An anatomical and functional analysis of cat biceps femoris and semitendinosus muscles

JOURNAL OF MORPHOLOGY 191:161-175 (1987)

An Anatomical and Functional Analysis of Cat Biceps Femoris and Semitendinosus Muscles

ARTHUR WM. ENGLISH AND OPHELIA I. WEEKS Department of Anatomy and Cell Biology, Emory University School of Medicine, Atlanta, Georgia 30322

ABSTRACT The anatomy, architecture, and innervation patterns of the hamstring muscles, biceps femoris, and semitendinosus were examined in adult cats using microdissection and glycogen-depletion techniques. The biceps femoris muscle consists of two heads. The anterior head, which attaches mainly to the femur, is divided into two parts by the extramuscular branches of its nerve. The posterior head is innervated by a single nerve. Semitendinosus is composed of two heads, one proximal and one distal to a tendonous inscription, each of which is separately innervated. The extramuscular branches of the nerves to these hamstring muscles thus partition them into innervation subvolumes termed parts. The available evidence suggests that each of the parts of these muscles so innervated is not equivalent to the collections of single motor units that have been described for ankle extensors as neuromuscular compartments. It is quite likely that each of the parts of the hamstring muscles may contain more than one neuromuscular compartment. Using chronically implanted EMG electrodes, the activation patterns of different parts of the hamstring muscles were analyzed during locomotion. The anterior and middle parts of biceps femoris are active during the early stance phase, probably producing hip extensor torque. The posterior part of biceps femoris and semitendinosus act most consistently as flexors, during the early swing phase, but also may function in synergy with hip, knee, and ankle joint extensors near the time of foot placement. Greater variability is found in the activity patterns of posterior biceps femoris and semitendinosus with respect to the kinematics of the step cycle than is observed for anterior and middle biceps femoris. It is suggested that this variation may reflect a larger role of sensory feedback in shaping the timing of activity in posterior biceps femoris and semitendinosus than in "monarticular" muscles.

Recent studies have provided evidence that romuscular compartment and are each mammalian skeletal muscles are composed thought to be composed of a unique set of of divisions called neuromuscular compart- motor axons. The afferent axons, at least ments. The results of studies of several mus- from muscle sense organs (spindles and ten- cles in different species (Burke et al., '73; don organs), are thought to be organized in a Carlson, '78; English and Letbetter, '82a,b; similar fashion (e.g., Cameron et al., '81 and English and Weeks, '84; Gonyea and Ericson, references). '77; Letbetter, '74; Richmond and Abrahams, These results have provided an interesting '75; Richmond et al., '85; Wineski and Her- peripheral substrate for partitioning of mus- ring, '83) have shown that the muscle fibers cle function. Detailed studies of the patterns innervated by individual alpha motoneurons of central connections of these compartments (muscle units; Burke and Tsairis, '73) are not using cat medial and lateral gastrocnemius distributed equally to all parts of muscles but muscles (Lucas and Binder, '84; Lucas et al., are distributed to subvolumes of muscle supplied by primary muscle nerve branches. The

the nerve to a muscle each innervate a neu- GA 30322.

Address reprint requests to Dr. Arthur Wm. English, Depart- Occurring first Order branches Of ment of Anatomy and Cell Biology, Emory University, Atlanta,

Q 1987 ALAN R. LISS, INC.

162 A. WM. ENGLISH AND 0.1. WEEKS

'84; Vanden Noven et al., '86) have provided clear examples of localized weighting of muscle spindle primary connections within the motor nuclei subserving these muscles. Sim- ilar analyses using the extramuscular branches of the cat hamstring muscles, biceps femoris (BF) (Botterman et al., '83a), and semimembranosus (SM) (Hamm et al., '85) have yielded similar results. In contrast, studies using the two divisions of semitendinosus (ST) (Botterman et al., '83b) failed to reveal a reflex partitioning. Except for a brief description in an early paper by Eccles et al. ('571, which formed the basis for the physiological studies cited above, and a recent paper on the properties of ST (Bodine et al., '82) describing its peculiar in-series architecture, little is known about the anatomy, architecture, and innervation patterns of cat BF and ST except the division of the hamstring muscle nerve into anterior biceps femoris-semimembranosus (AB-SM) and posterior biceps femoris-semitendinosus (PB-ST) divisions (e.g., Eccles et al., '57). Thus, it is of consid- erable interest to know whether BF and ST are organized peripherally into compart- iiients in a manner similar to the triceps surae muscles.

