Brain Research, 536 (1990) 69-78 Elsevier
69
BRES 16134
Modulation of cutaneous cortical evoked potentials during isometric and isotonic contractions in the monkey W. Jiang 1, Y. L a m a r r e I and C.E. C h a p m a n ~'2 1Centre de Recherche en Sciences Neurologiques, D~partement de Physiologie and 2Ecole de Rdadaptation, Universit~ de Montreal, Station A, Montreal, Que. (Canada)
(Accepted 3 July 1990) Key words: Somatosensory evoked potential; Sensory gating; Somatosensory cortex; Isotonic; Isometric; Monkey
The effects of the direction of movement (flexion vs extension) and the nature of the motor task (isotonic vs isometric) on the modulation of sensory cortical evoked responses to cutaneous stimulation were investigated in one monkey. Sensory responses were assessed by measuring the magnitude of the short latency component of air puff-evoked potentials recorded intracortically in the arm representation of areas 3b and 1 in the primary somatoseusory cortex. At most recording sites, it was found that the amplitude of the air puff-evoked potential was decreased in a non-specific manner by motor activity. Neither the timing nor the depth of the modulation were found to vary with either the direction or the type of contraction. The effects were widespread since inputs from practically the entire forelimb (hairy skin) were diminished during the motor tasks. These results thus show that the modulation was more closely linked to the central motor output than to the peripheral input generated by muscle force and/or limb displacement. It is suggested that signals originating from central motor structures, acting in a feedforward manner, play a major role in 'gating' cutaneous inputs during movement. It is further suggested that the centrally mediated effects are exerted via a final common pathway upon which the 'gating' signals converge. INTRODUCTION A n u m b e r of studies have demonstrated that the amplitude of somatic sensory evoked potentials is reduced prior to and during voluntary movement and that this modulation is present as early as the level of the first relay in the lemniscal system, the dorsal column nuclei 6' 8,n.12. We have recently investigated the extent and time course of the movement-related modulation occurring at rostral levels of the lemniscal system 4. These experiments demonstrated that air puff-evoked potentials recorded from 3 levels of the neuraxis - - the medial lemniscus, the sensory thalamus (VPLc, ventral posterior lateral nucleus, caudal division) and the primary somatosensory cortex - - are reduced prior to and during voluntary elbow flexion in the monkey. The time course of the modulation was similar at all 3 levels, but the depth of modulation during m o v e m e n t was most pronounced at the cortical level. A t the level of the dorsal column nuclei, the effects appear to be mainly central in origin because passive movements, attempting to mimic the feedback associated with movement, have little or no effect 4'6"12. In contrast, the additional 'gating' of sensory transmission at higher levels, occurring only during movement, appears to be peripheral in origin since
passive movements elicited a similar decrease in the amplitude of the evoked response in most experiments. However, a number of questions remain unanswered. First, how specific is the movement-related decrease in cutaneous transmission to primary somatosensory cortex with regard to the location of the receptive field on the moving limb and the direction of m o v e m e n t ? In our previous study, it was suggested that inputs from the operant forearm were diminished in a non-specific manner since elbow flexion produced a similar decrease in the evoked responses both from the radial side of the forearm (in the path of movement) and from the ulnar side. However, these recordings were taken from different recording sites in separate experiments. A n y subtle differences in either the timing or the depth of modulation might not have been detected. Furthermore, the monkey's performance can vary from day to day, and it is known that factors such as the speed of m o v e m e n P '12 influence the depth of the modulation. Thus, the question of specificity needs to be addressed using recordings from one single site, in a single experimental session. In the present series of experiments this issue was addressed by examining the influence that different directions of movement have on the modulation of input from a single receptive field site on the moving arm. It was expected
Correspondence: Y. Lamarre, Centre de Recherche en Sciences Neurologiques, Facult6 de M6decine, Universit6 de Montr6al, P.O. Box 6128, Station A, Montr6al, Qu6. H3C 3J7, Canada.
0006-8993/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
70 that if the m o d u l a t i o n is really non-specific, then changing the relationship between the area of skin and the direction of m o v e m e n t (flexion or extension of the elbow in the present study) would have no influence on either the timing or the d e p t h of the m o v e m e n t - r e l a t e d modulation of cutaneous e v o k e d potentials recorded from primary s o m a t o s e n s o r y cortex. Second, how i m p o r t a n t is peripheral feedback to the m o d u l a t i o n of sensory transmission at higher levels of the lemniscai system during m o v e m e n t ? Passive limb movements (see above) have been r e p o r t e d to be practically as effective as active m o v e m e n t s in reducing the amplitude of sensory cortical potentials e v o k e d by stimulation of either the p e r i p h e r y 1"4A3"19'29, or central structures (medial lemniscus and VPLc) 4. Such observations might lead one to conclude that much of the decrease during m o v e m e n t is due to reafferent signals from the moving limb. In our previous study 4, however, the depth of m o d u l a t i o n at the thalamic or the cortical level p r o d u c e d by passive m o v e m e n t s was sometimes much less than that p r o d u c e d by active movements. F u r t h e r m o r e , and as p o i n t e d out above, central signals also contribute to the decrease, being entirely responsible for the decrease occurring at the level of the medial lemniscus 4'6"12. The relative contribution of p e r i p h e r a l feedback to the modulation at higher levels may, in fact, have been overest i m a t e d in previous studies since passive movements might actually elicit m o r e intense feedback than do active m o v e m e n t s . Thus, Prochazka 27 has observed that passive m a n i p u l a t i o n of an awake animal's limb results in greatly e n h a n c e d dynamic fusimotor action, and so enhanced spindle feedback, as c o m p a r e d to during voluntary s t e r e o t y p e d m o v e m e n t s , an effect which can be explained by the arousing effects of handling an animal. H e suggested that this observation could explain Soso and Fetz's 3° finding that units in p r i m a r y somatosensory cortex receiving input from proprioceptive afferents are m o r e vigorously activated by passive than active movements. It therefore appears that the results o b t a i n e d with passive m o v e m e n t s of the limb should be cautiously i n t e r p r e t e d as this m a y not provide a good m e t h o d to d e t e r m i n e the exact role of peripheral feedback in the m o d u l a t i o n of cutaneous transmission. M o v e m e n t - r e l a t e d f e e d b a c k can, however, be modified in o t h e r ways. The a p p r o a c h used in the present study was to c o m p a r e the m o d u l a t i o n during two m o t o r tasks, one in which limb displacement occurs (isotonic contraction) and the o t h e r in which limb displacement is mechanically blocked (isometric contraction). F e e d b a c k from the limb should be considerably different in the two situations. We anticipated that if peripheral f e e d b a c k contributes in any substantial way to the modulation during m o v e m e n t , then a noticeable difference would be
observed during isometric and isotonic contractions. A l t h o u g h the n e t input from the limb may or may not be different in the two types of contraction, the origin and balance of inputs from skin and from flexor and extensor p r o p r i o c e p t o r s is likely to differ considerably in these two m o t o r tasks. The present experiments were thus designed to compare the time course and the extent of the modulation of air puff-evoked cortical potentials during 4 m o t o r tasks: isotonic m o v e m e n t s in flexion and extension and isometric contractions in flexion and extension. We used the e v o k e d potential m e t h o d since only this technique provides the necessary stability of recording conditions over the p r o l o n g e d p e r i o d of time required to complete the testing under the above conditions. In a g r e e m e n t with our previous suggestion 4, the results indicated that the m o v e m e n t - r e l a t e d m o d u l a t i o n of cortical e v o k e d potentials is indeed non-specific with respect to the direction of m o v e m e n t . The results also d e m o n s t r a t e d that the m o d u l a t i o n appears to d e p e n d m o r e u p o n the central m o t o r output than the reafferent input, suggesting that much of the modulation at the cortical level during m o v e m e n t is central in origin. A preliminary account of the results has been p r e s e n t e d 17.
