Changes in short and long latency stretch responses during the transition from posture to movement

Brain Research, 229 (1981) 337-351

337

Elsevier/North-Holland Biomedical Press

C H A N G E S IN S H O R T A N D L O N G L A T E N C Y STRETCH RESPONSES D U R I N G T H E T R A N S I T I O N F R O M P O S T U R E TO M O V E M E N T

JAMES A. MORTIMER, DAVID D. WEBSTER and THOMAS G. DUKICH Geriatric Research, Education and Clinical Center, and Neurology Service, Veterans Administration Medical Center, Minneapolis, M N 55417 and Department of Neurology, University of Minnesota Medical School, Minneapolis, M N 55455 ( U.S.A.).

(Accepted June 4th, 1981) Key words: stretch reflex - - posture - - movement -- electromyography -- human - - motor set

SUMMARY Experiments were performed in 18 normal subjects to estimate the time course of changes in the gains of pathways mediating short- and long-latency responses to muscle stretch during the transition from a maintained posture against a steady load to a rapid ballistic movement. Subjects were instructed to rapidly flex or extend their forearm in response to a tone from an initial position of 90 ° of elbow flexion. Torque pulses stretching the biceps muscle were applied to the forearm at 8 different times before and after the signal to initiate the movement, and the gains of short- and longlatency pathways were estimated from averages of rectified biceps E M G activity for 20 trials at each time interval between the onsets of the tone and torque pulse. The findings demonstrate that changes in the magnitude of long-latency responses (M2, M3) occur during the period between the onset of the auditory signal and the voluntary motor response. However, the magnitude of the short-latency response (M1) remains unchanged until after the onset of voluntary motor activity. The differences in the timing of short- and long-latency stretch responses suggests that activity in long-latency pathways may play an important preparatory role in facilitating the transition from posture to movement.

INTRODUCTION Despite considerable investigation, the role played by the stretch reflex in the regulation of movement and posture remains controversial. Numerous studies have examined motor responses to changes in load during attempted maintenance of posture 1,4,11, 14,15,16,23,28,29,33,40,41,43,48 and during the course of voluntary movements6,S,9,12,13,1a, 16,za,34,36,38,40.In addition, E M G responses to applied perturbation s have been studied 0006-8993/81/0000-0000/$02.75 © Elsevier/North-Holland Biomedical Press

338 in relation to the intended motor response to imposed load changes 4,11,15,17-19,z6,28, 30,3z,38,41,4z,46-4s. Until very recently, however, little attention has been given to the changes that may occur in the gains of stretch refex pathways during the transition from posture to movement25, zT. Studies in man of the electromyographic response to step load changes show that the stretch reflex is composed of two main components: an early component (M 1), occurring at a latency similar to that of the tendon jerk, and a longer latency component (M2) 19,37. Because of its brief latency, the M 1 component is generally assumed to involve only segmental pathways. Evidence for the origin of the M2 component favors a combined influence of segmental and suprasegmental pathwaysZO,3L A third peak in the motor response to muscle stretch (M3) has been described in both alert and tranquillized monkeys 47 and in some human subjects35,4a,4L The size of this component depends primarily upon the motor set of the subject and is influenced little by the initial load or displacement amplitudO 1. Several investigators 1~,j9,24 have suggested that this late component is a preprogrammed response, since its latency, magnitude and even polarity can depend on the volitional set of the subject. In the experiments reported here, we investigated changes in stretch reflex gain during the period of transition from posture to movement. A reaction time paradigm was employed, in which normal human subjects performed rapid forearm movements in response to an auditory signal and the apparent gains of short- and long-latency pathways involved in the stretch lesponse were estimated from averaged motor responses to torque perturbations elicited before and after the signal to initiate the movement. Our findings indicate that changes in the magnitude of long-latency responses (M2, M3) occur during the period between the onset of the auditory signal and the voluntary motor response. However, the magnitude of the short latency reflex (M1) remains unchanged until after the onset of voluntary motor activity. The differences in the timing of changes of short- and long-latency stretch responses suggests that activity in long-latency pathways may play an impoTtant preparatory role in facilitating the transition from posture to movement. MATERIALS AND METHODS Eighteen right-handed subjects with normal motor function, ranging in age from 20 to 57 years, were studied. During the experiments, subjects were seated in a chair and had their right forearm strapped into a rigid, lightweight arm support that could be rotated about the elbow in a horizontal plane by two brushless DC torque motors (Aeroflex TQ82W) coupled to a single shaft (Fig. 1). Subject's hands were strapped to a flat vertical handle located on the arm support with the forearm supinated. E M G was recorded bipolarly from DISA electrodes pasted over the belly of the biceps and the lateral head of the triceps. An interelectrode separation of 3 cm was used. Following differential amplification and high-pass filtering (passband 30 Hz-10 kHz), the E M G signals were full-wave rectified. Amplified signals for torque, angular position and biceps and triceps rectified E M G were sampled at intervals of 1.6 ms and averaged on a MED-80 computer. To permit precise determination of the times of change of E M G signals, no additional filtering was performed on individual or averaged traces.

