Exp Brain Res (1991) 85:271-280

Experimental BrainResearch 9 Springer-Verlag1991

Functionally complex muscles of the cat hindlimb III. Differential activation within biceps femoris during postural perturbations C.M. Chanaud 1' * and J.M. Macpherson 2" **

1 Laboratory of Neural Control, National Institute of Neurological and Communicative Disorders and Stroke, National Institutes of Health, Bethesda, MD 20892, USA, 2 Department of Anatomy, Queen's University, Kingston, Ontario, K7L 3N6 Canada Received February 16, 1990 / Accepted January 22, 1991

Summary. The biceps femoris (BF) muscle is divided into

three neuromuscular compartments defined by the innervation patterns of the main nerve branches (English and Weeks 1987). The goals of this study were i) to determine how different regions of the biceps femoris muscle are activated in the intact cat during a broad range of limb movements evoked by perturbations of stance posture, and ii) to determine the relationship between the anatomical compartments of biceps femoris and the functional units as defined in this task. Cats were trained to stand on a moveable platform with each paw on a triaxial force plate. The animal's stance was perturbed by linear translation of the platform in each of sixteen different directions in the horizontal plane. E M G activity was recorded from eight sites across the width of the left biceps femoris muscle. During quiet stance only the anterior compartment was tonically active, presumably contributing to hip extensor torque in the maintenance of stance. During platform translation, evoked E M G activity was recorded from each electrode pair for a wide range of directions of perturbation; as direction changed progressively, the amplitude of evoked activity from any electrode pair increased to a maximum and then decreased. When the E M G amplitude was plotted in polar coordinates as a function of translation direction, the region of response formed a petal shaped area in the horizontal plane, termed the E M G tuning curve. The compartments of the BF muscle were not activated homogeneously. The tuning curve of the anterior BF compartment was similar to that of other hip extensors, and coincided with the region of postero-lateral force production by the hindlimb against the support. The tuning curve of the middle BF compartment was shifted in a Present addresses: * Paralyzed Veterans of America, Spinal Cord Research Foundation, 801 18th St. N.W., Washington, D.C. 20006, USA 9* R.S. Dow Neurological Sciences Institute, Good Samaritan Hospital, 1120 NW 20th Ave., Portland, OR 97209, USA Offprint requests to: G.E. Loeb, Biomedical Engineering, Abramsky Hall, Queen's University, Kingston, Ontario K7L 3N6, Canada

counterclockwise direction from that of the anterior compartment, but overlapped extensively with it; the middle BF tuning curve was similar to that of anterior gracilis. The tuning curve of the posterior biceps compartment was rotated further counterclockwise and overlapped very little with that of the middle BF compartment. The posterior BF was activated in a pattern similar to that of other knee flexors. The functional units of BF activation were not identical with the neuromuscular compartments defined by the main nerve branches. As direction of the perturbation changed, the region of BF that was activated moved progressively across the muscle. This progression of the active region was continuous across BFa and BFm, whereas there was a jump, or discontinuity at the border between BFm and BFp. Thus, differences in activation were observed not only across compartments, but also within compartments, and different regions of the BF muscle were activated independently during responses to postural perturbations. Key words: Biceps femoris - Muscles Posture

Stance Kinesiology - Electromyography - Differential - Cat

Introduction

Numerous mammalian skeletal muscles are divisible into distinct neuromuscular compartments according to the distribution of separate nerve branches (Letbetter 1974; English and Letbetter 1981 ; Richmond et al. 1985; Hammond 1987; Balice-Gordon and T h o m p s o n 1988). Observations o f this anatomical compartmentalization have led to the question of whether there is a corresponding functional compartmentalization. This question is important since a segregation of muscle regions according to function implies that there is likely a segregation of the motoneurone pool, characterized perhaps by different distributions of the descending and/or afferent inputs.

272 T h e biceps femoris (BF) muscle o f the cat h i n d l i m b is i n n e r v a t e d b y three, a n d o c c a s i o n a l l y f o u r m a i n b r a n c h e s f r o m the sciatic nerve which s u p p l y 3 - 4 c o m p a r t m e n t s with distinct b o u n d a r i e s , a n t e r i o r (BFa), m i d d l e ( B F m ) , a n d p o s t e r i o r ( B F p ) biceps (English a n d W e e k s 1987; C h a n a u d et al. 1991a). P r e v i o u s studies o f B F activity have e x a m i n e d three r h y t h m i c m o v e m e n t s , l o c o m o t i o n , p a w shake, a n d ear scratch ( E n g b e r g a n d L u n d b e r g 1969; E n g l i s h a n d W e e k s 1987; C h a n a u d et al. 1991b) a n d have d e m o n s t r a t e d some differential a c t i v a t i o n o f B F c o m p a r t m e n t s . I n general, the activity o f B F a a n d B F m was similar to t h a t o f o t h e r hip extensors, whereas the activity o f B F p was similar to t h a t o f o t h e r knee flexors. H o w e v e r , in all these m o v e m e n t s the r a n g e o f m o t i o n o f the h i n d l i m b was restricted p r i m a r i l y to the sagittal plane. T h e c u r r e n t s t u d y was designed to investigate B F activity d u r i n g m o v e m e n t s with a b r o a d r a n g e o f h i n d limb m o t i o n , i n c l u d i n g b o t h sagittal a n d transverse planes. W e wished to d e t e r m i n e w h e t h e r the p a t t e r n o f a c t i v a t i o n o f B F was segregated i n t o discrete regions a n d , if so, w h e t h e r the b o r d e r s o f these regions corres p o n d e d to the b o r d e r s o f the n e u r o m u s c u l a r c o m p a r t ments. T h e p o s t u r e p l a t f o r m p a r a d i g m was utilized to evoke p o s t u r a l responses to p e r t u r b a t i o n s o f quiet stance ( M a c p h e r s o n et al. 1987). Cats were exposed to t r a n s l a t i o n o f the s u p p o r t i n g p l a t f o r m in, each o f sixteen different directions i n the h o r i z o n t a l plane, while E M G activity was r e c o r d e d at eight sites across the w i d t h o f the B F muscle. A c t i v i t y was e v o k e d in different s u b r e g i o n s o f BF, d e p e n d i n g o n the d i r e c t i o n o f t r a n s l a t i o n . Differential a c t i v a t i o n was observed b o t h w i t h i n a n d across neuromuscular compartments. These findings have b e e n r e p o r t e d in a b s t r a c t f o r m ( C h a n a u d a n d M a c p h e r s o n 1987).