MATERIALS AND METHODS One monkey, Macaca mulata (weight 3.6 kg), was first trained to perform an isotonic motor task and then trained to perform an isometric motor task. The same monkey was also used in two other studies14'15. The pronated operant forearm and hand were strapped onto a manipulandum (see Fig. 1 in ref. 4). The isotonic task has been fully described elsewhere2°. Briefly, the animal was conditioned to perform rapid elbow movements, flexion or extension, in response to, respectively, a tone (800 Hz) or a light (3 x 3 array of 9 light-emitting diodes). Movements had to reach a predetermined amplitude (25°) within 500 ms after the GO cue (Fig. 1). Successful trials were rewarded with a drop of fruit juice. There were no mechanical stops and no reference points for starting or ending the movements. After this training was completed, the isometric task was introduced. Displacement was eliminated by blocking the manipulandum at a position of 80 __ 10° of elbow flexion. The animal made contractions against the wrist and forearm supports in the direction of flexion or extension (same GO cues as above). Contractions had to reach a predetermined level of force (3 N) within 500 ms after the onset of the GO cue. The animal mastered the latter task within one training session (about 300 trials)i Angular displacement at the elbow was measured by a potentiometer, while the force exerted during the isometric task was measured by a strain gauge located at the level of the wrist. Following training, the monkey was anesthetized with pentobarbital and, under aseptic conditions, a recording chamber which permitted access to the primary somatosensory cortex was implanted over the cerebral hemisphere contralateral to the trained arm22. Teflon-coated, multi-stranded stainless steel wires were chronically implanted into selected arm and shoulder muscles for electromyographic (EMG) recordings. Additional recordings were taken with insulated copper wires inserted percutaneously into the muscle of interest. The muscles studied had actions around the shoulder (pectoralis major, anterior and posterior deltoid), elbow (long head of biceps brachii, lateral head of triceps brachii,
71 brachialis, brachioradialis) or wrist (flexor carpi radialis, extensor carpi ulnaris).
ment or the force signal, were digitized on-line (2 kHz). For experiments in which only E M G activity was recorded, 4 muscles were recorded simultaneously (see Results); these signals were digitized at a frequency of 200 Hz. In each case, reaction time (onset of displacement or force change for the isotonic and isometric conditions, respectively) was detected by a simple algorithm that measured the occurrence of the first of 50 ms of consecutive samples that changed in the same direction. The speed of movement, or the rate of the force change, in the isotonic and isometric tasks, respectively, was computed by 3-point numerical differentiation and appropriate digital filtering of the position and force traces, respectively. For all recordings, the amplitude of the first component of the evoked potential (onset-to-peak, see Fig. 1) was measured from individual trials. The results were expressed as a percentage of the value of the corresponding control trials and plotted as a function of the delay of-the stimulus with respect to: (1) the onset of the GO cue; (2) the onset of the displacement or the force change (Figs. 3 and 4); and (3) the onset of E M G activity in the agonist (Fig. 5C,D) (onset measured by hand from individual trials). In the latter two cases, trials were regrouped into 20 ms bins relative to the time in question. The data were analyzed using two-tailed paired t-tests (level of significance, P < 0.01). The onset of the modulation, relative to the reaction time (RT), was determined as the first of at least two consecutive delays at which a significant effect was seen. Student's t-test was used to test for statistically different effects of either movement direction or the nature of the motor task (level of significance, P < 0.01).
Data recording and analysis The methods for data recording and analysis have been described elsewhere 4. Briefly, intracortical somatosensory evoked potentials were recorded from the arm representation of the primary somatosensory cortex contralateral to the operant arm with glass-coated, tungsten microelectrodes (impedance 0.2-0.8 MI2). Multi-unit recordings were first taken to identify and map the peripheral receptive field of the cortical recording site using a hand-held probe. All recording sites used in this study had a receptive field on the hairy skin of the upper arm, forearm or hand. Short latency evoked potentials were elicited by an air puff (duration, 20 ms) delivered into the centre of the receptive field. The air puff delivery, through a solenoid valve, was controlled by a laboratory mini-computer (PDP 11-23). The computer also controlled the task and the data acquisition. The tubing which delivered the air puff was mounted on the manipulandum so that a constant relationship between the stimulus and the limb was maintained throughout the recordings. During data acquisition, movement and control (no GO cue, animal at rest) trials were alternated. In each movement trial, the air puff was applied at a delay following the onset of the GO cue. The tested delays varied from 20 up to 360 ms in different experiments (range indicated by the horizontal arrows in Fig. 1), so that the cortical evoked response was evaluated during both the premovement period and the initial part of the muscle contraction. Control responses were recorded during the intertrial interval, at a time unpredictable to the monkey. For each delay tested, a block of 60 trials (30 movement, 30 control) was recorded, consisting of equal numbers of trials in flexion and extension (taken in two separate series along with their respective control trials). During the experiment, the task (isometric or isotonic) was switched after testing from 1-3 delays, thus controlling for any possible changes in the recording conditions during the session which usually lasted at least 1 h. Routinely, simultaneous recordings of cortical evoked potentials and EMGs of one elbow muscle (biceps for flexion, triceps for extension) were taken. The individual evoked potentials, EMGs (full-wave rectified and integrated), as well as the angular displace-
Histological methods Towards the end of the recording period, electrolytic lesions were made in selected electrode tracks. After the final recording session, the monkey was deeply anesthetized and perfused through the heart with buffered formal-saline solution. After removing the dura, known stereotaxic points on the cortical surface were marked with India ink and the whole brain was subsequently photographed. Electrode tracks were reconstructed from 30/~m sagittal sections stained with Cresyl violet. The criteria described by Powell and Mountcastle 25 and Jones et al. TM were used to distinguish between areas 3b and 1. In 18 experiments, the cortical recording sites were
MOVEMENT
AT REST
Cortical Evoked Potential Air puff Stimulus
] 50 I~V I I-I
"
Qocue I Position
Biceps EMG - -
I
I I I
I 17
=
1 ] 20 ° X
t,
I
.
