- Email: [email protected]
Interlimb reflexes evoked in human arm muscles by ankle displacement
Electroencephalography and Clinical Neurophysiology, 1981, 5 2 : 6 5 - - 7 1 Elsevier/North-Holland Scientific Publishers, Ltd.
65
INTERLIMB REFLEXES EVOKED IN HUMAN ARM MUSCLES BY ANKLE DISPLACEMENT R.E. KEARNEY and C.W.Y. CHAN
Biomedical Engineering Unit, and Aviation Medical Research Unit, McGill University, Montreal, Que. H3G 1 Y6 (Canada) (Accepted for publication: March 4, 1981 )
The control of posture, locomotion and voluntary movement requires the coordinated regulation of muscles acting at different joints and limbs. It has been suggested that interlimb reflexes may play an important role in this coordination (Halbertsma et al. 1977; Roberts 1979). For example, in the cat, reflex connections between hind limb and forelimb muscles are mediated by both spino-bulbospinal (Shimamura and Livingston 1963; Shimamura and Akert 1965) and propriospinal (Gernandt and Megirian 1961; Miller et al. 1973) pathways, and the later connections appear to be appropriate for the coordination of hind and forelimb movements during locomotion (Miller et al. 1975). We have previously demonstrated that reflex responses to cutaneous electrical stimulation of the foot can be detected in both leg and arm muscles of normal human subjects by averaging stimulus-related changes in tonic electromyographic (EMG) activity (Kearney and Chan 1979). We suggested that these responses could be involved in the coordination of a generalized response to noxious stimuli. If such interlimb reflexes are to play an important role in the coordination of movement, then a physiologically meaningful stimulus such as the displacement of a joint should be effective in evoking them. Furthermore, displacement of the joint in different directions should evoke significantly different responses. The present report describes a study which has demonstrated that ankle displacement evokes a substantial reflex response in human arm muscles whose nature depends upon the direction in which the ankle is displaced.
Methods Six subjects between the ages of 20 and 25 with no history of neuromuscular disease were examined. EMGs from gastrocnemius (G), tibialis anterior (TA), biceps brachii (BB) and triceps brachii (TB) were measured with surface electrodes, amplified, rectified and smoothed before recording. Subjects lay supine with the right foot attached to the pedal of a high performance electrohydraulic actuator by means of a custom-fitted plastic boot. The upper arm was externally rotated and abducted at 90 °, with the elbow flexed at 90 ° and held supine by a rigid frame. Subjects were instructed to maintain a steady, tonic contraction of the relevant muscle aided by a meter display of the smoothed and rectified EMG. Stimuli consisted of step displacements of ankle position (0.1 tad amplitude; 45 msec rise time) in either the dorsiflexing or plantarflexing direction. Stimuli were only applied when the tonic EMG activity was within 15% of the desired level. A PDP-11/40 computer sampled the EMG at 2000 Hz for 70 msec prior to, and 170 msec after, the stimulus. An on-line data analysis program computed the ensemble average response and Wiener-filtered it (Walter 1969; Kearney 1979). Stimuli were applied until an acceptable signal-to-noise ratio was obtained, as judged by comparing the pre- and poststimulus portions of the average. Anywhere from 50 to 500 responses were required depending upon the signal-to-noise ratio of the evoked EMG. The mean value and
0013-4649/81/0000--0000/$02.50 © Elsevier/North-Holland Scientific Publishers, Ltd.
66
R.E. KEARNEY, C.W.Y. CHAN
standard deviation of the prestimulus portion of the average were computed. Stretch-evoked responses were only considered to be significant if their amplitude was greater than twice the standard deviation of the prestimulus portion of the record. Latencies of responses were measured from the onset of displacement to the time at which the magnitude of the response exceeded twice the prestimulus standard deviation. Response magnitudes were expressed as the percentage modulation of the mean prestimulus activity. Means and standard errors (S.E.) presented below are group values computed from data from all 6 subjects.
Results
Responses to dorsiflexing displacements A typical set of responses evoked by dorsi-
A
flexing displacements of the ankle is illustrated in Fig. 1A. The response of G, which was stretched by the displacement, was dominated by a period of intense excitation (group mean modulation 800 + S.E. 130%) lasting from about 40 to 80 msec (mean 37 ± S.E. 2 msec to 84 ± S.E. 5 msec), followed by a period of decreased activity. This pattern of response is in agreement with that previously described (Kearney and Chan 1976; Chan et al. 1979b; Gottlieb and Agarwal 1979). Dorsiflexion of the ankle did not evoke a significant response in the elbow flexor, BB. Small, long latency responses were occasionally observed but these were not significantly different from control responses. In contrast, a consistent response was observed in TB, an elbow extensor, in all 6 subjects. A short latency (peak at 55 ± S.E. 3 msec) excitation was observed which although small (mean modulation 25 ± S.E. 4%) was significantly
i
I
G
i
I
j!
I
~
400 % i
TA
I
',
~
_ _ _
_
~
_,~L_..,
200
...........~:_ _,____~_+-:~
I
BB
I ,
TB
~[
5o~
j
I
I
I
,
I I
I
POSITION I~ / //~ [. /
......
1
.1 rad
POSITION I 1
I I
20
ms
Fig. 1. Patterns of EMG potentials evoked in ipsilateral arm and leg muscles by ankle displacements. Baseline for each record corresponds to mean level of prestimulus EMG activity. Positive EMG rectification was used so that an upward deflection of the trace corresponds to increased EMG activity. EMG scales are given as a percentage of the tonic activity. Bottom trace is ankle position with dorsiflexion denoted by an upward deflection. The dashed line in this and all subsequent figures indicates the start of the ankle displacement. A: responses evoked by dorsiflexing displacements in gastrocnemius (G), biceps brachii (BB), and triceps brachii (TB}. B: responses evoked by plantarflexing displacements in tibialis anterior (TA), biceps brachii (BB), and triceps brachii {TB).