Methods

Trainin9 and electrode implantation Two female cats (3.6 kg and 4.9 kg) were trained to stand quietly and unrestrained on force plates mounted upon a hydraulicallydriven, moveable platform. Each plate measured the forces produced by one paw in three orthogonal directions, vertical (z-axis), longitudinal (y-axis), and lateral (x-axis), as shown in Fig. 1A. The cats were required to distribute their weight such that the difference in force between the left and right limbs of each girdle was less than 12% of total body weight. Detailed descriptions of the posture platform and the force plates are available elsewhere (Lywood et al. 1987; van Eyken et al. 1987). Once the animals were trained, EMG electrodes were implanted using standard, aseptic techniques. Cats were anaesthetized with Saffan (alphaxalone-alphadolone, 0.75 ml/kg). A single Silastic patch configured with eight electrode pairs was sutured to the fascial surface of the underside of the left biceps femoris muscle (Fig. 1B) (Hoffer and Loeb 1980). The Silastic patch provided a dielectric backing to reduce the chance of cross-talk from adjacent muscles (Loeb and Gans 1986). Each electrode consisted of a Teflon-coated stainless steel wire sewn into the patch material. The exposed contacts of each electrode pair were 3 mm in length and spaced 5 mm apart. The eight electrode pairs were spaced evenly across the width of the muscle in order to record the activity of each of the three compartments, BFa, BFm, and BFp. A separate ground electrode

was sutured to a nearby fat pad. All wire leads were passed subcutaneously to an exit site through the skin over the head and soldered to a 17-pin connector attached to the skull.

Recordin9 Once the cat had recovered from the surgery (5-7 days), recording sessions were begun. During periods of quiet stance, the cat was perturbed by a linear translation of the platform in each of sixteen directions in the horizontal plane. The direction of translation was specified in a polar coordinate system, with 0~ defined as a tailward translation, 90 ~ a leftward translation, 180~ a headward translation and 270~ rightward translation (Fig. 1A). The platform was translated at an average velocity of 15 cm/s. Since the cat is more stable in the sagittal plane than the transverse plane (Macpherson and Craig 1986), the amplitude of platform movement was scaled linearly from a minimum of 2 cm in the lateral direction to a maximum of 3 cm in the longitudinal direction. When a new direction of platform movement was chosen, the cat was given several practice trials to allow for adaptation of the response. During any given trial the animals were not cued or warned about the onset time of platform movement. Data obtained during 5 trials were stored on disk and used later for off-line analysis. Four sets of data were collected for each cat for a total of 20 trials at each direction. New directions were presented to the cats in varying orders. The data that were collected included the three orthogonal components of force exerted by each paw against the ground, the EMG signals from BF, and the position of the platform. EMG signals were differentially amplified, bandpass filtered (200-2000 Hz), full-wave rectified, and low-pass filtered prior to digitization. Forces, EMGs, and platform position were recorded on-line with a PDP 11/73 computer at 500 samples/s. Each trial consisted of at least 100 ms of quiet stance followed by platform movement and the postural response, for a total sampling time of 1 s. Averages and rasters of the data were plotted off-line using the PDP 11/73 computer.

Data analysis To eliminate force transients associated with the mass of the force plate, data recorded from the force plates alone were subtracted from those recorded with a cat on the platform. Averages of the subtracted force data (n = 5) were used for the following analysis. The force exerted by each paw was analyzed in terms of the change from the quiet stance, or background level, and was measured as the area under the curve for each component, x, y, and z. The area was divided by the time of integration to give an average change in force. The average change in force was computed for two epochs of time, a passive period and an active period. Changes in force during the initial 50 ms following onset of platform movement, were considered to be purely passive since there is not enough time for the cat to generate force as a result of evoked muscle activity (Macpherson 1988a). The active period began at the end of the passive phase and continued to the end of the rapid force response, usually 100-200 ms in duration. These force changes were termed the active response, although they include both passive and active components. For each EMG trace, the average activity level and standard deviation (SD) were computed for the background period (100 ms of quiet stance). Periods of significant increase or decrease from background were identified on a trial-by-trial basis using thresholds of + 2.5 SD and - 1.5 SD, respectively. The times of onset and offset of each response were determined using the cumulative sums technique (Ellaway 1978). For each EMG recording at each direction of translation, a probability of response was calculated: the number of responsive trials divided by the total number of trials, giving a number between 0 and 1. Responses that occurred in at least 50% of the trials and within 60 ms of the onset of platform

273 movement were included for further analysis. EMG responses were quantified by summing the differences from background level (area under the curve). The EMG areas were averaged over each set of 5 trials and the means from 4 sets of data were averaged to obtain a grand mean and standard error of the mean (SEM). For each electrode pair, the grand mean and SEM of the EMG were normalized to the maximum mean response of the sixteen translation directions. The normalized EMG data were plotted as a function of the direction of platform movement in polar coordinates, and such plots were termed EMG tuning curves.