.
.
.
T 1 00 IIV
.
i
I
=
t
1 0 0 ms
o n s e t of air puff
*
exp. no. 24
Fig. 1. Illustration of the experimental paradigm (isotonic task). A i r puffs were applied to the center of the receptive field (ulnar forearm in this case) either at a variable delay following the onset of the GO signal (left, movement trial; delay 207 ms), or with the:animal at rest (right, control trial). The evoked potential traces are averages of 38 trials, aligned on the onset of the air puff stimulus (negative up). The corresponding displacement traces and E M G were averaged either on the onset of displacement (left), to show where the air puffs were applied relative to the motor task, or on the onset of the trial (right). F, flexion; X, extension.
72 flexion (comparing B with A), and in the isotonic task for extension (comparing C with D). In the present study, however, the delays at which the air puff stimuli were applied did not exceed 200 ms after the onset of EMG (approximate limits of time of testing indicated by the broken vertical lines in A - D ) . As has been described previously4'23, the EMG activity of the prime movers in the task of isotonic elbow flexion and extension (biceps and triceps brachii) was correlated with peak velocity (Fig. 2E,G). The wrist muscles demonstrated a similar, but weaker, correlation (not shown). It should be noted that the range of peak velocities was similar for the two directions of movement in this monkey. This indicates that the two oppositely directed movements had similar kinematics since the amplitude of movement and peak velocity covary in this task 23. For the isometric task (Fig. 2F, H), EMG activity was also significantly correlated with the peak rate of force change (dF/dt). Again, the strongest correlations were observed for the elbow muscles. In addition, it was a fairly consistent observation in this animal that isometric flexions and extensions were not symmetrical: lower
clearly either in area 1 (n = 10) or area 3b (n = 8). In 3 experiments, the recording sites were located in the boundary zone between areas 1 and 3b.
RESULTS
Isotonic and isometric tasks: EMG and kinematics In 10 experimental sessions, the pattern of EMG activity in a number of muscles during the isometric and isotonic tasks was studied. The general pattern during isotonic flexion and extension is shown in Fig. 2A,C; the number of muscles active, and their pattern of activation, were similar to what has been reported previously for this task 24. Although muscles acting at the shoulder, elbow and wrist were all coactive, the elbow muscles showed the greatest activity and the best temporal relationship with the movement z1'24 and so were considered the prime movers. The same muscles were also active in isometric flexion (Fig. 2B) and extension (Fig. 2D) with the overall pattern of activity being similar to that observed under isotonic conditions. The duration of activity in the major agonists was slightly longer in the isometric task for
EXTENSION
FLEXION A
ISOTONIC
B ISOMETRIC
C
i
Position/Force
-~~i ~'/ir ~.,,
Biceps
. . . . . . . . . . . . . .
D
ISOTONIC
ISOMETRIC
1 50 °
,
!
1 20N
....+~'" ~i ~-' ,,,~ j..
. . . . .
Wrist Flexor Triceps Wrist Extensor Position/Force Anterior Deltoid Posterior Deltoid
.....
4
...... 500 ms
E w O
1201 r=0.64
ol :~ w
BICEPS
F 120
i n=80 60 0
r=0.58
TRICEPS
120 r1 ,0.51
n=82 • .. °°~" 60t'
.....,
100
G
BICEPS
250
400
PEAK VELOCITY (Deg/s)
0 0
.
120
PEAK dF/dt (N/s)
r =0.51 n=63
60
0 100
TRICEPS
Ln=80 60
50
H
0 1oo
250
400
PEAK VELOCITY (Oeg/s)
0
50
100
PEAK dF/dt (N/s)
Fig. 2. Comparison of EMG activity during isotonic and isometric contractions in both flexion and extension. A - D : averaged position (isotonic) and force (isometric) traces, and the averaged EMGs. The data of the upper 4 muscles (elbow and wrist) were collected simultaneously in one experiment (n = 82 trials); EMGs of the lower two muscles (shoulder) were recorded in another experiment (n = 60 trials) (corresponding averaged position and force traces shown just above). All records are aligned (vertical solid fines) on the onset of displacement (isotonic), or the onset of force change (isometric). The broken vertical lines illustrate the time window within which the cutaneous stimuli were ~ l i v e r e d . E - H : correlations between the EMG activity of the agonist muscles (flexion, E and F, extension, G and H) and the associated kinematics. The amplitude of the EMG was integrated over a time window extending from the onset of EMG to the time of peak velocity or peakdF/dt,
73 mean peak dF/dt values were generally observed in extension, and the total force change was less. The major difference observed between the isotonic and isometric tasks was the relative timing between the onset of detectable displacement or force change (RT) and the onset of the agonist muscle activity. The RT was consistently about 50-60 ms longer in the isotonic task than in the isometric task (P < 0.001). The EMG latencies following the GO cue, however, were identical for isotonic and isometric contractions. Thus, the difference in RT was entirely explained by the difference in the delay between the onset of EMG activity and the onset of the first detectable mechanical event. This delay was about 70 ms for isotonic contractions and 15 ms for isometric contractions. Data base
For 18 out of 21 experiments, recordings from a single site in the somatosensory cortex were obtained during the performance of the isotonic and the isometric tasks (both in flexion and in extension). In the remaining 3 experiments, cortical responses were tested in the two directions with only the isometric (n = 2) or the isotonic (n = 1) tasks. The peripheral receptive fields of the cortical recording sites were located on the shoulder (n = 1
A
experiment), the upper arm (n = 4), the forearm and the wrist (n = 11), and the dorsum of the hand (n = 5). A consistent and significant decrease in the magnitude of the short latency component of the cortical evoked potential was observed in all experiments (see e.g. Fig. 1) except the one in which the receptive field of the cortical recording site was located on the shoulder. No facilitatory effects were seen. Flexion versus extension
Fig. 3 illustrates the data obtained from one experiment. The receptive field was located on the dorsum of the wrist and the recording site was in area 3b. Despite the large variation in the amplitude of individual evoked potentials, which was typical in this type of experiment, one can readily observe a clear decrease in the amplitude of the air puff-evoked potential for movements in either direction. In the isotonic task (Fig. 3A), the decrease occurred at about 50 ms prior to onset of displacement (just after the onset of EMG activity). The evoked response was maximally depressed about 20-30 ms before the onset of movement. A slightly, but significantly (P < 0.01), greater decrease was observed in extension (mean response amplitude during movement, 18 + 13% of control) than in flexion (29 + 17%). In the
ISOTONIC
ISOMETRIC
175
175-
• Flexion
O ILl D I-_J
ca. .jo < E
150
150 0
125 100
8•
75
O > uJ
°
50
0 O
25
EMG
0
i
•
01
Oleo
•O
!.