339

Torque¢ Position Position

Pillow

Full Wove Rectified Biceps $ Triceps EMG

Position ,.~--,--~ nsducers

Med-~80 L Computer I

Control Electronics

]

Audio Stimulus Toique Pulse

I

Fig. 1. Schematic diagram of the experimental setup. See text for description.

During the experiments, subjects were requested to maintain a 90 ° elbow angle against a steady clockwise 2 N m preload on biceps, and instructed to respond to a 2.9 kHz tone by flexing their forearm as rapidly as possible. Tones were 500 ms in duration and delivered through a S O N A L E R T located immediately behind the subject's head at an intensity of 80 db SPL (as measured at a position corresponding to the center of the head). This intensity was selected to provide an adequate stimulus without evoking startle responses. For each trial, the desired initial position of 90 ° elbow angle was indicated to the subject by the vertical position of a needle on a meter. Trials were initiated when the subject held his arm within a 1° window and demonstrated electrical silence of the triceps muscle for a period of 5 s. Arm movements were arrested by a sponge rubber pillow. Clockwise torque pulses of 2 N m amplitude and 500 ms duration (loading the biceps) were presented at 8 different times relative to the onset of the tone signal (100 and 25 ms prior, at the same time, and 25, 50, 75,100 and 125 ms following tone onset). These times were selected to sample the interval between tone onset and the onset of the voluntary E M G response to the tone, which averaged 137.6 ms in the 18 subjects studied. For each of the 8 different intervals between tone onset and onset of the torque pulse 20 trials were given, and the tone was presented without torque perturbation on 20 trials. Trials were given in blocks of 30, with a 3-5 min test period between blocks. The 9 types of trials were presented in a random order and sorted on-line by the MED-80 to yield ensemble averages of the 20 trials for each condition.

340 Data collected on each trial were displayed on a video scope, permitting on-line artifact rejection. For those trials in which only the tone was presented, the onset of voluntary biceps E M G activity was located with a cursor and the reaction time stored in memory for later analysis. Following completion of an experiment, the following output was obtained: (l)a histogram of the E M G r e a c t i o n times on the 20 trials in which the tone was presented alone; (2) ensemble averages of torque, position, biceps and triceps E M G on the same 20 trials; and (3) ensemble averages of torque, position, biceps and triceps E M G for each time interval between tone onset and torque pulse onset. Averaged E M G data collected from one subject are shown in Figure 2. Trace I is

I

~

AVERAGE BICEPS EMG RESPONSETO TONE ALONE

RT~--~,] I

"~-- T ~ J ~

TOROUE

II

AVERAGEBICEPS EMG RESPONSETO TONE + TORQUEPULSE

II-I COMPUTED AVERAGEBICEPS EMG RESPONSETO TORQUEPULSE

i

w .... . "t'=T + D i ~

PULSE DELAY T - -i00,-25,0,25,50,75,100,125 MS DI- 37.5 mS (25-50 InS) D2= 62.5 mS (50-75 mS) D3- 87.5 mS (75-100 mS) Fig. 2. Computation of EMG responses to torque pulses. Average biceps EMG responses are shown for one particular pulse delay (75 ms) between tone onset (vertical axis) and torque onset. Subject D4.