In recent kinesiological studies it has been speculated that the organization of the ankle extensor muscles of cats into compartments may have functional significance (English, '84). Each compartment has been postulated to form a module onto which the CNS control of movement might be focused. The evidence leading to the proposal of this hypothesis is that motor units in different compartments can be activated with a certain amount of independence during posture (Russell et al., '82) and locomotion (English, '83, '84). It is thus of interest to determine whether different parts of BF are activated with some degree of independence during normal function, whereas different parts of ST are not, as the results of the reflex localization experiments would predict.

The purpose of this paper is to describe the anatomy, architecture, and innervation patterns of cat BF and ST muscles, to examine the patterns of electromyographic (EMG) activity in different parts of these muscles during normal locomotion, and to compare these results to those of cat triceps surae muscles. A preliminary report of some of these data was made (English and Letbetter, '81).

MATERIALS AND METHODS Anatomical Studies

To examine the anatomy and innervation patterns of BF and ST, methods similar to

those used to describe compartments in calf muscles were used. In six adult cats, the hindlimbs were removed immediately postmortem by cutting through the sacroiliac joints, carefully dissecting the perineal mus- culature and cutting through the pubic symphysis, after which the skin and sub- cutaneous fascia were removed. This procedure ensured that the entire hamstring complex could be studied intact. The limbs were then stained en bloc for acetylcholinesterase (AChE) (for details see English and Letbet- ter, '82a) and were immersed in neutral buff- ered formalin. They were later used to determine muscle architecture and muscle nerve branching patterns.

In six other cats, patterns of innervation were studied using the method of glycogen depletion (Edstrom and Kugelberg, '68). An- imals were food deprived for 24 hours prior to surgery, but no other dietary restrictions or additions were imposed. Under pentobarbital anesthesia (40 mgiKg, IP) the hamstring nerves were exposed just anterior to the anterior margin of BF. In each experiment a single nerve branch to both BF and ST was isolated and all others cut. The re- mainder of the hamstring nerves and the sciatic nerve were also cut. The intact nerve branches were then stimulated according to a paradigm known to deplete muscle fibers of their glycogen (see e.g., English and Let- better, '82a for details). Stimulus strength was adjusted to be at least four times that needed to evoke a visible twitch in the innervated muscle. Care was taken to elevate the stimulating electrodes in a pool of mineral oil made by the flaps of skin about the sur- gical incision so that stimulus current did not spread passively to denervated muscle fibers. Stimulation was continued for 1-1 112 hours. After a few minutes of stimulation the magnitude of muscle contraction was visibly reduced, and at that time intermittent vascular occlusion was begun. This was produced by application of pressure to a ligature placed about either the abdominal aorta or the external iliac artery for 2-3 minutes every 10 minutes. Vascular occlusion is thought to cause muscle fibers that normally use mainly oxidative metabolic pathways to use glycolytic mechanisms, and thus aid in depleting their stores of glycogen. Occlusion without stimulation is said not to result in glycogen depletion (Kugelberg, '73).

At the conclusion of stimulation, the oper- ated legs were removed in the manner described above and quickly immersed in a bath of isopentane cooled in liquid nitrogen. The frozen leg was then placed on a block of dry

CAT BICEPS FEMORIS AND SEMITENDINOSUS MUSCLES 163

ice and cut into smaller blocks with a hack- saw cooled in liquid nitrogen. Care was taken to attempt to make the saw cuts as close as possible to perpendicular to the orientation of fascicles in BF and/or ST so that serial histologic sections could be reconstructed in terms of muscle architecture. However, since both muscles are pennate and arise from a deep tendon to insert on a superfkial tendon, it was impossible to cut the muscles in a plane truly perpendicular to the direction of their fascicles. From these blocks, BF and ST were removed from the underlying bone and thigh muscle, using cooled razor blades. The resulting frozen muscle blocks were then cut on a cryostat. Sections 40 pM thick were taken every 100 pM throughout the length of the block. Thus, except for saw cuts (ca. 2 mm thick the entire muscle was sampled every 1,000 pM from origin to insertion. Sec- tions were reacted for the demonstration of glycogen using the periodic acid-Schiff (PAS) reaction. Muscle fibers that were innervated and had been depleted of their glycogen were either weakly stained or unstained. Dener- vated fibers stained intensely for glycogen. Regions of glycogen depletion were determined from tracings of sections made using a projection miscroscope. The tracings were then used to provide a three-dimensional re- construction of the innervation territories of each of the branches to BF and ST.