°o8 o
o° ~ 000%08 ° o08 •@•Coo •
o•
0°08 • °o
100
0
2'0
onset of displacement
40
6'0
80 1;0 120 DELAY (ms)
0
00~ ,., 0 00~O'm 0 QWaI~ • 0 l - •6 o~
%
o~ o
50
0
O
°'il= ".I
o
o
o
I
I
O•
!!
o~o°8
25
ON O i
2'0
0
~_o~oo.
0
r
-120 -100 -80 -60 4 0
0
•
8o
o
_.dR o ,oOo O • '~ ~'~-(:~o I - - o •
! i
Io
". o°
I°° o
ON
I
• *
•
oo,
75
I •
0
---
I I
O o •
oO
i"
1 •
•
@
0 0
125
•~" ° o @e
8
• 0
oUJ
O°o o
;=8 '"'6 ~ o ,.a ~
o Extension
i
/
i
i
-120 -100 -80 -60 -40
m
-20
II I
0
onset of force
o l
•
-~. i
20. 40
=o, i
60
i
,
i
80 100 120 DELAY (ms)
Fig. 3. Comparison of the effects of the two directions of movement, flexion and extension, on the modulation of the air puff-evoked cortical potentials in both the isotonic (A) and the isometric (B) tasks in one experiment (area 3b). Results plotted as a function of the delay r e l a t i v e to the onset of the relevant mechanical event (A, displacement; B, force change). Each point represents the amplitude of a tingle evoked potential (as a % of its immediately adjacent control response). The mean control amplitudes (+ 1 S.D.) are shown on the right of each plot. The receptive field is shown above (shaded area).
74 isometric task (Fig. 3B), the decrease occurred at about the onset of the force change (i.e. just after the onset of E M G activity); the depth of modulation was the same for the two directions of contraction (24 + 19% in flexion; 28 + 26% in extension). Similar data were obtained from 16 additional experiments in which 'gating' was observed in both directions and in both tasks. In 13 of the 17 experiments, the depth of modulation was the same for the two directions, flexion and extension, in both the isotonic and the isometric tasks. In 4 experiments, the depth of modulation varied with direction: a significantly greater decrease was observed during isometric flexion as compared to isometric extension in 3 experiments; isotonic extension was associated with a greater modulation than isotonic flexion in two experiments (e.g. Fig. 3A). In only one of these experiments was there a directional difference for both the isotonic and the isometric tasks, and in this case direction did not have the same effect for the two tasks (as above). For the 4 experiments showing a difference in the depth of modulation with respect to direction, the mean difference was 24% (range, 11-39%). Their receptive fields were located either on the upper arm (n = 2) or the forearm (n = 2). Two of the recording sites were located in area 3b, one in area 1 and one in the area 3b/1 border zone.
Fig. 4 summarizes the results from the 17 experiments in which the evoked response was significantly reduced during movement, and in which both tasks were tested. In the isotonic task (Fig. 4A), the time course and the depth of modulation were very similar for the two directions of movement. In the isometric task (Fig. 4B), the depth of the modulation tended to be somewhat greater in flexion than in extension but this was not significant. One likely explanation for the occasional directional difference may lie in the kinematics of the associated movements. In the 17 experiments summarized in Fig. 4, the mean peak velocity in the isotonic task was almost identical for flexion and extension (284 and 322 deg/s, respectively), but in the isometric task the mean peak dF/dt for extension was only about one-half of that in flexion (43 vs 84 N/s). Further to this, in 12 out of 16 experiments subjected to linear regression analysis, we found a significant negative correlation (P < 0.02) between the amplitude of the air puff-evoked responses during the motor activity and the corresponding peak velocity and/or peak dF/dt values (correlation coefficients, r, ranging from - 0 . 2 to -0.7). Thus, faster movements or contractions resulted in a greater decrease in the amplitude of the air puff-evoked response during the motor task.
n=17
A w ~3 I--
-3
ISOTONIC
B
125
ISOMETRIC
:
125 7
100
100
5O
50
-
-- Flexion
o----~o
Extension
--
13.
~) o~ D. v
w "~
25
"'
o
25
EMG
£
r
l
-160 -120
oN l
-80
DELAY (ms)
[
-40
0
40
I
t
80
120
o
E(~NG
I
I
l
-120
-80
-40
t'
I
I
]
I
I
0
40
80
120
160
{
f
onset of displacement
onset of force
DELAY (ms)
Fig. 4. Summary of the effects of the two directions of movement, flexion and extension, on the modulation of air puff-evoked .potentials in the isotonic (A) and the isometric (B) tasks (n = 17 experiments). Data were subdivided into 20 ms bins (see Materials and Methods) and are aligned on the onset of the mechanical event (A, displacement; B, force change). Each point is an average of 4-15 experiments (+ 1 S.D.), as data from each experiment were not always available for a particular time bin (mean of 9 values for each point). This depended upon: (1) the number of tested delays; and (2) the distribution of the animal's RTs. The receptive fields of the 17 experiments are shown above.
75 EXTENSION
FLEXION
B
A
:
W 0
125-
125
O.
100
100
75.
75
50.