RESPONSE OF HUMAN ARM MUSCLES TO ANKLE DISPLACEMENT
greater than the prestimulus level in all cases, according to the criterion described in Methods. This was followed by a decrease in activity of somewhat greater magnitude (mean modulation 38-+ S.E. 7%) with an onset at 72.5 + S.E. 8 msec and a peak at 90.2 + S.E. 4 msec. In our previous study (Kearney and Chan 1979) we found t h a t cutaneous electrical stimulation of the sole of the f o o t evoked a similar decrease in TB activity at about this latency.
Responses to plantarflexing displacements
67
Consistency of response pattern The pattern of response evoked in the arm muscles by ankle displacement was very consistent both from subject to subject and from day to day in the same subject. This is illustrated in Fig. 2 which shows the TB response to ankle plantarflexion in 4 subjects, two of whom (top two traces) were studied on two different days. Note the similarity o f the responses. In general, as noted above, there was significantly more variation in the amplitudes of the responses than in their latencies.
Sensory origin of the responses Plantarflexing displacements of the ankle evoked a somewhat different pattern of response as illustrated in Fig. l B . The response of TA, which was stretched in this case, consisted of a prolonged, biphasic excitation (mean modulation 418 + S.E. 111%) lasting from about 40 to more than 100 msec (mean 42 -+ S.E. 1 msec to 110 + S.E. 2 msec), followed by a period of decreased activity. Again, this pattern of response is consistent with what has been previously reported (Kearney and Chan 1976; Chan et al. 1979a; Gottlieb and Agarwal 1979)• Plantarflexing displacements of the ankle evoked responses consisting o f an early excitation followed by a period of decreased activity in both TB and BB. In TB, the early excitation started at 54 -+ S.E. 2 msec, peaked at 70 -+ S.E. 3 msec and modulated the tonic activity an average of 74 +- S.E. 18%. The subsequent decrease in activity started at 89 + S.E. 3 msec, peaked at 102 + 2 msec and modulated the tonic activity an average of 48 -+ S.E. 11%. This pattern is in contrast to t h a t evoked by dorsiflexion in which the early excitation was always o f much smaller magnitude than the subsequent decrease in activity. The pattern o f response in BB was similar to that of TB although always of much smaller magnitude. The early excitation modulated the tonic activity a mean of only 30 -+ S.E. 8%, although the latency was similar (onset of 51 -+ 2 msec and a peak at 67 -+ S.E. 3 msec).
There was some noise and vibration associated with the stimulus which might in itself evoke a response (e.g., Rossignol and Melvill Jones 1976). Control experiments were therefore run with the subject in the same position as t h a t used in the displacement experiments, but with the foot removed from the actuator
TB
I
I I
A^AAAA
A
A ..... I I I
20
J
,~
V , r
7~
,
A
,
ms
Fig. 2. Repeatability of triceps brachii response to plantarflexing displacements of the ankle. Responses observed in 4 different subjects are shown. Superimposed responses for subjects T B and SF were obtained on different days. All responses have been normalized to have the same peak-to-peak amplitude.
68
R.E. K E A R N E Y , C.W.Y. CHAN
100
I
Fig. 3. A: typical mental situation pedal t o p r e v e n t t i o n s as in A b u t p e d a l to p e r m i t it
r e s p o n s e o f s u b j e c t in t h e experiwith the foot removed from the a n k l e d i s p l a c e m e n t . B: same condit h e s u b j e c t ' s f o o t was placed o n t h e to be displaced.
pedal. The responses observed in these cases were always much smaller than those evoked by ankle displacement and had a longer latency (greater than 100 msec). This is illusGrated in Fig. 3 which shows responses evoked in TB by plantarflexing displacements of the ankle (Fig. 3B) and by a control stimulus in which everything was the same except that the subject's foot was removed from the actuator pedal (Fig. 3A). Clearly, the control response is insignificant when compared to that evoked by ankle displacement. Thus it seems certain that the responses observed in the arm muscles in these experiments were evoked by ankle displacement rather than by vibration or sound. These displacements must have stimulated both proprioceptors and exteroceptors whose relative contributions to the response remain to be determined. The initial component of the TB response had a latency (mean 54 msec) only
12--17 msec longer than those of TA {mean 42 msec) and G (mean 37 msec). In view of their short latencies, the initial components of leg muscle responses are almost certainly mediated by muscle spindle afferents (Kearney and Chan 1976; Gottlieb and Agarwal 1979). If central conduction time is allowed for, it seems likely that at least the initial component of the TB response is also mediated by muscle afferents. A contribution from cutaneous afferents, in particular from pressure receptors which must have been excited by our stimuli, cannot be ruled out. However, due to the lower conduction velocity of these afferents, their effects would be expected to occur at longer latencies than those of muscle afferents. This idea is supported by our previous finding (Kearney and Chan 1979) that cutaneous electrical stimulation never evoked responses in arm muscles at latencies less than 70 msec.
