/ quadriceps femoris
Condition
AM
SD
AM
SD
Normals Cerebellum
24718 31 1.6
66.7 98.8
10.9 117.7*
33.5 186.2
TABLE 4. Linear correlation of EMG latencies between the two legs and TA and TS on the same side
TA - TA TS - TS TA - TS"
0.7645'' 0.8807" + correlation, n correlation, n
Cerebellum 0.7695" = =
12; 8
+ a
0.7300" + correlation, n - correlation, n
= significant positive correlation (p < 0.01); Calculated for individual subjects and legs.
Movement Disorders, Vol. 7. N o . 1 , 1992
=
= =
Tibialis anterior
Normals Cerebellum
36.2 (15.868) 43.9* (6.1-161)
Quadriceps femoris 31.4 (12.4-53)
46.0* (16.0-103)
Triceps surae 53.6 (32-98) 67.3 (19-152)
~
addition, the temporal coupling becomes less reproducible, not only between preparation and execution, but also and even more so within preparation, i.e., TA and QUA. The basic structure, with tibialis first followed by quadriceps and triceps surae, however, was preserved in 16 of 18 patients. This suggests that the basic structure of the motor sequence is programmed outside the cerebellum, or maybe the cerebellar disorders were not severe enough to produce timing errors large enough to change the order. The latencies of preparational and executional EMG activity were correlated in 12 of 20 measurements in normals, but only in 4 of 36 in patients (Table 4). The temporal structure of the motor pattern, therefore, seems at least partly to be preprogrammed in normals, which is not the case in patients. The duration of tibialis activity was significantly increased by 100 ms or more compared to normals (393.1 k 196 vs. 289.1 ? 93.3 ms). Integrals of tibialis activity, however, were normal. The single patient with a unilateral lesion of the right cerebellar hemisphere had normal tibialis latencies on both sides. Early quadriceps was missing on the affected side. The onset of EMG in the TS was bilaterally delayed.
Normals
Condition
* Significant difference compared to controls.
* Significant difference compared to normals.
Correlation
TABLE 5. Mean intraindividual standard deviations (ms) across trials f o r preparatory (tibialis anterior, quadriceps femoris) and executional (triceps surae) activity and range of standard deviations (in parentheses) of individual subjects and patients
4; 32
no correlation.
DISCUSSION Upright stance in humans is maintained by muscle tone and corrective muscular action based on sensory input from proprioceptive, visual, and vestibular receptors and also by feedforward control. The latter is part of any voluntary motor action, since the task of keeping the projection of the center of gravity inside the area of stability delineated by the surface of the feet must have priority. Postural adjustments, therefore, precede and accompany rapid voluntary body or arm movements (8-18). An example of preparatory postural activity is the early bilateral activity of the extensors of the foot and knee prior to ankle flexors when the task is to raise as quickly as possible to the toes (1). This study in patients with cerebellar disease showed that the sequence of preparatory and executional activity is basically preserved in terms of order, but is severely disordered in terms of its precise timing and the fast buildup of EMG activity and therefore the proper scaling of force of each of the participating muscles. Moderately affected patients often still exhibit preparation by activity of the extensors of the foot (TA) but show no or delayed EMG activity in muscles that stabilize the knee. Severely disabled patients may show no preparation at all and consequently are unable to perform the task. These observations indicate that the cerebellum not only participates in the execution of the voluntary movement, but also in the coordination of execution and the concomitant postural adjustments that have been shown to form an integral and necessary part of the volitional act. Furthermore, the preparatory response does not come as a package with a preprogrammed temporal sequence of properly scaled EMG responses, but is within itself coordinated by the cerebellum. The temporal coordination between preparation and execution is strictly fixed in two-thirds of normals. This temporal relationship is lost in cerebellar disease. This
COORDINATION OF POSTURE AND MOVEMENT
indicates a role for internal (corollary discharge) or external feedback about preparatory efficacy prior to the release of the final execution. The bilateral temporal coordination of the pattern between the two legs, however, seems not to be controlled by the cerebellum. The most critical parameters controlled by the cerebellum in this task are the absolute timing of onset latencies, durations of EMG responses, timing of the proper interval between consecutive EMG responses, and the recruitment of force. If the time interval between preparation and execution is too short (as observed in one of our patients), the shift of the body cannot occur prior to TS activity and remains partially ineffective. More frequently, the time interval between preparation and execution is too long, again causing preparation to be ineffective, or even leading to a forward fall of the patient (hypermetria). This observation makes an important role of the feedback transmitted to the cerebellum unlikely. Another important feature of cerebellar dysfunction is the altered shape of EMG responses. At least in young normal subjects who perform faster movements, these responses are phasic with a rapid but transient buildup of force performing the “rise” followed by tonic activity for the “hold.” EMG responses in cerebellar patients exhibit a much slower rise in activity, frequently no discernable phasic component, and are increased in duration. The pathological features observed seem to be specific for cerebellar disease. Ten patients with Parkinson’s disease had normal latencies of all EMG responses, normal interreponse intervals, and normal EMG durations and integrals (7). Patients with cerebellar atrophy exhibit delayed motor reaction times in terms of the latency of the triceps surae. A similar delay of the reaction time in cerebellar patients was observed earlier for arm and hand movements (5,6). We additionally confirmed the work of Pal’tsev and El’ner (19), who described delayed anticipatory activity in some patients with cerebellar damage. Their observation of its absence in others was repeated for quadriceps activity and rarely also for the TA with our paradigm. The buildup of EMG activity in the TA and QUA is slower and less effective in patients, which makes it useful to delay the execution of the task until the body is shifted far enough forward. Alternatively, the delays could be due to learned motor set through everyday experience changing the motor program according to the experiences made before with inadequate motor preparation.