A

0o

J 90~

Relation between B F compartments and electrode placement In a terminal experiment, cat nol was deeply anaesthetized (Nembutal) and the nerve supply to BF was exposed and dissected free. Individual primary nerve branches were stimulated (0.1 ms pulses @ 3 Hz) with hand-held bipolar electrodes slightly above threshold for just noticeable twitches. EMGs were recorded from the eight electrode pairs in order to determine the relationship between the innervation compartment boundaries and the region of muscle from which each signal was recorded. At the time of the terminal experiment, several pairs of electrodes in cat no2 were no longer functional. The relationship between innervation compartment boundaries and electrode position was determined on a gross morphological basis using anatomical dissection of the primary and secondary nerve branches (see Chanaud et al. 1991a).

Results The left B F o f cat n o l (4.9 kg) was innervated by four main nerve branches, anterior (BFa), the first middle ( B F m 1), the second middle (BFm2), and posterior (BFp). Sequential stimulation o f these f o u r branches, f r o m anterior to posterior, p r o d u c e d E M G signals that were recorded f r o m the following g r o u p s o f electrodes: 1 4 , 4-6, 6-7, and 7-8, respectively. T h e relationships between the innervation c o m p a r t m e n t territories and the electrode recording sites that were determined by electrical stimulation are s h o w n in Fig. 1B (left). The left B F o f cat no2 was innervated by three main nerve branches (Fig. 1B, right). Based on dissection, electrodes 1-3 were situated over the anterior c o m p a r t m e n t while electrode 4 appeared to be positioned over the b o r d e r o f the anterior and middle c o m p a r t m e n t s . Electrodes 5 and 6 were within the boundaries o f the middle c o m p a r t m e n t and electrode 7 was over the b o r d e r o f the middle and posterior c o m p a r t m e n t s . Finally, electrode 8 was within the limits o f the posterior c o m p a r t m e n t .

Forces at the ground D u r i n g quiet stance, the vertical forces o f cats n o l a n d no2 were distributed such that 52% and 54% o f b o d y weight respectively, was supported by the forelimbs. Consistent with a previous report ( M a c p h e r s o n 1988a), the center o f mass o f each cat was situated slightly anterior to the m i d p o i n t between the fore- a n d h i n d p a w s (or front and rear force plates).

Yx ~ z BFa

B

BFa BFml ~ BFm2 ~

.~

CAT .# 1

BFrn

~

BFp,

~

POSTERIOR

CAT "#2 X , ~

Fig. 1. A Platform paradigm and recording apparatus. The cat stood with each paw on a separate, triaxial force plate. The x-y-z axes illustrate the coordinate system for the lateral, longitudinal, and vertical forces, respectively, with the arrows pointing in the positive directions. The direction of platform movement was specified in a polar coordinate system, such that 180~ was an anterior translation, and 0~ posterior. B Biceps femoris (BF) innervation, compartment boundaries and EMG electrode placement. A single Silastic patch containing 8 pairs of electrodes was placed underneath the muscle with the electrodes against the inner surface of the muscle. Electrode pairs were numbered 1-8 from anterior to posterior. Each bipolar electrode consisted of two contacts, 3 mm in length, and 5 mm apart. The distance between each electrode pair was 4 mm. The relationship between the main nerve branches, the muscle territory innervated by each branch, and the electrode positions are shown for each cat. Both cats: BFa= anterior and BFp =posterior. Cat nol: BFml=first middle, BFm2=second middle. Cat no2: BFm =middle

F o r each cat and each limb, the longitudinal and lateral b a c k g r o u n d forces were c o m b i n e d to p r o d u c e horizontal plane vectors for the quiet stance period prior to each translation. All sixteen vectors were averaged for each limb to give a g r a n d m e a n o f force distribution in the horizontal plane. D u r i n g quiet stance, b o t h cats exerted a force anteriorly a n d laterally with each forelimb and posteriorly a n d laterally with each hindlimb. The pattern o f the horizontal plane vectors was similar across cats a l t h o u g h the amplitude and direction o f each vector could vary, as n o t e d in a previous study ( M a c p h e r s o n 1988a). B o t h the direction and amplitude o f the quiet stance vectors were relatively c o n s t a n t for each cat and did n o t v a r y as a function o f the direction o f p l a t f o r m m o v e m e n t ; thus, the cats did n o t alter their quiet stance force distribution despite being able to anticipate the direction o f the next perturbation.

274 CAT 1

@

18o

CAT 2

270

0 360

"o

200 ~

Z

_O 180

9 160 ~

I--

o ,'~ uuJ

_~ ~

PO

0

i

i

i

u

i

i

i

n 360

v

< ~.