Discussion These findings demonstrate that displacements of the ankle evoked significant reflex responses in arm muscles. The relatively small ankle displacements {less than 6 ° ) used in our study still modulated the tonic activity of TB by as much as 80%. The possible significance of these responses is apparent when it is noted that in some cases their magnitude was as much as 20% of that evoked in the muscle that was actually stretched. Furthermore, the stretch evoked responses occurred in the arm muscles at very short latencies. For example, the early excitatory component of the TB response had a mean latency of 54 msec only 12--17 msec greater than those in TA (mean 42 msec) and G (mean 37 msec). In contrast, in our previous study (Kearney and Chan 1979), interlimb responses evoked by cutaneous electrical stimulation, at levels more than 3 times the sensory threshold, rarely modulated the tonic activity by more than 15--20% and never had a latency of less than 70 msec. It is difficult to compare the relative magni-
RESPONSE OF HUMAN ARM MUSCLES TO ANKLE DISPLACEMENT tudes of the displacement stimuli used in the present experiments to those of the electrical stimuli used previously. Nevertheless, it is notable that small displacements evoked interlimb responses which were larger and of shorter latency than those evoked b y electrical stimuli o f the highest intensity subjects would tolerate. This would seem to indicate that the interlimb reflex connections o f proprioceptive afferents are stronger and more direct than those of cutaneous origin. If this is so, then interlimb connections o f proprioceptive afferents may play an important role in the coordination of movement. The interlimb reflexes described in this paper should be compared to the 'anticipatory postural reflexes' studied b y Marsden and co-workers (Marsden et al. 1978; Traub et al. 1980). The original observation was that the unexpected perturbation of an arm while executing a simple manual tracking task evoked reflex responses in muscles of the opposite arm, trunk and leg engaged in supporting the b o d y (Marsden et al. 1978). Subsequently, a more detailed study was made of reflexes evoked in the ankle muscles o f normal and pathological subjects b y displacements of the arm (Traub et al. 1980). The essential features of these descending reflexes appear to be the same as those of the ascending reflexes we have studied. Thus, Marsden et al. (1978) concluded that the reflexes originated in the arm that was pulled and that the afferent input was not cutaneous since the response was not modified b y anesthesia o f the hand and wrist. We have reached similar conclusions a b o u t the reflexes evoked in the arm muscles b y ankle displacements based on the results of the control experiments and the finding that ankle displacement is a more effective stimulus than cutaneous electrical stimulation. Furthermore, our finding that the activity in TB is predominantly increased b y plantarflexing displacements of the ankle and predominantly decreased b y dorsiflexing displacements is consistent with the finding that the polarity o f the triceps surae response depends u p o n the direction is which the arm
69
is pulled (Traub et al. 1980). Thus it appears that the ascending reflexes linking arm muscle activity to ankle displacement are organized in much the same way as the descending reflexes linking the activity of ankle muscles to arm displacement. One difference between the two studies is in the latencies of the responses. The 'anticipatory postural reflexes' evoked in triceps surae b y arm displacement had a mean latency of 83 msec (Traub et al. 1980). In contrast, we found that ankle displacement evoked reflex responses in TB at 55 msec. These different latencies might indicate that the descending and ascending reflexes utilize different pathways. Traub et al. {1980) suggested that a transcortical pathway may be involved in generating the descending reflexes. In contrast the short latency of the responses we have observed makes a propriospinal pathway seem likely. However, the different latencies might also arise from differences in experimental methods. We used rapid changes in ankle position in our experiments which would be expected to evoke a synchronous afferent volley. Little temporal summation would therefore be required to evoke a reflex response and its latency should be determined primarily b y conduction delay. Our finding that the reflex responses o f TA and G occurred at latencies close to those of their tendon jerks supports this argument. In contrast, in the experiments of Traub et al. (1980) relatively slow displacements were used which might give rise to an asynchronous afferent volley. Consequently, some temporal summation would be required before a reflex response was evoked resulting in an increased latency. It remains to consider the functional significance of these responses. Marsden and co-workers considered the responses t h e y observed to be postural in nature since t h e y occurred in postural muscles, were in the appropriate direction to compensate for the postural disturbances evoked b y the stimuli, and were generally reduced in parkinsonian patients who displayed postural instability
70 (Marsden et al. 1978; Traub et al. 1980). The functional role of the responses observed in our experiments is less clear. The predomiant response occurred in TB and was dominated by excitation for plantarflexing displacements and b y a decrease in activity for dorsiflexing displacements. This pattern had no obvious role in postural stabilization but then postural stabilization was not required in our paradigm. Alternatively, interlimb reflexes could be involved in the coordination of movement. For example, the excitation of TB (an elbow extensor) when TA (an ankle flexor) is stretched, and the predominant decrease in TB activity which occurs when G (an ankle extensor) is stretched, appears to be appropriate for the coordination of reciprocal arm and leg movements during locomotion. However, the pattern of interlimb reflexes evoked in other muscles will have to be determined before any such functional hypothesis can be substantiated. Furthermore, it should be realized that, in the cat, interlimb reflexes have been shown to vary with both static limb position (Grillner and Rossignol 1978), and the phase of the step cycle during locomotion (Halbertsma et al. 1977). Such effects will have to be investigated in man before the functional significance of human interlimb reflexes can be properly understood.
Summary Interlimb reflexes evoked by ankle displacements were studied in arm muscles of 6 normal subjects. EMGs from grastrocnemius (G), tibialis anterior (TA), biceps brachii (BB), and triceps brachii (TB) were amplified, rectified and low-pass filtered before recording. Averaging and Wiener filtering were used to detect changes in tonic EMG activity evoked b y dorsiflexing or plantarflexing displacements of the ankle. A consistent pattern of response was observed in all subjects. In the leg muscles, the responses to stretch were consistent with previous reports. In the arm muscles, the
R.E. KEARNEY, C.W.Y. CHAN response of TB was dominant. Dorsiflexing displacements of the ankle evoked a small excitation followed by a more marked decrease in TB activity but had no effect on BB. In contrast, plantarflexing displacements of the ankle resulted in a large, early period of excitation followed by a decreased level of activity in TB. A similar but smaller pattern of activity was observed in BB. It is notable that the TB responses to displacement were sizable, often modulating the tonic EMG activity by as much as 80%. Interlimb reflexes evoked b y ankle displacement were larger and of shorter latency than those evoked by cutaneous electrical stimulation of the foot reported previously. This suggests that proprioceptive afferents may have stronger and more direct interlimb reflex connections than cutaneous afferents and may therefore play an important role in the coordination of movement.