21
Patients with delayed preparatory activity show a disproportional increase in intermuscular latency of executional with respect to preparatory activity. This could be an argument against a fixed central program. A lesion within central structures generating fixed central programs could manifest itself in a delay of preparatory movement, but then the time interval to focal muscular activity should be equally delayed. Indeed, the delays observed are best explained by changes in biomechanical necessities as explained earlier. In summary, our results indicate that the cerebellum contributes to the scaling of size and duration of preparatory and executional motor activity and controls their temporal relationships. Acknowledgment: The research was supported by the Deutsche Forschungsgemeinschaft (SFB 307 A3). The authors thank S. Gandevia, J. Hore, and D. Tweed for critical comments. We acknowledge the technical help of K. Wessel and E. Scholz.
REFERENCES I . Lipshits MI, Mauritz K, Popov KE. Quantitative analysis of anticipatory postural components of a complex voluntary movement. Fiziol Chelov 1982;7:411-419. 2. Massion J. Postural changes accompanying voluntary movement. Normal and pathological aspects. Hum Neurobiol 1984;2:261-267. 3. Brown JE, Frank JS. Influence of event anticipation of postural actions accompanying voluntary movement. Exp Brain Res 1987;67:645-650. 4. Dichgans J, Diener HC. Clinical evidence for functional compartmentalization of the cerebellum. In: Bloedel JR, Dichgans J, Precht W, eds. Cerebellar functions. Heidelberg: Springer, 1985:126-147. 5 . Hallett M, Shahani BT, Young RR. EMG analysis of patients with cerebellar deficits. J Neurol Neurosurg Psychiatry 1975;38:1163-1 169. 6. Holmes G. The symptoms of acute cerebellar injuries due to gunshot injuries. Brain 1917;40:461-535. 7. Diener HC, Dichgans J, Guschlbauer B, Bacher M, Rapp H , Langenbach P. Associated postural adjustments with body movements in normal subjects and patients with parkinsonism and cerebellar disease. Rev Neurol (Paris) 1990;146: 555-563. 8. Bazalgette D, Zattara M, Bathien N , Bouisset S , Rondot P. Postural adjustments associated with rapid voluntary arm movements in patients with Parkinson’s disease. In: Yahr MD, Bergmann KJ, eds. Advances in Neurology, Vol. 45. New York: Raven Press, 1986:371-374. 9. Belen’kii VY, Gurfinkel VS, Pal’tsev YI.Elements of control of voluntary movements. Biofizika 1967;12:134-141. 10. Bouisset S, Zattara M. A sequence of postural movements precedes voluntary movement. Neurosci Lett 1981;22:263270. 1 1 . Bouisset S, Zattara M. Anticipatory postural movements related to a voluntary movement. In: Centre National D’ktudes Spatiales, ed. Space physiology. Toulouse: Cepadues, 1983:137-141. 12. Cordo PJ, Nashner LM. Properties of postural adjustments
Movement Disorders, Vol. 7, N o . 1. 1992
22
H.-C. DIENER ET AL.
associated with rapid arm movement. J Neurophysiol 1982; 47~187-302. 13. Dick JPR, Rothwell JC, Berardelli A, et al. Associated postural adjustments in Parkinson’s disease. J Neurol Neurosurg Psychiatry 1986;49:1378-1385. 14. Friedli WG, Hallett M, Simon SR. Postural adjustments associated with rapid voluntary arm movements. I: Electromyographic data. J Neurol Neurosurg Psychiatry 1984;47: 61 1422. 15. Horak FB, Esselman PE, Anderson ME, Lynch MK. The effects of movement velocity, mass displaced and task certainty on associated postural adjustments made by normal and hemiplegic individuals. J Neurol Neurosurg Psychiatry 1984;47:1020-1 028.
Movement Disorders, Vol. 7 , No. 1, 1992
16. Lee WA. Anticipatory control of postural and task muscles during rapid arm flexion. J Motor Behav 1980;12:185-196. 17. Lee WA, Buchanan TS, Rogers MW. Effects of arm acceleration and behavioural conditions on the organization of postural adjustments during arm flexion. Exp Brain Res 1987;66:257-270. 18. Zattara M, Bouisset S. I h d e chronometrique du programme posturo-cinktique liC au mouvement volontaire. J Physiol (Paris) 1986;81:14-16. 19. Pal’tsev YI, El’ner AM. Preparatory and compensatory period during voluntary movement in patients with involvement of the brain of different localization. Biofizika 1967;12: 142-147.