2 PLATFORM

DIRECTION

(deg)

During the initial 50 ms following onset of platform movement (passive period), there was little change in the vertical forces exerted by the limbs. During the following active phase, the left hindlimb showed an increase in vertical force (loading) for translations ranging in direction from 180~ to 315 ~ A decrease in vertical force (unloading) was observed for translations from 0 ~ to 135 ~ (Figs. 3B and 4B, Vertical). Minimal changes were seen in vertical forces exerted by the left hindlimb for translations along the axis of 157-337 ~. Forces exerted in the horizontal plane showed the characteristic pattern that was more fully described in a previous study (Macpherson 1988a). During the passive period, the horizontal force vector exerted by the left hindlimb was opposite in direction to the platform movement for each direction o f perturbation. This was not true for the active force vectors. The average changes in horizontal force exerted by the left hindlimb during the active phase are shown in Fig. 2 (top) as a series o f vectors, one for each direction of translation. The points on the polygons in Fig. 2 represent equidistant points along the axes of translation; each horizontal plane vector takes origin at its respective point. It is evident that the active vectors do not directly oppose the perturbation. Plots of the vector direction and amplitude as a function of'the direction of platform movement are shown in Fig. 2 (bottom). There were two preferred directions for the correction force vector, indicated by the regions of zero slope in the linear plots of Fig. 2 (bottom). The direction of the horizontal vector remained constant over a broad range o f translation angles, and then underwent a rapid change to the other preferred direction. In contrast, the amplitude of the vector was modulated continuously; the amplitude curve showed a maximum at the midpoint o f each constant direction segment and a minimum at the points o f rapid change in vector direction (Fig. 2, bottom).

Fig. 2. Horizontal plane force vectors durin9 the active postural response. Top: the active horizontal force vectors at each translation angle are shown for the implanted left hindlimb. The points on the polygons represent equidistant points on the axes of translation in the horizontal plane. The force exerted by each paw is represented as a vector taking origin at its respective point on the polygon. Bottom: direction and amplitude of the active horizontal force plotted as a function of the direction of platform translation. During the corrective postural response, the left hindlimb produced horizontal forces in two preferred directions (160~ and 45~ for Cat nol, 200~ and 45~ for Cat no2). The horizontal dotted line indicates the average direction of the horizontal force exerted by the left hindlimb during quiet stance

B F E M G activity

During quiet stance, tonic activity was observed only in the anterior compartment (see Fig. 6, electrodes 1-3). The middle and posterior compartments were relatively silent. Following platform translation, E M G activity was evoked at an average latency of 38 ms in cat nol and 52 ms in cat no2. The average latency recorded from each electrode pair of the two cats is shown in Table 1. Within each cat, there was no difference in latency among the eight recordings. Cat no2, however, had longer latencies than cat no 1 and lower levels of activation recorded at each electrode pair. The tuning curves of the normalized E M G activity recorded from the eight electrode pairs are shown in Fig. 3 for cat no l and Fig. 4 for cat no2. The tuning curves o f vertical and horizontal force changes for the recorded limb are included for comparison. Each compartment was activated over a broad range o f translation angles with a monotonic increase in relative activity to a maximum, followed by a decrease back to zero. The tuning curves from the BFa electrodes coincided with the vertical and horizontal force tuning curves in the upper right quadrant of the plane (Figs. 3 and 4). In other words, the BFa compartment was activated during those perturbations in which the limb generated a horizontal force backward and outward (45 ~ and a vertical force downward. The translation direction that evoked maximum activation of the BFa compartment (225 ~ also evoked the maximum force production by the limb against the ground. Although similar, the tuning curves of BFa and BFm were not superimposable. The BFm tuning curves were rotated in a counterclockwise manner relative to those of BFa. Even within a compartment there was a gradual but small shift in the location of the tuning curves. The most

275

A EMG AMPLITUDE I

ANTERIOR

180

,,~

./..:. ~

2

i3

4

9. , , .

9

Fig. 3A, B. EMG and force tuning curves of cat nol. A EMGs: The relative amplitude of E M G activation is plotted as a function of translation direction (deg) for each electrode pair. The central region represents normalized mean amplitudes of evoked E M G (dark); the surrounding envelope (light) illustrates + 1 SEM. The BFa and BFm compartments were active during translations of 180-293 ~ and 113-270 ~ respectively. The BFp compartment was active at translations of 23-203 ~ (The BF7 electrode was situated directly over the BFm/BFp border and resulted in a combined tuning curve.) B Forces: Vertical force exerted by the left hindlimb increased during translations of 180-315 ~ (loading), and decreased during translations of 0-135 ~ (unloading). The limb produced a horizontal plane force vector at a constant direction of 45 ~ for translations of 157 315~ the horizontal force had a constant direction of 160 ~ for translations of 337-135 ~ . The tuning curve illustrates the amplitude of each constant-direction force vector as a function of translation direction

ib

270

90

0 (

6

:::

:~

\

.....

i~

~iiiiiiii:~

- -

MIDDLE 1

I

MIDDLE 2

I

I B

POSTERIOR

LH FORCE AMPLITUDE VERTICAL

~

UNLOADING

HORIZONTAL

~

LOADING

~

160 ~

~

45~

A EMG AMPLITUDE I

ANTERIOR t 180

3

4 :

90

i

270

0 7 f ................................. i,~

!