R6sume Rdflexes inter-segmentaires dvoquds par des ddplacements de la cheville dans des muscles du bras humain Les r~flexes inter-segmentaires dvoquds par des d6placements de la cheville ont 6t6 ~tudids au niveau des muscles du bras chez 6 sujets normaux. Les dlectromyogrammes (EMG) des muscles gastrocn6miens (G), tibial antdrieur (TA), biceps brachial (BB), et triceps brachial (TB) ont 6t6 amplifids, rectifi6s et filtr~s avant l'enregistrement. Le moyennage et le filtrage Wiener ont 6t6 utilis6s afin de ddtecter les changements de l'activitd EMG tonique ~voqu~s par les d6placements en dorsiflexion ou en flexion plantaire de la cheville. Un 'pattern' constant de r6ponse a 6td observ6 chez tous les sujets. Dans les muscles de la jambe, les r6ponses ~ l'6tirement sont conformes aux 6tudes ant6rieures. Dans les muscles du bras, la r6ponse du TB est dominante. Les d6placements en dorsiflexion de la cheville ont 6voqu6 une diminution significa-
RESPONSE OF HUMAN ARM MUSCLES TO ANKLE DISPLACEMENT
tire de l'activit~ du TB mais n ' o n t pas eu d'effet sur le BB. Par contre, les d~placements de la cheville en flexion plantaire entrai'nent une p~riode initiale d'excitation suivie d'une diminution du niveau d'activit~ du TB. Un 'pattern' d'activit~ similaire mais r~duit s'observe au niveau du BB. I1 est ~ noter que les r~ponses du TB au d~placement sont mesurabies, modulant souvent l'activit~ EMG tonique d'environ 80%. Les r~flexes inter-segmentaires ~voqu~s par le d~placement de la cheville sont plus importants et de latence plus courte que ceux ~voqu~s par les stimulations cutan~es ~lectriques du pied pr~c~demment rapport~es. Ceci sugg~re que les aff~rences proprioceptives pourraient avoir des liens plus directs et plus ~troits avec les r~flexes intersegmentaires que les aff~rences cutan~es, et peuvent ainsi jouer un rSle important dans la coordination des mouvements. This work was supported by grants from the Canadian Medical Research Council and La Counseil de la Recherche en Sant~ du Qu4bec.
References Chan, C.W.Y., Kearney, R.E. and Melvill Jones, G. Tibialis anterior responses to sudden ankle displacements in normal and parkinsonian subjects. Brain Res., 1979a, 173: 303--314. Chan, C.W.Y., Melvill Jones, G., Kearney, R,E. and Watt, D.G.D. The 'late' electromyographic response to limb displacement in man. I. Evidence for supraspinal contribution. Electroenceph. clin. Neurophysiol., 1979b, 46: 173--181. Gernandt, B.E. and Megirian, D. Ascending propriospinal mechanisms. J. Neurophysiol., 1961, 24: 364--376. Gottlieb, G.L. and Agarwal, G.C. Response to sudden torques about the ankle: myotatic reflex. J. Neurophysiol., 1979, 46: 173--181. Grillner, S. and Rossignol, S. Contralateral reflex reversal controlled by limb position in the acute spinal cat injected with clonidine i.v. Brain Res., 1978, 144: 411--414.
71
Halbertsma, J.M., Miller, S. and Van der Mechd, F.G.S. Basic programs for the phasing of flexion and extension movements of the limbs during locomotion. In: R.M. Herman, S. Grillner, P.G.S. Stein and D.G. Stuart (Eds.), Neural Control of Locomotion. Plenum Press, New York, 1977: 489--517. Kearney, R.E. Evaluation of the Wiener filter applied to evoked EMG potentials. Electroenceph. clin. Neurophysiol., 1979, 46 : 475--478. Kearney, R.E. and Chan, C.W.Y. A quantitative study of the electromyographic responses to sudden ankle displacements in man. Neurosci. Abstr., 1976, 2: 5544. Kearney, R.E. and Chan, C.Y.C. Reflex response of human arm muscles to cutaneous stimulation of the foot. Brain Res., 1979, 170: 214--217. Marsden, C.D., Merton, P.D., and Morton, H.B., Anticipatory postural responses in the human subject. J. Physiol. (Lond.), 1978, 259: 47--48P. Miller, S., Reitsma, D.J. and Van der Mech,, F.G.A. Functional organization of long ascending propriospinal pathways linking lumbro-sacral and cervical segments in the cat. Brain Res., 1973, 62: 169-188. Miller, S., van der Burg, J. and Van der Mech,, F.G.A. Coordination of hindlimb and forelimbs in different forms of locomotion in normal and decerebrate cats. Brain Res., 1975, 91: 217--237. Roberts, T.D.M. Neurophysiology of Postural Mechanisms, 2nd edition. Butterworths, London, 1979. Rossignol, S. and Melvill Jones, G. Audio-spinal influence in man studied by the H-reflex and its possible role in rhythmic movements synchronized to sound. Electroenceph. clin. Neurophysiol., 1976, 41: 83--92. Shimamura, M. and Akert, K. Peripheral nervous relations of propriospinal and spino-bulbo-spinal reflex systems. Jap. J. Physiol., 1965, 15: 638-647. Shimamura, M. and Livingston, R.B. Longitudinal conduction systems serving spinal and brain-stem coordination. J. Neurophysiol., 1963, 26: 258-272. Traub, M.M., Rothwell, J.C. and Marsden, C.D. Anticipatory postural reflexes in Parkinson's disease and other akinetic-rigid syndromes and in cerebellar ataxia. Brain, 1980, 103: 393--412. Walter, D.O. A posteriori 'Wiener filtering' of average evoked responses. Electroenceph. clin. Neurophysiol., 1969, Suppl. 27: 61--70.