341 the average E M G response to the tone presented alone, trace II, the average response to the tone plus a torque pulse presented 75 ms following tone onset. The average biceps response to the torque pulse was computed by subtracting trace II from trace I. While no claim can be made for superposition, subtraction of these traces yields a discrete response (bottomtrace, Fig. 2), lasting from approximately 25 msto 100ms following torque pulse onset. The absence of appreciable E M G activity in the subtracted trace after 100 ms suggests that the torque pulse elicited only an early 'reflex-like' response (shown in black) without subsequent voluntary activity. Although this was a common finding when torque pulses followed tone onset by 75 ms or more, a prolonged discharge was frequently seen when the change in load occurred before or slightly after the beginning of the tone, suggesting that subjects initiate voluntary responses to torque pulses when they occur well before the onset of the voluntary motor response to the tone. In order to quantify the computed E M G responses during the first 100 ms following the onset of the torque pulse, motor response indices were calculated for each set of 20 trials. The methods for their computation have been described in previous reports42,4L Briefly, the average rectified E M G voltage displacement relative to a 50 ms pre-torque pulse baseline was obtained for three different time intervals following torque pulse onset (25-50 ms, 50-75 ms, and 75-100 ms). These average displacements were then divided by the mean rectified E M G voltage level when subjects maintained a 90 ° elbow angle against a constant 2 Nm load on the biceps for an equivalent period of time. The three time intervals selected correspond approximately to the M1, M2 and M3 intervals described by Mortimer and Johnson 41 and Lee and Tatton 35. In most cases, separate peaks for M2 and M3 could not be distinguished. However, because of differences in the dependence of M2 and M3 on the motor set of the subject 41, an independent assessment of the response magnitude in these two time intervals was obtained. The method for evaluating changes in the motor response indices with time is illustrated in Fig. 2. The 3 indices provide a sample of motoneuronal excitability at an average delay from tone onset, ~, equal to the sum of the time difference between torque pulse onset and the midpoint of the interval over which the motor response is calculated (D1, D2, D3), and T, the interval between tone and torque onsets. Thus, for the 25-50 ms interval, the 'gain' of the pathways governing the magnitude of the response during this interval is effectively being sampled at an average time delay of 112.5 ms (37.5 ÷ 75 ms), whereas the 'gain' of pathways governing the magnitude of the response from 50 to 75 ms is being sampled at a longer delay from tone onset of 137.5 ms (62.5 ÷ 75 ms). It should be noted that the inferred times of change of the 'gains' must exceed the times at which the gains of central pathways are actually adjusted by at least the conduction time between the motoneurons and the muscle where the change in magnitude of specific motor responses is detected. For responses involving long loop pathways, the change in gain may precede its detection at the muscle by several tens of milliseconds if the site of gain change is located at a relay prio~ to the final common pathway.

342

-100 ms

25 ms

~>

d iv-

75ms

{l-

er-

125 ms

0

5O

~ I--

100

TIME AFTER ONSET OF TORQUE PULSE (ms)

Fig. 3. Ensemble averages of biceps EMG responses to torque pulses presented 100 ms before tone onset and 25, 75 and 125 ms followingtone onset. Traces were computed by subtracting the response to the tone presented alone from the responses to a combination of the tone and the torque pulse. Subject D20. RESULTS

Changes in the stretch responses of the agonist muscle in preparation for a rapid movement Ensemble averages of biceps E M G responses to torque pulses presented 100 ms before tone onset, and 25, 75 and 125 ms following tone onset are shown for one subject in Fig. 3. The 4 traces were produced by subtracting the response to the tone presented alone from the responses to a combination of the tone and the torque pulse. In each trace, the E M G activity begins to deviate from the background level about 25 ms after torque pulse onset. Three separate increases in E M G activity can be seen: an early increase in E M G activity lasting from about 25-50 ms after pulse onset; a larger increase from approximately 50-100 ms; and a third increase beginning at around 100 ms and continuing beyond the end of the trace. The second increase in activity is considerably larger than the first, and this was the case for all subjects tested. By comparing the responses to the torque pulse delivered at different times relative to the tone signal, it can be seen that both short- and long-latency responses initially increase in size as the transition from maintenance of posture against a constant load to performance of a rapid flexion movement is accomplished. However, substantial differences in the timing of these changes are apparent. No change is seen in the short-latency (25-50 ms) component until the torque pulse lags the tone onset by 125 ms. By contrast, an increase in the biceps E M G response between 75 and 100 ms is observed when the torque pulse lags tone onset by as little as 25 ms.

343 In order to quantify the changes in EMG stretch responses during the period of preparation for a movement, motor response indices for biceps activity from 25-50 ms, 50-75 ms and 75-100 ms following the load change (hereafter referred to as A, B and C) were computed for each of the 8 different intervals between the onsets of the tone and the torque pulse. Fig. 4 shows data from another subject, where the relative magnitudes of the 3 motor response indices are plotted against z, the time delay from tone onset. For comparison, the averaged response to the tone presented alone is plotted as the bottom trace. The following observations can be made. First, the magnitudes of both B and C increase before the onset of the voluntary EMG response to the tone presented alone (shown by arrow in bottom trace). Second, the magnitude of A does not change until after the onset of voluntary EMG activity. Third, A, B and C decrease together at about the same time, approximately 50 ms following the onset of the voluntary EMG response to the tone. Finally, the EMG responses to the torque pulse are considerably larger during the period of preparation for a rapid movement than during the maintenance of posture (represented by points to the left of time 0).