8

t ~;

MIDDLE

a:

I I

POSTERIOR

g

LH FORCE AMPLITUDE VERTICAL

HORIZONTAL

x. ~ ~LOADING UNLOADING

~

'

:. ~

200 ~

~

5o

Fig. 4A, B. EMG and force tunino curves of cat no2. A EMGs: The BFa and BFm compartments were active during translations of 180-293 ~ and 90-247 ~ respectively. The BFp compartment was active during translations of 67-180 ~ B Forces: Vertical force exerted by the left hindlimb increased from 180-315 ~ (loading), and decreased during translations of 0-135 ~ (unloading). The limb produced a horizontal plane force vector at 45 ~ during translations o f 157-293 ~ the horizontal vector had a constant direction of 200 ~ during translations of 315-135 ~

276 Table 1. Latencies of EMG responses with respect to onset of platform movement

1

2

3

4

5

6

7

8

42 5

37 6

36 4

38 6

36 4

38 5

37 3

37 2

Electrode pair Cat n o l Mean SEM Innervation:

BFa I

I B F m l I-

I BFm2I

I BFpI-

I

Cat no2 Mean SEM

51 7

51 5

51 4

44 4

55 6

55 5

54 3

56 3

Innervation: BFa I BFm I

I BFp I-

I

ELECTRODE PAIR o--ol

10

.

9 --.2 zx--

0.8

.

.

.

dx ~

u--n5 0.6

Fig. 5. Probability of E M G response as a function of translation direction. The probability of an evoked E M G response is shown for electrode pairs 1-6 of Cat n o l . Electrode 1 was the most anterior pair on BFa and electrode 6 was the most posterior pair on BFm. Differences in the probability of response are evident across all electrode pairs, even within compartments

s - - s

p_~ -.~

.

A--A4

~) ~"

.

T

...................................................................... 0,4

8 0.2

0.0 -90

I 0

90

180

270

360

TRANSLATION DIRECTION ( d e g )

anterior recordings in both cats (electrodes 1 and 2) showed responses in the right, upper quadrant of the plane, from 180 to 293 ~. Subsequent electrode pair recordings showed a counterclockwise progression o f the tuning curves. The most posterior region o f the BFm compartment (electrode 6) was activated from 113 ~ to 270 ~ in cat n o l , and 90 ~ to 225 ~ in cat no2. This gradual progression of responsiveness both within and across the anterior and middle compartments is more clearly observed in plots o f the probability of response versus translation direction (Fig. 5). A response probability of 0.5 or greater (above horizontal dotted line in Fig. 5) was chosen as the criterion for inclusion o f the response in the tuning curves. A gradual change in responsiveness was observed within the BFa (electrodes 1,2,3) and within the BFm (electrodes 5,6) for both cats. Since the tuning curves o f the anterior and middle compartments do not superimpose, there were certain

perturbations for which one compartment was activated independent o f the other. BFa was activated without BFm for translations at 293 ~ (e.g. Fig. 6F); BFm was activated without BFa for translations at 113 ~ to 157 ~ for cat nol (e.g. Fig. 6B, C) and 90 ~ to 157 ~ for cat no2. During synchronous activation of BFa and BFm, the degree of activation of BFa was equal to or greater than that of BFm (e.g. Fig. 6D, E, note that the scale bar denotes 0.2 mV for electrodes 1-3 and 0.1 mV for 4-6). The tuning curve of the posterior compartment, BFp, was in the leftward region o f the plane and overlapped only slightly with those of the anterior and middle compartments. The region of BFp activation overlapped somewhat with th.e region of anterior horizontal force and decreased vertical force exerted by the limb (Figs. 3 and 4). However, the E M G tuning curve extended beyond the boundaries of these force regions to overlap slightly with the areas of increased vertical force and

277

A BFa

"~

B

900

~

135"

C

~

157"

2 3

^,

BFml BFm2I

,L

BFpI

A

,

]

2930

2470

~/~ 180~

,

1 i..,:,~.9 BFa

BFml BFm2I BFpI

8 ~ ,

9

PLAT

.\

.\

,

0

200

,

,

,

400

600

800

-\

,

lO00ms

Fig. 6A-F. Averaged EMG traces from BF compartments of cat nol at selected translation angles. The relationship between the innervating nerve branches and the recording electrodes is shown at the left, with electrode 1 on the anterior edge of BFa and electrode 8 on the posterior edge of BFp. The vertical dotted lines indicate the onset of platform movement. Evoked E M G responses included in the analysis are shaded. A During translation at 90 ~ BFa activity (electrodes 1-4) was decreased relative to background, BFm (electrodes 4-7) showed no change, and BFp (electrodes 7-8) was ac-

tivated. B At 135 ~ BFa showed little or no change in activity, BFm and BFp showed slight activation. C At 157 ~ BFa showed no change in activity but the BFm and BFp compartments were activated. D At 180 ~ BFa, BFm and BFp were active. E At 247 ~ the BFa and BFm compartments were strongly activated but BFp remained inactive. F At 293 ~ the BFa compartment was active and BFm and BFp remained inactive. Calibration bar: 0.2 mV for electrodes 1-3 and 0.1 mV for electrodes 4-8

278 postero-lateral horizontal force. Thus, the BFp tuning curve also covered the region of transition in active force response, when the change in force exerted by the limb was minimal. The greatest activation of BFp was observed for translations of 90 ~ for cat nol and 113 ~ for cat no2; at these angles the anterior compartment was often inhibited (Fig. 6A), and the middle compartment was either not activated (cat no 1) or minimally activated (cat no2). BFp could be activated independent of the other two compartments (Fig. 6A), in combination with BFm only (Fig. 6B, C), or with both BFa and BFm (Fig. 6D). Furthermore, as the direction of translation increased from about 90 ~ onward, the activation of BFp gradually decreased while that of BFm and BFa gradually increased (Fig. 6A-E).