65
INTERLIMB REFLEXES EVOKED IN HUMAN ARM MUSCLES BY ANKLE DISPLACEMENT R.E. KEARNEY and C.W.Y. CHAN
Biomedical Engineering Unit, and Aviation Medical Research Unit, McGill University, Montreal, Que. H3G 1 Y6 (Canada) (Accepted for publication: March 4, 1981 )
The control of posture, locomotion and voluntary movement requires the coordinated regulation of muscles acting at different joints and limbs. It has been suggested that interlimb reflexes may play an important role in this coordination (Halbertsma et al. 1977; Roberts 1979). For example, in the cat, reflex connections between hind limb and forelimb muscles are mediated by both spino-bulbospinal (Shimamura and Livingston 1963; Shimamura and Akert 1965) and propriospinal (Gernandt and Megirian 1961; Miller et al. 1973) pathways, and the later connections appear to be appropriate for the coordination of hind and forelimb movements during locomotion (Miller et al. 1975). We have previously demonstrated that reflex responses to cutaneous electrical stimulation of the foot can be detected in both leg and arm muscles of normal human subjects by averaging stimulus-related changes in tonic electromyographic (EMG) activity (Kearney and Chan 1979). We suggested that these responses could be involved in the coordination of a generalized response to noxious stimuli. If such interlimb reflexes are to play an important role in the coordination of movement, then a physiologically meaningful stimulus such as the displacement of a joint should be effective in evoking them. Furthermore, displacement of the joint in different directions should evoke significantly different responses. The present report describes a study which has demonstrated that ankle displacement evokes a substantial reflex response in human arm muscles whose nature depends upon the direction in which the ankle is displaced.
Methods Six subjects between the ages of 20 and 25 with no history of neuromuscular disease were examined. EMGs from gastrocnemius (G), tibialis anterior (TA), biceps brachii (BB) and triceps brachii (TB) were measured with surface electrodes, amplified, rectified and smoothed before recording. Subjects lay supine with the right foot attached to the pedal of a high performance electrohydraulic actuator by means of a custom-fitted plastic boot. The upper arm was externally rotated and abducted at 90 °, with the elbow flexed at 90 ° and held supine by a rigid frame. Subjects were instructed to maintain a steady, tonic contraction of the relevant muscle aided by a meter display of the smoothed and rectified EMG. Stimuli consisted of step displacements of ankle position (0.1 tad amplitude; 45 msec rise time) in either the dorsiflexing or plantarflexing direction. Stimuli were only applied when the tonic EMG activity was within 15% of the desired level. A PDP-11/40 computer sampled the EMG at 2000 Hz for 70 msec prior to, and 170 msec after, the stimulus. An on-line data analysis program computed the ensemble average response and Wiener-filtered it (Walter 1969; Kearney 1979). Stimuli were applied until an acceptable signal-to-noise ratio was obtained, as judged by comparing the pre- and poststimulus portions of the average. Anywhere from 50 to 500 responses were required depending upon the signal-to-noise ratio of the evoked EMG. The mean value and
0013-4649/81/0000--0000/$02.50 © Elsevier/North-Holland Scientific Publishers, Ltd.
66
R.E. KEARNEY, C.W.Y. CHAN
standard deviation of the prestimulus portion of the average were computed. Stretch-evoked responses were only considered to be significant if their amplitude was greater than twice the standard deviation of the prestimulus portion of the record. Latencies of responses were measured from the onset of displacement to the time at which the magnitude of the response exceeded twice the prestimulus standard deviation. Response magnitudes were expressed as the percentage modulation of the mean prestimulus activity. Means and standard errors (S.E.) presented below are group values computed from data from all 6 subjects.
Results
Responses to dorsiflexing displacements A typical set of responses evoked by dorsi-
A
flexing displacements of the ankle is illustrated in Fig. 1A. The response of G, which was stretched by the displacement, was dominated by a period of intense excitation (group mean modulation 800 + S.E. 130%) lasting from about 40 to 80 msec (mean 37 ± S.E. 2 msec to 84 ± S.E. 5 msec), followed by a period of decreased activity. This pattern of response is in agreement with that previously described (Kearney and Chan 1976; Chan et al. 1979b; Gottlieb and Agarwal 1979). Dorsiflexion of the ankle did not evoke a significant response in the elbow flexor, BB. Small, long latency responses were occasionally observed but these were not significantly different from control responses. In contrast, a consistent response was observed in TB, an elbow extensor, in all 6 subjects. A short latency (peak at 55 ± S.E. 3 msec) excitation was observed which although small (mean modulation 25 ± S.E. 4%) was significantly
i
I
G
i
I
j!
I
~
400 % i
TA
I
',
~
_ _ _
_
~
_,~L_..,
200
...........~:_ _,____~_+-:~
I
BB
I ,
TB
~[
5o~
j
I
I
I
,
I I
I
POSITION I~ / //~ [. /
......
1
.1 rad
POSITION I 1
I I
20
ms
Fig. 1. Patterns of EMG potentials evoked in ipsilateral arm and leg muscles by ankle displacements. Baseline for each record corresponds to mean level of prestimulus EMG activity. Positive EMG rectification was used so that an upward deflection of the trace corresponds to increased EMG activity. EMG scales are given as a percentage of the tonic activity. Bottom trace is ankle position with dorsiflexion denoted by an upward deflection. The dashed line in this and all subsequent figures indicates the start of the ankle displacement. A: responses evoked by dorsiflexing displacements in gastrocnemius (G), biceps brachii (BB), and triceps brachii (TB}. B: responses evoked by plantarflexing displacements in tibialis anterior (TA), biceps brachii (BB), and triceps brachii {TB).