I

0

o

~

O--

0

.e~e..---e~,

45-go-25 o 2'5 5~9 7'5

I 0 0 125 150 175 2 0 0

TIME (ms)

Fig. 4. Motor response indices for biceps activity from 25-50 ms following torque pulse onset (A), 50-75 ms (B), and 75-100 ms (C) plotted against 3, the time delay from tone onset. For comparison, motor response indices are normalized to per cent of maximum value attained. Peak values shown are 100 ~ . Bottom trace shows the average response to the tone presented alone. Points to the left of time 0 show size of motor response indices when the subject is engaged in the postural task. Computed times of change of motor response indices are indicated by vertical lines, with B changing first, followed by C and then A. Subject D6.

344

Timing of changes in components of the motor response As shown in Fig. 4, the magnitude of the 3 motor response indices (A, B, C) increased from a baseline value during the maintenance of posture to a maximum value before abruptly declining. The time at which each motor response index changed in value was arbitrarily taken to be the time at which that index reached a value equal to the sum of the baseline value and 20 % of the difference between the maximum value attained and the baseline. To assure that baseline noise would not be identified as a change in the value of the index, an additional requirement was applied: that the motor response index must at least double in magnitude from its baseline value for a change to be recognized as having occurred. The times of change of A, B and C determined in this way are illustrated by vertical lines in Fig. 4. These agree well with those times that would be selected from visual inspection, and this was the case for all subjects. Frequency histograms of the differences between the computed times of change of B and C, A and C, and A and B are plotted in Fig. 5. With few exceptions, these differences were greater than zero, suggesting a temporal ordering of the times of change of the 3 indices, with the earliest change occurring in the index representing activity between 75 and 100 ms and the latest time of change being that for activity in 25-50 ms interval. This observation was confirmed statistically by analysis of variance. The null hypothesis, that the times of change of A, B and C were the same, was rejected (F = 21.75 on 2,34 dr, P < 0.01). Additional analyses indicated that the time of change of A was different from that of B and C taken together (F ~ 35.6 on 1,34 df,

ii2 i B_,c 6z

4 -

tA - I c

o tx.

2-40

I Iq III 0

40

80

120

T~ME DIFFERENCE (ms)

Fig. 5. Frequency histograms of differences between computed times of change of motor response indices (tA, tB, tc). See text for description.

345 Subject D20 (~ 7O

Subject D4 d 300

9o_

260 60

c

220 5O x z -

~

"C {80

40

_~ 140 30

I00

/jit

~°

m w

/

60

o

20 0 ~ -20 20

b

10 0

\.~.

~-. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

20

• ----A

0 -20 i

-75 -50 -25

0

25

50

75

TIME (msl

I00

f25

150 I75 2 0 0

--

--o--

_

,~-"~

AC

•

i- . . . . . . . . . . . . . . . :': ...... ~ ~ - - - - " ~ , ~ /

i

-75 -50 -25

i

i

0

25

/

50

i

7.5

~

i

i

.

i

100 125 150 175 2 0 0

TIME (ms)

Fig. 6. Comparison of changes in the biceps motor response indices in two subjects performing rapid flexion (a and c) and extension (b and d) movements of the forearm in response to the auditory signal. Note the absence of an increase in motor response indices when the task is to extend the forearm. In all experiments, subjects maintained the initial position against a 2 N m preload on the biceps muscle and responses were evoked by torque pulses stretching the biceps.

P < 0.01), and that the time of change of B was different from that of C (F =- 7.87 on 1,34 df, P < 0.01). Thus, the statistical findings demonstrate a highly significant temporal ordering (C, then B, then A) of the times of change of the different motor response indices.

Changes in the stretch responses of the antagonist muscle in preparation for a rapid movement

The very early changes in the magnitude of long-latency EMG stretch responses (B, C) preceding rapid forearm voluntary movements raises the question whether these changes are associated with the subject's intent to perform a rapid movement with the stretched muscle or whether they reflect a non-specific increase in the gain of many different muscles associated with an orienting response to the tone signal. This question was addressed by having subjects respond to the tone with rapid extension of the forearm, in place of flexion. EMG responses were evoked as before by torque pulses stretching the biceps muscle, which was now the antagonist of the intended movement. For comparison, the biceps muscle was also preloaded during the experiment by having the subject maintain the initial position against a 2 Nm load. Results of experiments carried out in two of the subjects are shown in Fig. 6, where a and c illustrate the time course of change of A, B and C when the task was to rapidly flex the arm, and b and d (plotted for comparison with the same vertical scales as a and c), the changes when the task was to rapidly extend the arm to the tone cue. In contrast to the large increase seen in the motor response indices when the task was to flex the arm, only a late decrease in magnitude is evident when subjects were asked to respond by

346

80

60 40 20

C~ ~

0 -20

/\ -ioo-sb

o

5o

ioo

1go ado

TrME (ms)

Fig. 7. Temporal comparison of changes in motor response indices (A, B, C), EMG responses to the tone presented alone (middle trace) and acceleration of the limb in the tone-only condition (lower trace). Scales: motor response indices: actual values; EMG response: arbitrary scale; angular acceleration: 5000*/sL Subject D7.