Discussion

During quiet stance, only the anterior compartment of the biceps femoris muscle was active. Following platform translation, activity was evoked from subregions of the muscle in a direction-dependent manner. As the direction of platform movement changed from 293 ~ in a counterclockwise manner, the region of activation of BF progressed across the muscle from the anterior to the posterior border. Whereas this progression was continuous across the anterior and middle compartments, there was a more discrete change in response profile at the BFm-BFp border. Various combinations of compartment activation were observed progressively; BFa alone, BFa with BFm, all three together, BFm with BFp, or BFp alone. When the platform was translated horizontally, the limbs moved with the platform and the trunk remained in position due to inertia (Rushmer et al. 1983). The limbs that were translated away from the body were unloaded and the limbs that were moved under the body were loaded. In order for the cat to maintain upright stance, the four limbs had to produce forces in the horizontal plane that propelled the body back to the original position over the feet. Each hindlimb exerted against the support an active horizontal force vector in one of only two preferred directions. For a detailed discussion of the cat's postural response strategy, see Macpherson (1988a). Despite the discrete postural strategy (i.e. only two directions of active horizontal force for a range of directions of perturbation), the evoked EMG activity showed a continuous response profile across the BFa and BFm that varied both within and across compartment boundaries. Furthermore, although there was a rather discrete change in response profile from BFa/BFm to BFp, the boundaries of the EMG tuning curves did not match those of the force responses. The tuning curve of the anterior compartment coincided with the regions of increased vertical force and the postero-lateral horizontal force. However, the tuning curves of both the middle and posterior compartments did not coincide with either horizontal or vertical force tuning curves. Rather, BFm and

BFp tuning curves overlapped the region of transition from one force direction to another. These data support the previous finding that the discrete force strategy is not the result of a discrete pattern of muscle activation, or synergy (Macpherson 1988b). The continuous nature of the response profile in BF, especially evident within the BFa and BFm compartments, is in marked contrast to findings in sartorius (Pratt and Loeb 1991), another complex muscle that crosses both hip and knee. The sartorius is functionally divided into anterior and posterior subregions; each subregion displayed a distinctive pattern of activation that was homogeneous within the subregion. The boundary between areas of differential activation was discrete, and coincided with the anatomical division of sartorius into knee flexor and knee extensor subregions. It is possible that sartorius may show a more continuous change in response profile like BF, during postural reactions that are not restricted to the sagittal plane, but this has not yet been examined. It may be that a broader range of motion of the limb is necessary in order to reveal the full potential for independent activation within these complex muscles. Considering the differences in the mechanical relationship of each BF compartment to the skeleton (Chanaud et al. 1991a), there is an obvious advantage to the nervous system to be able to selectively activate individual or adjacent compartments. Although the three compartments of BF have origins with similar lever arms to the hip joint, the distributed insertion results in qualitatively and quantitatively different lever arms to the knee joint. BFa inserts on the femur and has no lever arm across the knee. BFm inserts on crural fascia of the proximal shank and has a hip-to-knee lever arm ratio of 1 : 1. Finally, BFp inserts at the distal shank and has a hip-to-knee lever arm ratio of 3 : 5, giving this BF compartment the greatest mechanical advantage for knee flexion. In general, the EMG tuning curves showed that as the biomechanical situation varied, the activity level of individual compartments also varied. Each muscle region exhibited a peak activation at a specific translation angle and decreased progressively as the translation angle decreased or increased. Thus, for postural adjustments, each compartment appeared to be independently controlled. It is likely that the continuous nature of the muscle response profile was related to the limb kinematics. As the direction of the translation changed, so too did the trajectory traversed by the hindlimb. The position of the left hindlimb at the end of translation varied from extended to abducted, to flexed and finally to adducted, as translation angle changed from 0 ~in a clockwise manner. Since we were unable to record the kinematics, the joint angle changes occurring at the hip and the knee for each direction of translation are not known. Therefore, we cannot speculate on the nature of the change in length of the muscle and .afferent input with change in platform direction. The tuning curve for the anterior compartment was located in the upper right quadrant of the plane. As mentioned above, this tuning curve was more or less