RESPONSE OF HUMAN ARM MUSCLES TO ANKLE DISPLACEMENT
greater than the prestimulus level in all cases, according to the criterion described in Methods. This was followed by a decrease in activity of somewhat greater magnitude (mean modulation 38-+ S.E. 7%) with an onset at 72.5 + S.E. 8 msec and a peak at 90.2 + S.E. 4 msec. In our previous study (Kearney and Chan 1979) we found t h a t cutaneous electrical stimulation of the sole of the f o o t evoked a similar decrease in TB activity at about this latency.
Responses to plantarflexing displacements
67
Consistency of response pattern The pattern of response evoked in the arm muscles by ankle displacement was very consistent both from subject to subject and from day to day in the same subject. This is illustrated in Fig. 2 which shows the TB response to ankle plantarflexion in 4 subjects, two of whom (top two traces) were studied on two different days. Note the similarity o f the responses. In general, as noted above, there was significantly more variation in the amplitudes of the responses than in their latencies.
Sensory origin of the responses Plantarflexing displacements of the ankle evoked a somewhat different pattern of response as illustrated in Fig. l B . The response of TA, which was stretched in this case, consisted of a prolonged, biphasic excitation (mean modulation 418 + S.E. 111%) lasting from about 40 to more than 100 msec (mean 42 -+ S.E. 1 msec to 110 + S.E. 2 msec), followed by a period of decreased activity. Again, this pattern of response is consistent with what has been previously reported (Kearney and Chan 1976; Chan et al. 1979a; Gottlieb and Agarwal 1979)• Plantarflexing displacements of the ankle evoked responses consisting o f an early excitation followed by a period of decreased activity in both TB and BB. In TB, the early excitation started at 54 -+ S.E. 2 msec, peaked at 70 -+ S.E. 3 msec and modulated the tonic activity an average of 74 +- S.E. 18%. The subsequent decrease in activity started at 89 + S.E. 3 msec, peaked at 102 + 2 msec and modulated the tonic activity an average of 48 -+ S.E. 11%. This pattern is in contrast to t h a t evoked by dorsiflexion in which the early excitation was always o f much smaller magnitude than the subsequent decrease in activity. The pattern o f response in BB was similar to that of TB although always of much smaller magnitude. The early excitation modulated the tonic activity a mean of only 30 -+ S.E. 8%, although the latency was similar (onset of 51 -+ 2 msec and a peak at 67 -+ S.E. 3 msec).
There was some noise and vibration associated with the stimulus which might in itself evoke a response (e.g., Rossignol and Melvill Jones 1976). Control experiments were therefore run with the subject in the same position as t h a t used in the displacement experiments, but with the foot removed from the actuator
TB
I
I I
A^AAAA
A
A ..... I I I
20
J
,~
V , r
7~
,
A
,
ms
Fig. 2. Repeatability of triceps brachii response to plantarflexing displacements of the ankle. Responses observed in 4 different subjects are shown. Superimposed responses for subjects T B and SF were obtained on different days. All responses have been normalized to have the same peak-to-peak amplitude.
68
R.E. K E A R N E Y , C.W.Y. CHAN
100
I
Fig. 3. A: typical mental situation pedal t o p r e v e n t t i o n s as in A b u t p e d a l to p e r m i t it
r e s p o n s e o f s u b j e c t in t h e experiwith the foot removed from the a n k l e d i s p l a c e m e n t . B: same condit h e s u b j e c t ' s f o o t was placed o n t h e to be displaced.
pedal. The responses observed in these cases were always much smaller than those evoked by ankle displacement and had a longer latency (greater than 100 msec). This is illusGrated in Fig. 3 which shows responses evoked in TB by plantarflexing displacements of the ankle (Fig. 3B) and by a control stimulus in which everything was the same except that the subject's foot was removed from the actuator pedal (Fig. 3A). Clearly, the control response is insignificant when compared to that evoked by ankle displacement. Thus it seems certain that the responses observed in the arm muscles in these experiments were evoked by ankle displacement rather than by vibration or sound. These displacements must have stimulated both proprioceptors and exteroceptors whose relative contributions to the response remain to be determined. The initial component of the TB response had a latency (mean 54 msec) only
12--17 msec longer than those of TA {mean 42 msec) and G (mean 37 msec). In view of their short latencies, the initial components of leg muscle responses are almost certainly mediated by muscle spindle afferents (Kearney and Chan 1976; Gottlieb and Agarwal 1979). If central conduction time is allowed for, it seems likely that at least the initial component of the TB response is also mediated by muscle afferents. A contribution from cutaneous afferents, in particular from pressure receptors which must have been excited by our stimuli, cannot be ruled out. However, due to the lower conduction velocity of these afferents, their effects would be expected to occur at longer latencies than those of muscle afferents. This idea is supported by our previous finding (Kearney and Chan 1979) that cutaneous electrical stimulation never evoked responses in arm muscles at latencies less than 70 msec.