extending their arms. Thus, the increase in gain associated with the flexion movement was related to the intent of the subject, rather than to a non-specific increase in excitability resulting from the auditory signal. DISCUSSION The experiments described in this report demonstrate that during the period of transition from a maintained posture against a steady load to a rapid voluntary movement, there is a large, though transient, increase in the magnitude of long-latency E M G stretch responses (M2, M3) to stretch of the agonist muscle. This increase precedes the earliest voluntary increase in E M G activity. By contrast, the amplitude of the short-latency response (M1) is not altered until after the descending intentional volley reaches the motoneurons. These differences in the timing of changes in short- and long-latency stretch responses prior to movement have not been reported previously. However, two other studies have been performed to examine alterations in stretch responses before voluntary movements. Gottlieb and Agarwa125, studying the stretch responses of the soleus muscle in human subjects, observed an increase in the gain of the myotatic reflex at the initiation of a voluntary soleus contraction, followed by sudden and total inhibition of this reflex during the movement itself. It is

347 difficult to compare their findings with the present results, because of the uncertain correspondence between the stretch response of the soleus and the short- and longlatency responses recorded in the arm muscles. In another study, Hallett and Marsden 27 found no change in the M2 stretch response of the human thumb flexor during the interval between a warning stimulus and a signal to initiate movement, other than that related to subtle changes in the background EMG discharge. However, these results are not comparable to the present findings, since no investigation was made of changes in reflex gain following the signal to initiate muscular contraction. The present finding, that the amplitude of the M1 response is not altered until after the initiation of voluntary EMG activity, differs from observations of previous investigations, in which increases in the tendon jerk and H-reflex of the agonist muscle were reported to precede voluntary motor activation by approximately 100 ms10,22,32. In the latter studies, reflexes were evoked in electrically silent muscles, for which the level of excitability of the motoneuron pool was unknown. Subthreshold changes in excitability resulting from suprasegmental or segmental input therefore could have influenced the magnitude of the tendon jerk without being detected. By contrast, in the current study, both the initial tension and the level of discharge of the agonist and antagonist muscles were monitored to assure uniform initial conditions. The need to control the background EMG in addition to the pre-existing muscle load is supported by the finding of Evarts and Vaughn 19 that the size of tendon jerk can vary with the background discharge of the stretched muscle even with a constant preload. In the present study, subjects were encouraged to flex their forearm as rapidly as possible in response to the auditory signal, typically resulting in maximum velocities of 400-800°/s. Under these conditions, the increase in stretch response amplitude was followed by an abrupt decrease (cf. Figs. 4, 6 and 7). Similar decreases in the amplitudes of stretch responses of agonist muscles during very fast movements have been described by other investigators9A2,13,25. Desmedt and Godaux ~2 proposed that long loop reflex control is switched off after the ballistic impulse is delivered to the motoneurons and suggested that a gate is opened at either the thalamic or cortical level. Other investigators14, however, have reported that sufficiently large torque steps delivered during maximum velocity movements are capable of evoking reflex responses, implying that the reflex suppression seen during ballistic movements may be attributable to peripheral factors. Studies of the discharge of muscle spindle afferents during voluntary shortening contractions carried out at different speeds demonstrate that both primary and secondary endings fall silent above certain velocities of muscle shorteningTM. For unobstructed shortening, Prochazka, Stephens and Wand 44 have suggested that silencing of spindles occurs at rates of shortening exceeding0.2 resting lengths per second. The maximum rate of shortening of the biceps muscle in the present study was between 2 and 4 resting lengths per second, well above the velocity necessary to produce suppression of spindle discharge in unobstructed muscles. Velocities of 0.2 resting lengths per second were attained within 10 ms after the initiation of the movement, perhaps accounting for the very early decrease in both short- and longlatency responses to muscle stretch.

348 A possible contribution of Golgi tendon organ inhibition to the decrease in reflex amplitude observed during rapid voluntarily initiated movements must also be considered. During the period of maximum limb acceleration, the stretch response computed by subtracting the response to the tone alone from the response to the tone plus the torque pulse was frequently negative in sign (cf. Fig. 7). This finding may be explained by an unmasking of Golgi tendon organ inhibition during the interval in which spindle receptors are presumably unloaded. In several subjects, a recovery in the amplitude of short- and long-latency stretch responses was seen as the acceleration of the limb began to decrease (Fig. 7). This increase in gain may have resulted from the intrafusal contraction catching up with the shortening of extrafusal fibers or from a decrease in Golgi tendon organ inhibition as the force declined. Gottlieb and Agarwal ~5, studying much slower movements, found that the period of reflex suppression following muscular contraction could last several hundred milliseconds, suggesting that other, perhaps central, mechanisms may be involved in modulating reflex gain following movement initiation.