279 coincident with the region of increased vertical force and postero-lateral horizontal force exerted by the hindlimb. The BFa tuning curve was similar to the tuning curves found previously for other hip extensors such as gluteus medius, adductor femoris, and caudofemoralis as well as the ankle extensors, lateral gastrocnemius, soleus, and plantaris (Macpherson 1988b and unpublished observations). The origin and insertion of the BFa suggests that this muscle region produces an extensor and abductor torque at the hip joint (Chanaud et al. 1991a). It is likely that such a torque may contribute to the production of the postero-lateral horizontal force against the support, as well as the vertical support for the increased loading on the limb. Studies of BFa activity during locomotion have also concluded that it functions as a hip extensor (Engberg and Lundberg 1969; English and Weeks 1987; Chanaud et al. 1991b). The BFm compartment exhibited a tuning curve that was aligned more about the sagittal axis, in the upper region of the plane. Although the region of activation overlapped with that of BFa, it was not coincident. The tuning curve of BFm was similar to that of the anterior region of gracilis, characterized in a previous study (Macpherson 1988b). Although BFm has abductor functions and gracilis has adductor functions, both muscles produce extensor torques at the hip and flexor torques at the knee. BFm has been shown to have equal moment arms for hip extension and knee flexion (Chanaud et al. 1991a). The tuning curve for BFm crossed over the boundary of change in force production at the support, so it is difficult to speculate on the mechanical role that this muscle region played during postural responses. However, since the limb was flexed with respect to the trunk at the end of the translations in which BFm was active, and since the correction force vector was directed back and lateral for most of those same translations, it is conceivable that the vector of ground reaction force passed in front of the knee. In that case, one would expect to observe a hip extensor and a knee flexor torque (see van Ingen Schenau 1989). Both BFm and gracilis could contribute to this combination of joint actions. However, it is necessary to have the kinematic data in order to fully understand the function of BFm in this context. When BFa and BFm were active synchronously, BFa usually showed a higher amplitude of activation than BFm (Fig. 6). Recordings of BFa and BFm during locomotion have also demonstrated a similar relationship in activation levels at slow to moderate walking speeds (English and Weeks 1987; Chanaud et al. 1991b). These amplitude differences may merely reflect differences in the volume of muscle being recorded since BFa is thicker than BFm (see Fig 5, English and Weeks 1987). Alternatively, differences in activation levels may reflect the differences in the mechanics of BFa and BFm, related to the differences in insertion. It is likely that the strategy by which the different regions of BF are activated depends on the particular motor task that is being executed. The BFp tuning curve for the left hindlimb was located around the transverse axis, in the leftward region of the plane. Through most of this region, the hindlimb was unloaded and produced a horizontal force in the

anterior direction, although the BFp tuning curve did overlap the region of change in force direction. Without the kinematic data, it is difficult to speculate on the role of BFp during this range of perturbations. Due to its relatively long lever arm to the knee, BFp functions during locomotion as a knee flexor (Engberg and Lundberg 1969; English and Weeks 1987). During postural responses, BFp activation was similar to that of other knee flexors such as posterior gracilis and semitendinosus (Macpherson, unpublished observations). The tuning curves of knee extensors were reciprocally located, occupying the opposite, or mirror-image region of the plane (Macpherson 1988b). It is possible that knee flexors and extensors are required for lateral stabilization of the knee following translations in and around the transverse axis. During translations for which BFp was maximally activated, the tonically active BFa was inhibited while the silent BFm showed no change from background activity. This reciprocal activity pattern of BFa and BFp is interesting given the fact that both these regions have similar moment arms for extension at the hip (Chanaud et al. 1991a). Coactivation of the BFm and BFp compartments without the BFa was observed during translations of 135-180 ~ The set of translations for which BFm and BFp alone were coactivated spanned the region of changeover in both vertical and horizontal force directions. The vertical force changed from decreasing force to increasing force, and the horizontal force vector changed from 165 ~ to 45 ~. Over this range the relative activation of BFm and BFp changed inversely as translation angle increased from 135 ~ to 180 ~ with BFp amplitude decreasing towards 180~ A knowledge of the kinematics of the postural responses is required to understand more fully the roles of these BF compartments. Differences in activation patterns among the BF compartments cannot be attributed to recruitment factors related to differences in size and type of motoneurons (Henneman and Mendell 1981). A histochemical evaluation of BF found no major differences in the proportions of fiber types (SO, FOG, FG) across the compartments (Chanaud et al. 1991b). There is, however, a segregation in the Ia monosynaptic inputs to the BF motoneuronal pool. Eccles and Lundberg (1958) found that BFa received its greatest Ia input from other hip extensors, whereas BFp received its greatest Ia input from knee flexors. Furthermore, Botterman and colleagues (1983) showed that motoneurones subserving each of the three compartments received their largest Ia input from their own compartment, and lesser inputs from the other two compartments. Interestingly, this latter study also demonstrated a dichotomy in the BFm motoneurones. Some were classified as extensors whereas others were classified as flexors, based on differential Ia input from knee flexors. This may relate to the differences observed between BFa and BFm during postural responses. Differential activation of the three BF compartments has been observed for locomotion, scratch, and paw shake (English and Weeks 1987; Chanaud et al. 1991b; Pratt et al. 1991). The combinations of activity patterns across the compartments varied, depending on the move-

280 m e n t studied. These results along with the present findings for p o s t u r a l responses graphically illustrate the fact that patterns o f activation within a muscle are context dependent, a n d are likely to be finely-tuned according to the mechanical requirements o f the particular task being performed. In s u m m a r y , during postural reactions the cat BF muscle was n o t activated as a h o m o g e n e o u s muscle mass. As the direction o f translation changed, the activation travelled across the muscle in a c o n t i n u o u s m a n n e r within B F a a n d B F m , and with a m o r e discrete j u m p between B F m a n d BFp. E a c h c o m p a r t m e n t , BFa, B F m , a n d BFp, exhibited a different p a t t e r n o f activation, as has been previously r e p o r t e d for other movements. H o w e v e r , this study s h o w e d that differences in activation also occurred within a n a t o m i c a l c o m p a r t m e n t s , consistent with the findings o f B o t t e r m a n et al. (1983), a n d predicted by English a n d Weeks (1987). These activation differences m a y reflect the existence o f multiple functional units within a single a n a t o m i c a l c o m p a r t m e n t . Thus, the B F m a y consist o f several functional units that do n o t coincide with the a n a t o m i c a l c o m p a r t m e n t s , and indeed, m a y be smaller t h a n the a n a t o m i c a l c o m p a r t m e n t s . The presence o f functional differences within the B F muscle suggests that there are likely to be differences within the B F m o t o n e u r o n pool in the distribution o f b o t h afferent inputs and central inputs. The facility for differential activation o f the muscle regions w o u l d i m p a r t a great flexibility in activity patterns that could be a d v a n tageous for the p e r f o r m a n c e o f different types o f m o t o r tasks.