50.
25
25-
0
0
69
BRES 16134
Modulation of cutaneous cortical evoked potentials during isometric and isotonic contractions in the monkey W. Jiang 1, Y. L a m a r r e I and C.E. C h a p m a n ~'2 1Centre de Recherche en Sciences Neurologiques, D~partement de Physiologie and 2Ecole de Rdadaptation, Universit~ de Montreal, Station A, Montreal, Que. (Canada)
(Accepted 3 July 1990) Key words: Somatosensory evoked potential; Sensory gating; Somatosensory cortex; Isotonic; Isometric; Monkey
The effects of the direction of movement (flexion vs extension) and the nature of the motor task (isotonic vs isometric) on the modulation of sensory cortical evoked responses to cutaneous stimulation were investigated in one monkey. Sensory responses were assessed by measuring the magnitude of the short latency component of air puff-evoked potentials recorded intracortically in the arm representation of areas 3b and 1 in the primary somatoseusory cortex. At most recording sites, it was found that the amplitude of the air puff-evoked potential was decreased in a non-specific manner by motor activity. Neither the timing nor the depth of the modulation were found to vary with either the direction or the type of contraction. The effects were widespread since inputs from practically the entire forelimb (hairy skin) were diminished during the motor tasks. These results thus show that the modulation was more closely linked to the central motor output than to the peripheral input generated by muscle force and/or limb displacement. It is suggested that signals originating from central motor structures, acting in a feedforward manner, play a major role in 'gating' cutaneous inputs during movement. It is further suggested that the centrally mediated effects are exerted via a final common pathway upon which the 'gating' signals converge. INTRODUCTION A n u m b e r of studies have demonstrated that the amplitude of somatic sensory evoked potentials is reduced prior to and during voluntary movement and that this modulation is present as early as the level of the first relay in the lemniscal system, the dorsal column nuclei 6' 8,n.12. We have recently investigated the extent and time course of the movement-related modulation occurring at rostral levels of the lemniscal system 4. These experiments demonstrated that air puff-evoked potentials recorded from 3 levels of the neuraxis - - the medial lemniscus, the sensory thalamus (VPLc, ventral posterior lateral nucleus, caudal division) and the primary somatosensory cortex - - are reduced prior to and during voluntary elbow flexion in the monkey. The time course of the modulation was similar at all 3 levels, but the depth of modulation during m o v e m e n t was most pronounced at the cortical level. A t the level of the dorsal column nuclei, the effects appear to be mainly central in origin because passive movements, attempting to mimic the feedback associated with movement, have little or no effect 4'6"12. In contrast, the additional 'gating' of sensory transmission at higher levels, occurring only during movement, appears to be peripheral in origin since
passive movements elicited a similar decrease in the amplitude of the evoked response in most experiments. However, a number of questions remain unanswered. First, how specific is the movement-related decrease in cutaneous transmission to primary somatosensory cortex with regard to the location of the receptive field on the moving limb and the direction of m o v e m e n t ? In our previous study, it was suggested that inputs from the operant forearm were diminished in a non-specific manner since elbow flexion produced a similar decrease in the evoked responses both from the radial side of the forearm (in the path of movement) and from the ulnar side. However, these recordings were taken from different recording sites in separate experiments. A n y subtle differences in either the timing or the depth of modulation might not have been detected. Furthermore, the monkey's performance can vary from day to day, and it is known that factors such as the speed of m o v e m e n P '12 influence the depth of the modulation. Thus, the question of specificity needs to be addressed using recordings from one single site, in a single experimental session. In the present series of experiments this issue was addressed by examining the influence that different directions of movement have on the modulation of input from a single receptive field site on the moving arm. It was expected
Correspondence: Y. Lamarre, Centre de Recherche en Sciences Neurologiques, Facult6 de M6decine, Universit6 de Montr6al, P.O. Box 6128, Station A, Montr6al, Qu6. H3C 3J7, Canada.
0006-8993/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
70 that if the m o d u l a t i o n is really non-specific, then changing the relationship between the area of skin and the direction of m o v e m e n t (flexion or extension of the elbow in the present study) would have no influence on either the timing or the d e p t h of the m o v e m e n t - r e l a t e d modulation of cutaneous e v o k e d potentials recorded from primary s o m a t o s e n s o r y cortex. Second, how i m p o r t a n t is peripheral feedback to the m o d u l a t i o n of sensory transmission at higher levels of the lemniscai system during m o v e m e n t ? Passive limb movements (see above) have been r e p o r t e d to be practically as effective as active m o v e m e n t s in reducing the amplitude of sensory cortical potentials e v o k e d by stimulation of either the p e r i p h e r y 1"4A3"19'29, or central structures (medial lemniscus and VPLc) 4. Such observations might lead one to conclude that much of the decrease during m o v e m e n t is due to reafferent signals from the moving limb. In our previous study 4, however, the depth of m o d u l a t i o n at the thalamic or the cortical level p r o d u c e d by passive m o v e m e n t s was sometimes much less than that p r o d u c e d by active movements. F u r t h e r m o r e , and as p o i n t e d out above, central signals also contribute to the decrease, being entirely responsible for the decrease occurring at the level of the medial lemniscus 4'6"12. The relative contribution of p e r i p h e r a l feedback to the modulation at higher levels may, in fact, have been overest i m a t e d in previous studies since passive movements might actually elicit m o r e intense feedback than do active m o v e m e n t s . Thus, Prochazka 27 has observed that passive m a n i p u l a t i o n of an awake animal's limb results in greatly e n h a n c e d dynamic fusimotor action, and so enhanced spindle feedback, as c o m p a r e d to during voluntary s t e r e o t y p e d m o v e m e n t s , an effect which can be explained by the arousing effects of handling an animal. H e suggested that this observation could explain Soso and Fetz's 3° finding that units in p r i m a r y somatosensory cortex receiving input from proprioceptive afferents are m o r e vigorously activated by passive than active movements. It therefore appears that the results o b t a i n e d with passive m o v e m e n t s of the limb should be cautiously i n t e r p r e t e d as this m a y not provide a good m e t h o d to d e t e r m i n e the exact role of peripheral feedback in the m o d u l a t i o n of cutaneous transmission. M o v e m e n t - r e l a t e d f e e d b a c k can, however, be modified in o t h e r ways. The a p p r o a c h used in the present study was to c o m p a r e the m o d u l a t i o n during two m o t o r tasks, one in which limb displacement occurs (isotonic contraction) and the o t h e r in which limb displacement is mechanically blocked (isometric contraction). F e e d b a c k from the limb should be considerably different in the two situations. We anticipated that if peripheral f e e d b a c k contributes in any substantial way to the modulation during m o v e m e n t , then a noticeable difference would be
observed during isometric and isotonic contractions. A l t h o u g h the n e t input from the limb may or may not be different in the two types of contraction, the origin and balance of inputs from skin and from flexor and extensor p r o p r i o c e p t o r s is likely to differ considerably in these two m o t o r tasks. The present experiments were thus designed to compare the time course and the extent of the modulation of air puff-evoked cortical potentials during 4 m o t o r tasks: isotonic m o v e m e n t s in flexion and extension and isometric contractions in flexion and extension. We used the e v o k e d potential m e t h o d since only this technique provides the necessary stability of recording conditions over the p r o l o n g e d p e r i o d of time required to complete the testing under the above conditions. In a g r e e m e n t with our previous suggestion 4, the results indicated that the m o v e m e n t - r e l a t e d m o d u l a t i o n of cortical e v o k e d potentials is indeed non-specific with respect to the direction of m o v e m e n t . The results also d e m o n s t r a t e d that the m o d u l a t i o n appears to d e p e n d m o r e u p o n the central m o t o r output than the reafferent input, suggesting that much of the modulation at the cortical level during m o v e m e n t is central in origin. A preliminary account of the results has been p r e s e n t e d 17.