Functional significance The finding of an early change in long-latency responses to stretch raises the question of its possible functional significance. It has been proposed that long-latency stretch responses may provide load compensation 7. In one subject studied in the present experiments, the amplitude of the long-latency stretch response (M2, M3) measured just before movement initiation exceeded that of the initial burst of E M G activity during a maximum velocity movement (cf. top and bottom traces of Fig. 2). This finding suggests that substantial load compensation may be provided in particular cases. However, because of the rapid drop in the amplitude of short- and long-latency stretch responses during the accelerative phase of the movement, any load compensation through this mechanism would be effective for only a very brief time. It is difficult to understand how such a brief period of load compensation could serve a useful function. The role of long-latency stretch responses in the regulation of limb stiffness has been stressed by Kwan and his collaborators3L These investigators found that a stretch response with a latency similar to M2 helped to maintain muscular stiffness at a constant value when subjects attempted to resist ramp disturbances in torque. As demonstrated in the present study, the gain of segmental reflexes does not appear to be independently modifiable by the intention of the subject (see Fig. 6). Long-latency motor responses to muscle stretch, especially those occurring at an M3 latency, offer alternative mechanisms for regulation of stiffness that are perhaps more adaptable to the needs of the subject. It is likely, however, that such a mechanism for controlling stiffness may be more appropriate to slower and more finely graded movements rather than maximum velocity contractions. During the former type of movements the gain of long-latency pathways appears to be quite high 16. A third function that may be served by the early transient increase in the gain of pathways mediating long-latency stretch responses is to provide assistance in overcoming inertia at the initiation of movements. Considerable evidence has been

349 obtained for the coactivation of skeletomotor and fusimotor systems during isometric as well as shortening and lengthening contractions 2,3. Under certain situations, the discharge of muscle spindle afferents can be greater during muscle shortening than during passive stretch 3. Furthermore, the increase in spindle receptor activity accompanying muscular contraction does not require a divergence from intended length, since an increase in discharge is seen even during perfect tracking 31. Smith et al. 45 investigated the effect of selective fusimotor blockade in human subjects and found that it resulted in a marked decrease in acceleration of rapidly initiated arm movements, such as dart throwing. These findings can be interpreted as evidence for the failure of alpha motoneurons to receive excitatory support from muscle spindle afterents discharging in response to fusimotor input. Such assistive support would be most important during the initial phase of movements, when the inertia of the limb is to be overcome. Recordings from single Ia afferents provide some evidence for this view. For rapidly initiated movements, afferent discharge is usually very brief and restricted to the period before appreciable limb acceleration occurs 3,5,21. While input from either short- or long-latency reflex pathways could facilitate alpha motoneurons during the initial burst of E M G associated with a voluntarily initiated movement, the fact that the gain of long-latency pathways is much greater suggests that they may provide a more powerful assistive input. Participants in the current study included 6 subjects who were highly trained in karate. In this group, the transient increase in the magnitude of long-latency responses prior to movement was considerably greater than in the remaining subjects. The karate-trained subjects also exhibited the greatest acceleration and highest peak velocities of movement. These findings are consistent with the earlier observations made by Milner-Brown et al. 39 of prominent long loop reflexes in weightlifters, and suggest that the increase in the gain of pathways mediating long-latency stretch responses prior to rapid movements may be important in providing greater force to overcome inertia. Data to test this hypothesis more rigorously are presently being collected. REFERENCES 1 Allum, J. H. J., Responses to load disturbances in human shoulder muscles: The hypothesis that one component is a pulse test information signal, Exp. Brain Res., 22 (1975) 307-326. 2 Burke, D., Muscle spindle function during movement, Trends Neurosck, 3 (1980) 251-253. 3 Burke, D., Hagbarth, K. E. and L/Sfstedt, L., Muscle spindle activity in man during shortening and lengthening contractions, J. Physiol. (Lond.), 227 (1978) 131-142. 4 Chan, C. W. Y., Kearney, R. E. and Melvill Jones, G., Tibialis anterior response to sudden ankle displacements in normal and Parkinsonian subjects, Brain Research, 173 (1979) 303-314. 5 Cody, F. W. J., Harrison, L. M. and Taylor, A., Analysis of activity of muscle spindles of the jawclosing muscles during normal movements in the cat, J. PhysioL (Lond.), 253 (1975) 565-582. 6 Conrad, B., The motor cortex as a primary devicefor fast adjustment of programmed motor patterns to afferent signals. In J. E. Desmedt (Ed.), Cerebral Motor Control in Man: Long Loop Mechanisms, Progress in Clinical Neurophysiology, Vol. 4, Karger, Basel, 1978, pp. 123-140. 7 Conrad, B. and Meyer-Lohmann, J., The long-loop transcortical load compensating reflex, Trends Neurosci., 3 (1980) 269-272. 8 Conrad, B., Meyer-Lohmann, J., Matsunami, K. and Brooks, V. B., Precentral unit activity following torque pulse injections into elbow movements, Brain Research, 94 (1975) 219-236. 9 Cooke, J. D., The role of stretch reflexes during active movements, Brain Research, 181 (1980) 493-497.