Discussion These findings demonstrate that displacements of the ankle evoked significant reflex responses in arm muscles. The relatively small ankle displacements {less than 6 ° ) used in our study still modulated the tonic activity of TB by as much as 80%. The possible significance of these responses is apparent when it is noted that in some cases their magnitude was as much as 20% of that evoked in the muscle that was actually stretched. Furthermore, the stretch evoked responses occurred in the arm muscles at very short latencies. For example, the early excitatory component of the TB response had a mean latency of 54 msec only 12--17 msec greater than those in TA (mean 42 msec) and G (mean 37 msec). In contrast, in our previous study (Kearney and Chan 1979), interlimb responses evoked by cutaneous electrical stimulation, at levels more than 3 times the sensory threshold, rarely modulated the tonic activity by more than 15--20% and never had a latency of less than 70 msec. It is difficult to compare the relative magni-
RESPONSE OF HUMAN ARM MUSCLES TO ANKLE DISPLACEMENT tudes of the displacement stimuli used in the present experiments to those of the electrical stimuli used previously. Nevertheless, it is notable that small displacements evoked interlimb responses which were larger and of shorter latency than those evoked b y electrical stimuli o f the highest intensity subjects would tolerate. This would seem to indicate that the interlimb reflex connections o f proprioceptive afferents are stronger and more direct than those of cutaneous origin. If this is so, then interlimb connections o f proprioceptive afferents may play an important role in the coordination of movement. The interlimb reflexes described in this paper should be compared to the 'anticipatory postural reflexes' studied b y Marsden and co-workers (Marsden et al. 1978; Traub et al. 1980). The original observation was that the unexpected perturbation of an arm while executing a simple manual tracking task evoked reflex responses in muscles of the opposite arm, trunk and leg engaged in supporting the b o d y (Marsden et al. 1978). Subsequently, a more detailed study was made of reflexes evoked in the ankle muscles o f normal and pathological subjects b y displacements of the arm (Traub et al. 1980). The essential features of these descending reflexes appear to be the same as those of the ascending reflexes we have studied. Thus, Marsden et al. (1978) concluded that the reflexes originated in the arm that was pulled and that the afferent input was not cutaneous since the response was not modified b y anesthesia o f the hand and wrist. We have reached similar conclusions a b o u t the reflexes evoked in the arm muscles b y ankle displacements based on the results of the control experiments and the finding that ankle displacement is a more effective stimulus than cutaneous electrical stimulation. Furthermore, our finding that the activity in TB is predominantly increased b y plantarflexing displacements of the ankle and predominantly decreased b y dorsiflexing displacements is consistent with the finding that the polarity o f the triceps surae response depends u p o n the direction is which the arm
69
is pulled (Traub et al. 1980). Thus it appears that the ascending reflexes linking arm muscle activity to ankle displacement are organized in much the same way as the descending reflexes linking the activity of ankle muscles to arm displacement. One difference between the two studies is in the latencies of the responses. The 'anticipatory postural reflexes' evoked in triceps surae b y arm displacement had a mean latency of 83 msec (Traub et al. 1980). In contrast, we found that ankle displacement evoked reflex responses in TB at 55 msec. These different latencies might indicate that the descending and ascending reflexes utilize different pathways. Traub et al. {1980) suggested that a transcortical pathway may be involved in generating the descending reflexes. In contrast the short latency of the responses we have observed makes a propriospinal pathway seem likely. However, the different latencies might also arise from differences in experimental methods. We used rapid changes in ankle position in our experiments which would be expected to evoke a synchronous afferent volley. Little temporal summation would therefore be required to evoke a reflex response and its latency should be determined primarily b y conduction delay. Our finding that the reflex responses o f TA and G occurred at latencies close to those of their tendon jerks supports this argument. In contrast, in the experiments of Traub et al. (1980) relatively slow displacements were used which might give rise to an asynchronous afferent volley. Consequently, some temporal summation would be required before a reflex response was evoked resulting in an increased latency. It remains to consider the functional significance of these responses. Marsden and co-workers considered the responses t h e y observed to be postural in nature since t h e y occurred in postural muscles, were in the appropriate direction to compensate for the postural disturbances evoked b y the stimuli, and were generally reduced in parkinsonian patients who displayed postural instability
70 (Marsden et al. 1978; Traub et al. 1980). The functional role of the responses observed in our experiments is less clear. The predomiant response occurred in TB and was dominated by excitation for plantarflexing displacements and b y a decrease in activity for dorsiflexing displacements. This pattern had no obvious role in postural stabilization but then postural stabilization was not required in our paradigm. Alternatively, interlimb reflexes could be involved in the coordination of movement. For example, the excitation of TB (an elbow extensor) when TA (an ankle flexor) is stretched, and the predominant decrease in TB activity which occurs when G (an ankle extensor) is stretched, appears to be appropriate for the coordination of reciprocal arm and leg movements during locomotion. However, the pattern of interlimb reflexes evoked in other muscles will have to be determined before any such functional hypothesis can be substantiated. Furthermore, it should be realized that, in the cat, interlimb reflexes have been shown to vary with both static limb position (Grillner and Rossignol 1978), and the phase of the step cycle during locomotion (Halbertsma et al. 1977). Such effects will have to be investigated in man before the functional significance of human interlimb reflexes can be properly understood.
Summary Interlimb reflexes evoked by ankle displacements were studied in arm muscles of 6 normal subjects. EMGs from grastrocnemius (G), tibialis anterior (TA), biceps brachii (BB), and triceps brachii (TB) were amplified, rectified and low-pass filtered before recording. Averaging and Wiener filtering were used to detect changes in tonic EMG activity evoked b y dorsiflexing or plantarflexing displacements of the ankle. A consistent pattern of response was observed in all subjects. In the leg muscles, the responses to stretch were consistent with previous reports. In the arm muscles, the
R.E. KEARNEY, C.W.Y. CHAN response of TB was dominant. Dorsiflexing displacements of the ankle evoked a small excitation followed by a more marked decrease in TB activity but had no effect on BB. In contrast, plantarflexing displacements of the ankle resulted in a large, early period of excitation followed by a decreased level of activity in TB. A similar but smaller pattern of activity was observed in BB. It is notable that the TB responses to displacement were sizable, often modulating the tonic EMG activity by as much as 80%. Interlimb reflexes evoked b y ankle displacement were larger and of shorter latency than those evoked by cutaneous electrical stimulation of the foot reported previously. This suggests that proprioceptive afferents may have stronger and more direct interlimb reflex connections than cutaneous afferents and may therefore play an important role in the coordination of movement.