MATERIALS AND METHODS One monkey, Macaca mulata (weight 3.6 kg), was first trained to perform an isotonic motor task and then trained to perform an isometric motor task. The same monkey was also used in two other studies14'15. The pronated operant forearm and hand were strapped onto a manipulandum (see Fig. 1 in ref. 4). The isotonic task has been fully described elsewhere2°. Briefly, the animal was conditioned to perform rapid elbow movements, flexion or extension, in response to, respectively, a tone (800 Hz) or a light (3 x 3 array of 9 light-emitting diodes). Movements had to reach a predetermined amplitude (25°) within 500 ms after the GO cue (Fig. 1). Successful trials were rewarded with a drop of fruit juice. There were no mechanical stops and no reference points for starting or ending the movements. After this training was completed, the isometric task was introduced. Displacement was eliminated by blocking the manipulandum at a position of 80 __ 10° of elbow flexion. The animal made contractions against the wrist and forearm supports in the direction of flexion or extension (same GO cues as above). Contractions had to reach a predetermined level of force (3 N) within 500 ms after the onset of the GO cue. The animal mastered the latter task within one training session (about 300 trials)i Angular displacement at the elbow was measured by a potentiometer, while the force exerted during the isometric task was measured by a strain gauge located at the level of the wrist. Following training, the monkey was anesthetized with pentobarbital and, under aseptic conditions, a recording chamber which permitted access to the primary somatosensory cortex was implanted over the cerebral hemisphere contralateral to the trained arm22. Teflon-coated, multi-stranded stainless steel wires were chronically implanted into selected arm and shoulder muscles for electromyographic (EMG) recordings. Additional recordings were taken with insulated copper wires inserted percutaneously into the muscle of interest. The muscles studied had actions around the shoulder (pectoralis major, anterior and posterior deltoid), elbow (long head of biceps brachii, lateral head of triceps brachii,
71 brachialis, brachioradialis) or wrist (flexor carpi radialis, extensor carpi ulnaris).
ment or the force signal, were digitized on-line (2 kHz). For experiments in which only E M G activity was recorded, 4 muscles were recorded simultaneously (see Results); these signals were digitized at a frequency of 200 Hz. In each case, reaction time (onset of displacement or force change for the isotonic and isometric conditions, respectively) was detected by a simple algorithm that measured the occurrence of the first of 50 ms of consecutive samples that changed in the same direction. The speed of movement, or the rate of the force change, in the isotonic and isometric tasks, respectively, was computed by 3-point numerical differentiation and appropriate digital filtering of the position and force traces, respectively. For all recordings, the amplitude of the first component of the evoked potential (onset-to-peak, see Fig. 1) was measured from individual trials. The results were expressed as a percentage of the value of the corresponding control trials and plotted as a function of the delay of-the stimulus with respect to: (1) the onset of the GO cue; (2) the onset of the displacement or the force change (Figs. 3 and 4); and (3) the onset of E M G activity in the agonist (Fig. 5C,D) (onset measured by hand from individual trials). In the latter two cases, trials were regrouped into 20 ms bins relative to the time in question. The data were analyzed using two-tailed paired t-tests (level of significance, P < 0.01). The onset of the modulation, relative to the reaction time (RT), was determined as the first of at least two consecutive delays at which a significant effect was seen. Student's t-test was used to test for statistically different effects of either movement direction or the nature of the motor task (level of significance, P < 0.01).
Data recording and analysis The methods for data recording and analysis have been described elsewhere 4. Briefly, intracortical somatosensory evoked potentials were recorded from the arm representation of the primary somatosensory cortex contralateral to the operant arm with glass-coated, tungsten microelectrodes (impedance 0.2-0.8 MI2). Multi-unit recordings were first taken to identify and map the peripheral receptive field of the cortical recording site using a hand-held probe. All recording sites used in this study had a receptive field on the hairy skin of the upper arm, forearm or hand. Short latency evoked potentials were elicited by an air puff (duration, 20 ms) delivered into the centre of the receptive field. The air puff delivery, through a solenoid valve, was controlled by a laboratory mini-computer (PDP 11-23). The computer also controlled the task and the data acquisition. The tubing which delivered the air puff was mounted on the manipulandum so that a constant relationship between the stimulus and the limb was maintained throughout the recordings. During data acquisition, movement and control (no GO cue, animal at rest) trials were alternated. In each movement trial, the air puff was applied at a delay following the onset of the GO cue. The tested delays varied from 20 up to 360 ms in different experiments (range indicated by the horizontal arrows in Fig. 1), so that the cortical evoked response was evaluated during both the premovement period and the initial part of the muscle contraction. Control responses were recorded during the intertrial interval, at a time unpredictable to the monkey. For each delay tested, a block of 60 trials (30 movement, 30 control) was recorded, consisting of equal numbers of trials in flexion and extension (taken in two separate series along with their respective control trials). During the experiment, the task (isometric or isotonic) was switched after testing from 1-3 delays, thus controlling for any possible changes in the recording conditions during the session which usually lasted at least 1 h. Routinely, simultaneous recordings of cortical evoked potentials and EMGs of one elbow muscle (biceps for flexion, triceps for extension) were taken. The individual evoked potentials, EMGs (full-wave rectified and integrated), as well as the angular displace-
Histological methods Towards the end of the recording period, electrolytic lesions were made in selected electrode tracks. After the final recording session, the monkey was deeply anesthetized and perfused through the heart with buffered formal-saline solution. After removing the dura, known stereotaxic points on the cortical surface were marked with India ink and the whole brain was subsequently photographed. Electrode tracks were reconstructed from 30/~m sagittal sections stained with Cresyl violet. The criteria described by Powell and Mountcastle 25 and Jones et al. TM were used to distinguish between areas 3b and 1. In 18 experiments, the cortical recording sites were
MOVEMENT
AT REST
Cortical Evoked Potential Air puff Stimulus
] 50 I~V I I-I
"
Qocue I Position
Biceps EMG - -
I
I I I
I 17
=
1 ] 20 ° X
t,
I
.