Acknowledgements. This study was supported by grants to J.M.M. from the MRC of Canada and Queen's University.

References Balice-Gordon RJ, Thompson WJ (1988) The organization and development of compartmentalized innervation in rat extensor digitorum longus muscle. J Physiol 398 : 211-231 Botterman BR, Hamm TM, Reinking RM, Stuart DG (1983) Localization of monosynaptic Ia excitatory post-synapfic potentials in the motor nucleus of the cat biceps femoris muscle. J Physiol 338 : 355-377 Chanaud CM, Macpherson JM (1987) Independent activation of compartments of feline biceps femoris during postural responses to translations of the support surface. Neuroscience Abstr 13:370 Chanaud CM, Pratt CA, Loeb GE (1991a) Functionally complex muscles of the cat hindlimb. II. Mechanical and architectural heterogeneity within a parallel-fibered muscle. Exp Brain Res 85 : 257-270 Chanaud CM, Pratt CA, Loeb GE (1991b) Functionally complex muscles of the cat hindlimb. V. The roles of histochemical fiber-type regionalization and mechanical heterogeneity in differential muscle activation. Exp Brain Res 85:300-313 Eccles RM, Lundberg A (1958) Integrative pattern of Ia synaptic

actions on motoneurons of hip and knee muscles. J Physiol 144:271-298 Ellaway PH (1978) Cumulative sum technique and its application to the analysis of peristimulus time histograms. Electroenceph Clin Neurophysiol 45 : 302-304 Engberg I, Lundberg A (1969) An electromyographic analysis of muscular activity in the hindlimb of the cat during unrestrained locomotion. Acta Physiol Scand 75:614-630 English AW, Letbetter WD (1981) Intramuscular "compartmentalization" of the cat biceps femoris and semitendinosus muscles: anatomy and EMG patterns. Neuroscience Abstr 7:557 English AW, Weeks OI (1987) An anatomical and functional analysis of cat biceps femoris and semitendinosus muscles. J Morphol 91 : 161-175 Hammond CGM (1987) Motor-unit territories supplied by primary branches of the phrenic nerve: an electromyographic study of the cat diaphragm. Master's thesis. Queen's University, Kingston, Ontario Henneman E, Mendell JM (1981) Functional organization of motoneuron pool and its inputs. In: Handbook of physiology: the nervous system, Vol 2, Chapt 11, Am Physiol Soc, Bethesda, MD, pp 423-507 Holler JA, Loeb GE (1980) Implantable electrical and mechanical interfaces with nerve and muscle. Ann Biomed Eng 8:351-360 Letbetter WD (1974) Influence of intramuscular nerve branching on motor unit organization in medial gastrocnemius muscle. Anat Rec 178:402 Loeb GE, Gans C (1986) Electromyography for experimentalists. University of Chicago Press, Chicago Lywood DW, Adams DJ, van Eyken A, Macpherson JM (1987) Small, triaxial force plate. Med Biol Eng Comput 25: 698-70l Macpherson JM (1988a) Strategies that simplify the control of quadrupedal stance. 1. Forces at the ground. J Neurophysiol 60: 204-217 Macpherson JM (1988b) Strategies that simplify the control of quadrupedal stance. 2. Electromyographic activity. J Neurophysiol 60: 218-231 Macpherson JM, Craig LS (1986) Postural responses in cats to movements of the support surface in the horizontal plane: comparison of lateral and longitudinal displacements. Neuroscience Abstr 12:1300 Macpherson JM, Lywood DW, van Eyken A (1987) A system for the analysis of posture and stance in quadrupeds. J Neurosci Meth 20: 73-82 Pratt CA, Chanaud CM, Loeb GE (1991) Functionally complex muscles of the cat hindlimb. IV. Intramuscular distribution of central inputs and cutaneous reflexes in broad, bifunctional thigh muscles. Exp Brain Res 85:281 299 Pratt CA, Loeb GE (1991) Functionally complex muscles of the cat hindlimb. I. Patterns of activation across sartorius. Exp Brain Res 85 : 243-256 Richmond FJR, MacGillis DRR, Scott DA (1985) Muscle-fiber compartmentalization in cat splenius muscles. J Neurophysiol 53 : 868-885 Rushmer DS, Russell C J, Macpherson J, Phillips JO, Dunbar DC (1983) Automatic postural responses in the cat: responses to headward and tailward translation. Exp Brain Res 50:45-61 van Eyken A, Perlin S, Lywood DW, Macpherson JM (1987) Robotic force platform for the study of posture and stance in the quadruped. Med Biol Eng Comput 25:693-697 van Ingen Schenau GJ (1989) From rotation to translation: constraints on multi-joint movements and the unique action of bi-articular muscles. Hum Mov Sci 8 : 301-337