R6sume Rdflexes inter-segmentaires dvoquds par des ddplacements de la cheville dans des muscles du bras humain Les r~flexes inter-segmentaires dvoquds par des d6placements de la cheville ont 6t6 ~tudids au niveau des muscles du bras chez 6 sujets normaux. Les dlectromyogrammes (EMG) des muscles gastrocn6miens (G), tibial antdrieur (TA), biceps brachial (BB), et triceps brachial (TB) ont 6t6 amplifids, rectifi6s et filtr~s avant l'enregistrement. Le moyennage et le filtrage Wiener ont 6t6 utilis6s afin de ddtecter les changements de l'activitd EMG tonique ~voqu~s par les d6placements en dorsiflexion ou en flexion plantaire de la cheville. Un 'pattern' constant de r6ponse a 6td observ6 chez tous les sujets. Dans les muscles de la jambe, les r6ponses ~ l'6tirement sont conformes aux 6tudes ant6rieures. Dans les muscles du bras, la r6ponse du TB est dominante. Les d6placements en dorsiflexion de la cheville ont 6voqu6 une diminution significa-
RESPONSE OF HUMAN ARM MUSCLES TO ANKLE DISPLACEMENT
tire de l'activit~ du TB mais n ' o n t pas eu d'effet sur le BB. Par contre, les d~placements de la cheville en flexion plantaire entrai'nent une p~riode initiale d'excitation suivie d'une diminution du niveau d'activit~ du TB. Un 'pattern' d'activit~ similaire mais r~duit s'observe au niveau du BB. I1 est ~ noter que les r~ponses du TB au d~placement sont mesurabies, modulant souvent l'activit~ EMG tonique d'environ 80%. Les r~flexes inter-segmentaires ~voqu~s par le d~placement de la cheville sont plus importants et de latence plus courte que ceux ~voqu~s par les stimulations cutan~es ~lectriques du pied pr~c~demment rapport~es. Ceci sugg~re que les aff~rences proprioceptives pourraient avoir des liens plus directs et plus ~troits avec les r~flexes intersegmentaires que les aff~rences cutan~es, et peuvent ainsi jouer un rSle important dans la coordination des mouvements. This work was supported by grants from the Canadian Medical Research Council and La Counseil de la Recherche en Sant~ du Qu4bec.
References Chan, C.W.Y., Kearney, R.E. and Melvill Jones, G. Tibialis anterior responses to sudden ankle displacements in normal and parkinsonian subjects. Brain Res., 1979a, 173: 303--314. Chan, C.W.Y., Melvill Jones, G., Kearney, R,E. and Watt, D.G.D. The 'late' electromyographic response to limb displacement in man. I. Evidence for supraspinal contribution. Electroenceph. clin. Neurophysiol., 1979b, 46: 173--181. Gernandt, B.E. and Megirian, D. Ascending propriospinal mechanisms. J. Neurophysiol., 1961, 24: 364--376. Gottlieb, G.L. and Agarwal, G.C. Response to sudden torques about the ankle: myotatic reflex. J. Neurophysiol., 1979, 46: 173--181. Grillner, S. and Rossignol, S. Contralateral reflex reversal controlled by limb position in the acute spinal cat injected with clonidine i.v. Brain Res., 1978, 144: 411--414.
71
Halbertsma, J.M., Miller, S. and Van der Mechd, F.G.S. Basic programs for the phasing of flexion and extension movements of the limbs during locomotion. In: R.M. Herman, S. Grillner, P.G.S. Stein and D.G. Stuart (Eds.), Neural Control of Locomotion. Plenum Press, New York, 1977: 489--517. Kearney, R.E. Evaluation of the Wiener filter applied to evoked EMG potentials. Electroenceph. clin. Neurophysiol., 1979, 46 : 475--478. Kearney, R.E. and Chan, C.W.Y. A quantitative study of the electromyographic responses to sudden ankle displacements in man. Neurosci. Abstr., 1976, 2: 5544. Kearney, R.E. and Chan, C.Y.C. Reflex response of human arm muscles to cutaneous stimulation of the foot. Brain Res., 1979, 170: 214--217. Marsden, C.D., Merton, P.D., and Morton, H.B., Anticipatory postural responses in the human subject. J. Physiol. (Lond.), 1978, 259: 47--48P. Miller, S., Reitsma, D.J. and Van der Mech,, F.G.A. Functional organization of long ascending propriospinal pathways linking lumbro-sacral and cervical segments in the cat. Brain Res., 1973, 62: 169-188. Miller, S., van der Burg, J. and Van der Mech,, F.G.A. Coordination of hindlimb and forelimbs in different forms of locomotion in normal and decerebrate cats. Brain Res., 1975, 91: 217--237. Roberts, T.D.M. Neurophysiology of Postural Mechanisms, 2nd edition. Butterworths, London, 1979. Rossignol, S. and Melvill Jones, G. Audio-spinal influence in man studied by the H-reflex and its possible role in rhythmic movements synchronized to sound. Electroenceph. clin. Neurophysiol., 1976, 41: 83--92. Shimamura, M. and Akert, K. Peripheral nervous relations of propriospinal and spino-bulbo-spinal reflex systems. Jap. J. Physiol., 1965, 15: 638-647. Shimamura, M. and Livingston, R.B. Longitudinal conduction systems serving spinal and brain-stem coordination. J. Neurophysiol., 1963, 26: 258-272. Traub, M.M., Rothwell, J.C. and Marsden, C.D. Anticipatory postural reflexes in Parkinson's disease and other akinetic-rigid syndromes and in cerebellar ataxia. Brain, 1980, 103: 393--412. Walter, D.O. A posteriori 'Wiener filtering' of average evoked responses. Electroenceph. clin. Neurophysiol., 1969, Suppl. 27: 61--70.