.
.
.
T 1 00 IIV
.
i
I
=
t
1 0 0 ms
o n s e t of air puff
*
exp. no. 24
Fig. 1. Illustration of the experimental paradigm (isotonic task). A i r puffs were applied to the center of the receptive field (ulnar forearm in this case) either at a variable delay following the onset of the GO signal (left, movement trial; delay 207 ms), or with the:animal at rest (right, control trial). The evoked potential traces are averages of 38 trials, aligned on the onset of the air puff stimulus (negative up). The corresponding displacement traces and E M G were averaged either on the onset of displacement (left), to show where the air puffs were applied relative to the motor task, or on the onset of the trial (right). F, flexion; X, extension.
72 flexion (comparing B with A), and in the isotonic task for extension (comparing C with D). In the present study, however, the delays at which the air puff stimuli were applied did not exceed 200 ms after the onset of EMG (approximate limits of time of testing indicated by the broken vertical lines in A - D ) . As has been described previously4'23, the EMG activity of the prime movers in the task of isotonic elbow flexion and extension (biceps and triceps brachii) was correlated with peak velocity (Fig. 2E,G). The wrist muscles demonstrated a similar, but weaker, correlation (not shown). It should be noted that the range of peak velocities was similar for the two directions of movement in this monkey. This indicates that the two oppositely directed movements had similar kinematics since the amplitude of movement and peak velocity covary in this task 23. For the isometric task (Fig. 2F, H), EMG activity was also significantly correlated with the peak rate of force change (dF/dt). Again, the strongest correlations were observed for the elbow muscles. In addition, it was a fairly consistent observation in this animal that isometric flexions and extensions were not symmetrical: lower
clearly either in area 1 (n = 10) or area 3b (n = 8). In 3 experiments, the recording sites were located in the boundary zone between areas 1 and 3b.
RESULTS
Isotonic and isometric tasks: EMG and kinematics In 10 experimental sessions, the pattern of EMG activity in a number of muscles during the isometric and isotonic tasks was studied. The general pattern during isotonic flexion and extension is shown in Fig. 2A,C; the number of muscles active, and their pattern of activation, were similar to what has been reported previously for this task 24. Although muscles acting at the shoulder, elbow and wrist were all coactive, the elbow muscles showed the greatest activity and the best temporal relationship with the movement z1'24 and so were considered the prime movers. The same muscles were also active in isometric flexion (Fig. 2B) and extension (Fig. 2D) with the overall pattern of activity being similar to that observed under isotonic conditions. The duration of activity in the major agonists was slightly longer in the isometric task for
EXTENSION
FLEXION A
ISOTONIC
B ISOMETRIC
C
i
Position/Force
-~~i ~'/ir ~.,,
Biceps
. . . . . . . . . . . . . .
D
ISOTONIC
ISOMETRIC
1 50 °
,
!
1 20N
....+~'" ~i ~-' ,,,~ j..
. . . . .
Wrist Flexor Triceps Wrist Extensor Position/Force Anterior Deltoid Posterior Deltoid
.....
4
...... 500 ms
E w O
1201 r=0.64
ol :~ w
BICEPS
F 120
i n=80 60 0
r=0.58
TRICEPS
120 r1 ,0.51
n=82 • .. °°~" 60t'
.....,
100
G
BICEPS
250
400
PEAK VELOCITY (Deg/s)
0 0
.
120
PEAK dF/dt (N/s)
r =0.51 n=63
60
0 100
TRICEPS
Ln=80 60
50
H
0 1oo
250
400
PEAK VELOCITY (Oeg/s)
0
50
100
PEAK dF/dt (N/s)
Fig. 2. Comparison of EMG activity during isotonic and isometric contractions in both flexion and extension. A - D : averaged position (isotonic) and force (isometric) traces, and the averaged EMGs. The data of the upper 4 muscles (elbow and wrist) were collected simultaneously in one experiment (n = 82 trials); EMGs of the lower two muscles (shoulder) were recorded in another experiment (n = 60 trials) (corresponding averaged position and force traces shown just above). All records are aligned (vertical solid fines) on the onset of displacement (isotonic), or the onset of force change (isometric). The broken vertical lines illustrate the time window within which the cutaneous stimuli were ~ l i v e r e d . E - H : correlations between the EMG activity of the agonist muscles (flexion, E and F, extension, G and H) and the associated kinematics. The amplitude of the EMG was integrated over a time window extending from the onset of EMG to the time of peak velocity or peakdF/dt,
73 mean peak dF/dt values were generally observed in extension, and the total force change was less. The major difference observed between the isotonic and isometric tasks was the relative timing between the onset of detectable displacement or force change (RT) and the onset of the agonist muscle activity. The RT was consistently about 50-60 ms longer in the isotonic task than in the isometric task (P < 0.001). The EMG latencies following the GO cue, however, were identical for isotonic and isometric contractions. Thus, the difference in RT was entirely explained by the difference in the delay between the onset of EMG activity and the onset of the first detectable mechanical event. This delay was about 70 ms for isotonic contractions and 15 ms for isometric contractions. Data base
For 18 out of 21 experiments, recordings from a single site in the somatosensory cortex were obtained during the performance of the isotonic and the isometric tasks (both in flexion and in extension). In the remaining 3 experiments, cortical responses were tested in the two directions with only the isometric (n = 2) or the isotonic (n = 1) tasks. The peripheral receptive fields of the cortical recording sites were located on the shoulder (n = 1
A
experiment), the upper arm (n = 4), the forearm and the wrist (n = 11), and the dorsum of the hand (n = 5). A consistent and significant decrease in the magnitude of the short latency component of the cortical evoked potential was observed in all experiments (see e.g. Fig. 1) except the one in which the receptive field of the cortical recording site was located on the shoulder. No facilitatory effects were seen. Flexion versus extension
Fig. 3 illustrates the data obtained from one experiment. The receptive field was located on the dorsum of the wrist and the recording site was in area 3b. Despite the large variation in the amplitude of individual evoked potentials, which was typical in this type of experiment, one can readily observe a clear decrease in the amplitude of the air puff-evoked potential for movements in either direction. In the isotonic task (Fig. 3A), the decrease occurred at about 50 ms prior to onset of displacement (just after the onset of EMG activity). The evoked response was maximally depressed about 20-30 ms before the onset of movement. A slightly, but significantly (P < 0.01), greater decrease was observed in extension (mean response amplitude during movement, 18 + 13% of control) than in flexion (29 + 17%). In the
ISOTONIC
ISOMETRIC
175
175-
• Flexion
O ILl D I-_J
ca. .jo < E
150
150 0
125 100
8•
75
O > uJ
°
50
0 O
25
EMG
0
i
•
01
Oleo
•O
!.