350 10 Coquery, J. and Coulmance, M., Variations d'amplitude des reflexes monosynaptiques avant un mouvement volontaire, Physiol. Behav., 6 (1971) 65-69. 11 Crago, P. E., Houk, J. C. and Hasan, Z., Regulatory actions of human stretch reflex, J. Neurophysiol., 39 (1976) 925-935. 12 Desmedt, J. E. and Godaux, E., Ballistic skilled movements: Load compensation and patterning of the motor commands. In J. E. Desmedt (Ed.), Cerebral Motor Control in Man: Long Loop Mechanisms, Progress in Clinical Neurophysiology, VoL 4, Karger, Basel, 1978, pp. 21-55. 13 Desmedt, J. E. and Godaux, E., Voluntary motor commands in human ballistic movements, Ann. Neurol., 5 (1979) 415-421. 14 DuFresne, J. R., Gurfinkel, V. S., Soechting, J. F. and Terzuolo, C. A., Response to transient disturbances during intentional forearm flexion in man, Brain Research, 150 (1978) 103-115. 15 DuFresne, J. R., Soechting, J. F. and Terzuolo, C. A., Electromyographic response to pseudorandom torque disturbances of human forearm position, Neuroscience, 3 (1978) 1213-1226. 16 Evarts, E. V. and Fromm, C., The pyramidal tract neuron as suming point in a closed-loop control system in the monkey. In J. E. Desmedt (Ed.), Cerebral Motor Control in Man: LongLoop Mechanisms, Progress in Clinical Neurophysiology, Vol. 4, Karger, Basel, 1978, pp. 56--69. 17 Evarts, E. V. and Granit, R., Relations of reflexes and intended movements. In S. Homma (Ed.), Understanding the Stretch Reflex, Progress in Brain Research, Vol. 44, Elsevier, Amsterdam, 1976, pp. 1-14. 18 Evarts, E. V. and Tanji, J., Gating of motor cortex reflexes by prior instruction, Brain Research, 71 (1974) 479-494. 19 Evarts, E. V. and Vaughn, W. J., Intended arm movements in response to externally produced arm displacements in man. In J. E. Desmedt (Ed.), Cerebral Motor Control in Man: Long Loop Mechanisms, Progress in Clinical Neurophysiology, Vol. 4, Karger Basel, 1978, pp. 178-192. 20 Ghez, C. and Shinoda, Y., Spinal mechanisms of the functional stretch reflex, Exp. Brain Res., 32 (1978) 55-68. 21 Goodwin, G. M. and Luschei, E. S., Discharge of spindle afferents from jaw-closing muscles during chewing in alert monkeys, J. NeurophysioL, 38 (1975) 560-571. 22 Gottlieb, G. L. and Agarwal, G. C., Modulation of postural reflexes by voluntary movement. 2. Modulation at an inactive joint, J. Neurol. Neurosurg. Psychiat., 36 (1973) 540-546. 23 Gottlieb, G. L. and Agarwal, G. C., Response to sudden torques about ankle in man: myotatic reflex, J. Neurophysiol., 42 (1979) 91-106. 24 Gottlieb, G. L. and Agarwal, G. C., Response to sudden torques about ankle in man. II. Postmyotatic reactions, J. Neurophysiol., 43 (1980) 86-101. 25 Gottlieb, G. L. and Agarwal, G. C., Response to sudden torques about ankle in man. III. Suppression of stretch-evoked responses during phasic contraction, J. NeurophysioL, 44 (1980) 233-246. 26 Hagbarth, K. E., E M G studies of stretch reflexes in man, Electroenceph. clin. Neurophysiol., Suppl. 25 (1967) 74-79. 27 Hallett, M. and Marsden, C. D., Long-latency stretch reflex of the human thumb prior to the ini tiation of voluntary movement, Soc. Neurosci. Abstr., 6 (1980) 218. 28 Hammond, P. H., An experimental study of servo action in human muscular control. In Proc. Third Int. Conf. on Medical Electronics, Int. Fed. Med. Electron., London, 1960, pp. 190--199. 29 Hendrie, A. and Lee, R. G., Selective effects of vibration on human spinal and long-loop reflexes, Brain Research, 157 (1978) 369-375. 30 Houk, J. C., Participation of reflex mechanisms and reaction time processes in the compensatory adjustments to mechanical disturbances. In J. E. Desmedt (Ed.), Cerebral Motor Control in Man: Long Loop Mechanisms, Progress in Clinical Neurophysiology, Vol. 4, Karger, Basel, 1978, pp. 193-215. 31 Hulliger, M. and Vallbo, A. B., The responses of muscle spindle afferents during voluntary tracking movements in man. Load dependent servo assistance?, Brain Research, 166 (1979) 401-404. 32 Kots, Y. M., Supraspinal control of the segmental centres of muscle antagonists in man. I. Reflex excitability of the motor neurons of muscle antagonists in the period of organization of voluntary movement, Biophysics, 14 (1969) 176-183. 33 Kwan, H. C., Murphy, J. T. and Repeck, M. W., Control of stiffness by the medium latency electromyographic response to limb perturbation, Canad. J. PhysioL Pharmacol., 57 (1979) 277-285. 34 Lamarre, Y. and Lund, J. P., Load compensation in human masseter muscles, J. Physiol. (Lond.), 253 (1975) 21-35. 35 Lee, R. G. and Tatton, W. G., Motor responses to sudden limb displacements in primates with specific CNS lesions and in human patients with motor system disorders, Canad. J. Neurol. Sci., 2 (1975) 285-293.