°o8 o
o° ~ 000%08 ° o08 •@•Coo •
o•
0°08 • °o
100
0
2'0
onset of displacement
40
6'0
80 1;0 120 DELAY (ms)
0
00~ ,., 0 00~O'm 0 QWaI~ • 0 l - •6 o~
%
o~ o
50
0
O
°'il= ".I
o
o
o
I
I
O•
!!
o~o°8
25
ON O i
2'0
0
~_o~oo.
0
r
-120 -100 -80 -60 4 0
0
•
8o
o
_.dR o ,oOo O • '~ ~'~-(:~o I - - o •
! i
Io
". o°
I°° o
ON
I
• *
•
oo,
75
I •
0
---
I I
O o •
oO
i"
1 •
•
@
0 0
125
•~" ° o @e
8
• 0
oUJ
O°o o
;=8 '"'6 ~ o ,.a ~
o Extension
i
/
i
i
-120 -100 -80 -60 -40
m
-20
II I
0
onset of force
o l
•
-~. i
20. 40
=o, i
60
i
,
i
80 100 120 DELAY (ms)
Fig. 3. Comparison of the effects of the two directions of movement, flexion and extension, on the modulation of the air puff-evoked cortical potentials in both the isotonic (A) and the isometric (B) tasks in one experiment (area 3b). Results plotted as a function of the delay r e l a t i v e to the onset of the relevant mechanical event (A, displacement; B, force change). Each point represents the amplitude of a tingle evoked potential (as a % of its immediately adjacent control response). The mean control amplitudes (+ 1 S.D.) are shown on the right of each plot. The receptive field is shown above (shaded area).
74 isometric task (Fig. 3B), the decrease occurred at about the onset of the force change (i.e. just after the onset of E M G activity); the depth of modulation was the same for the two directions of contraction (24 + 19% in flexion; 28 + 26% in extension). Similar data were obtained from 16 additional experiments in which 'gating' was observed in both directions and in both tasks. In 13 of the 17 experiments, the depth of modulation was the same for the two directions, flexion and extension, in both the isotonic and the isometric tasks. In 4 experiments, the depth of modulation varied with direction: a significantly greater decrease was observed during isometric flexion as compared to isometric extension in 3 experiments; isotonic extension was associated with a greater modulation than isotonic flexion in two experiments (e.g. Fig. 3A). In only one of these experiments was there a directional difference for both the isotonic and the isometric tasks, and in this case direction did not have the same effect for the two tasks (as above). For the 4 experiments showing a difference in the depth of modulation with respect to direction, the mean difference was 24% (range, 11-39%). Their receptive fields were located either on the upper arm (n = 2) or the forearm (n = 2). Two of the recording sites were located in area 3b, one in area 1 and one in the area 3b/1 border zone.
Fig. 4 summarizes the results from the 17 experiments in which the evoked response was significantly reduced during movement, and in which both tasks were tested. In the isotonic task (Fig. 4A), the time course and the depth of modulation were very similar for the two directions of movement. In the isometric task (Fig. 4B), the depth of the modulation tended to be somewhat greater in flexion than in extension but this was not significant. One likely explanation for the occasional directional difference may lie in the kinematics of the associated movements. In the 17 experiments summarized in Fig. 4, the mean peak velocity in the isotonic task was almost identical for flexion and extension (284 and 322 deg/s, respectively), but in the isometric task the mean peak dF/dt for extension was only about one-half of that in flexion (43 vs 84 N/s). Further to this, in 12 out of 16 experiments subjected to linear regression analysis, we found a significant negative correlation (P < 0.02) between the amplitude of the air puff-evoked responses during the motor activity and the corresponding peak velocity and/or peak dF/dt values (correlation coefficients, r, ranging from - 0 . 2 to -0.7). Thus, faster movements or contractions resulted in a greater decrease in the amplitude of the air puff-evoked response during the motor task.
n=17
A w ~3 I--
-3
ISOTONIC
B
125
ISOMETRIC
:
125 7
100
100
5O
50
-
-- Flexion
o----~o
Extension
--
13.
~) o~ D. v
w "~
25
"'
o
25
EMG
£
r
l
-160 -120
oN l
-80
DELAY (ms)
[
-40
0
40
I
t
80
120
o
E(~NG
I
I
l
-120
-80
-40
t'
I
I
]
I
I
0
40
80
120
160
{
f
onset of displacement
onset of force
DELAY (ms)
Fig. 4. Summary of the effects of the two directions of movement, flexion and extension, on the modulation of air puff-evoked .potentials in the isotonic (A) and the isometric (B) tasks (n = 17 experiments). Data were subdivided into 20 ms bins (see Materials and Methods) and are aligned on the onset of the mechanical event (A, displacement; B, force change). Each point is an average of 4-15 experiments (+ 1 S.D.), as data from each experiment were not always available for a particular time bin (mean of 9 values for each point). This depended upon: (1) the number of tested delays; and (2) the distribution of the animal's RTs. The receptive fields of the 17 experiments are shown above.
75 EXTENSION
FLEXION
B
A
:
W 0
125-
125
O.
100
100
75.
75
50.
50.
25
25-
0
0