351 36 Marsden, C. D., Merton, P. A. and Morton, H. B., Servo action in human voluntary movement, Nature (Lond.), 238 (1972) 140-143. 37 Marsden, C. D., Merton, P. A. and Morton, H. B., Stretch reflex and servo action in a variety of human muscles, J. Physiol. (Lond.), 259 (1976) 531-560. 38 Marsden, C. D., Merton, P. A., Morton, H. B., Adam, J. E. R. and Hallett, M., Automatic and voluntary responses to muscle stretch in man. In J. E. Desmedt (Ed.), Cerebral Motor Control in Man: Long Loop Mechanisms, Progress in Clinical Neurophysiology, VoL 4, Karger, Basel, 1978, pp. 167-177. 39 Milner-Brown, H. S., Stein, R. B. and Lee, R. G., Synchronization of human motor units: Possible roles of exercise and supraspinal reflexes, Electroenceph. clin. NeurophysioL, 38 (1975) 245-254. 40 Monster, A. W., Loading reflexes during two types of voluntary muscle contraction. In R. B. Stein, K. G. Pearson, R. S. Smith and J. B. Redford (Eds.), Control of Posture and Locomotion, Plenum, New York, 1974, pp. 347-362. 41 Mortimer, J. A. and Johnson, M. J., Modifications in segmental and long-loop reflexes in man by preload and instruction. In J. E. Desmedt (Ed.), International Symposium on Human Reflexes and Motor Disorders, Brussels Univ. Press, Brussels, 1976, pp. 131-132. 42 Mortimer, J. A. and Webster, D. D., Evidence for a quantitative association between EMG stretch responses and Parkinsonian rigidity, Brain Research, 162 (1979) 169-173. 43 Mortimer, J. A. and Webster, D. D., Relationship between quantitative measures of rigidity and tremor and the electromyographic responses to load perturbations in unselected normal subjects and Parkinsonian patients. In J. E. Desmedt (Ed.), Cerebral Motor Control in Man: Long Loop Mechanisms, Progress in Clinical Neurophysiology, Vol. 4, Karger, Basel, 1978, pp. 342-360. 44 Prochazka, A., Stephens, J. A. and Wand, P., Muscle spindle discharge in normal and obstructed movements, J. Physiol. (Lond.), 287 (1979) 57-66. 45 Smith, J. L., Roberts, E. M. and Atkins, E., Fusimotor neuron block and voluntary arm movement in man, Amer. J. phys. Med., 51 (1972) 225-239. 46 Strick, P. L., Cerebellar involvement in 'volitional' muscle responses to load changes. In J. E. Desmedt (Ed.), Cerebral Motor Control in Man: Long Loop Mechanisms, Karger, Basel, 1978, pp. 85-93. 47 Tatton, W. G., Bawa, P., Bruce, I. C. and Lee, R. G., Long loop reflexes in monkeys: An interpretative base for human reflexes. In J. E. Desmedt (Ed.), Cerebral Motor Control in Man: Long Loop Mechanisms, Progress in Clinical Neurophysiology, VoL 4, Karger, Basel, 1978, pp. 229-245. 48 Thomas, J. S., Brown, J. and Lucier, G. E., Influence of task set on muscular responses to arm perturbations in normal subjects and Parkinson patients, Exp. Neurol., 55 (1977) 618-628.