EMG recordings To examine the overall activity of motor

units in different parts of BF and ST, EMG recordings were made from selective electrodes which were chronically implanted. The implantation procedure used was similar to that described previously (English, '84). In nine pentobarbital anesthetized cats, BF and ST were exposed following a posterior skin incision, and bipolar fine-wire electrodes were implanted into the muscles. In six cats, all electrodes were similar to those described previously in studies from this laboratory (English, '84). Teflon-coated, stranded, stain- less-steel wires were implanted into the belly of the muscle in question. Approximately 1.0 mm of insulation was stripped from the tip of each wire and the two wires were spaced such that their tips were approximately 1.0 mm apart. Wires were implanted such that the tips were perpendicular to the direction of the muscle fascicles. The placement of each of the wires was checked postmortem. Al- though some re-orientation of tips could have occurred postoperatively, no detailed examination of the tip orientation was conducted.

Electrode leads were passed subcutaneously to a multipin connector plug Microtech Inc., Boothwyn, PA) in a vitreous carbon percuta- neous interface CBiosnap). The connector plug was insulated with epoxy. Several recording sites were chosen for study, but all of the data reported in this study are from recording sites indicated in Figure 1. These sites were chosen because they are as close to the center of each muscle as possible and also because they are at some distance from other muscles. The latter point is important because myoelectric signals of large amplitude that are generated in surrounding muscles (such as LG) may be recorded in the hamstring muscles. Although close interelectrode spacing and differential recording with a high common mode rejection might minimize recording of such volume conducted signals, they form a serious source of contamination, especially in small, flat muscles such as the posterior head of BF. In such muscles, in- dwelling electrodes might actually lie as close or closer to the sources of electrical signals from surrounding muscles as from the implanted muscles (English and Weeks, '85).

Recordings were made beginning on the third postoperative day. Repeated recordings for up to 10 weeks postoperatively did not differ qualitatively from those made after 3 days. In each experiment film and EMG recordings were made during a relatively large number (e.g., 20) of passes along the walkway, in an attempt to record stepping se- quences at different speeds and to sample the range of variability in locomotor behavior. In each pass, step cycles in the center of the walkway were chosen for analysis because they would include neither initial or final steps, and therefore any nonsteady-state con- ditions, and because analysis of stepping movements from cine film records would avoid any effects of parallax. In general two or three step cycles could be analyzed in each pass. All recordings were made using high input impedance (22 mQ) differential ampli- fiers at a gain of 1,000. All signals were band pass filtered (50-500 Hz) prior to processing. In most cases EMG signals were recorded online using a laboratory computer system that sampled activity at a rate of 1 kHz. These records were synchronized with high- speed (200 fps) cinematographic records as described previously (English, '84) and used to determine the activity patterns of BF and ST regions as the animals stepped overground. Although no attempts were made to analyze EMG intensity quantitatively in this study, all EMG signals were rectified and averaged (as described previously) in order


A bbreviatwns BF, Biceps femoris ST, Semitendinosus SM, Semimembranosus EMG, Electromyography AB, Anterior biceps femoris MB, Middle biceps femoris PB, Posterior biceps femoris AChE, Acetylcholinesterase BFa, Innervation territories of AB, MB, and PB

BFm, nerves BFP,

STp,

a-d, LG, Lateral gastrocnemius G. Gracilis

Innervation territories of proximal and distal branches of ST nerve tendons of origin and insertion of BF and ST

STd,

that certain qualitative assessments of EMG intensity might be made.

RESULTS Muscle architecture

The results of analysis of the anatomy and architecture of BF and ST are shown in Fig- ures 1 and 2. Figure 1 is a semidiagrammatic

drawing of BF and ST muscles showing the general architecture of their fascicles in relation to their tendons of origin and insertion. Both BF and ST are attached proximally to the ischial tuberosity but arise mainly from a common tendon that extends distad from the ischial tuberosity (Fig. 1:a). Al- though this tendon does not course exactly in a sagittal plane, the entire anterior head of BF arises from its more laterally directed surface. The entire posterior head of BF and the proximal fascicles of ST originate from the more medially directed surface of this tendon. ST arises from the proximal-most portion of the tendon, the posterior head of BF arises from the distal part of the tendon. Muscle fascicles of BF course laterad and distad and end on the deep surface of a broad tendon of insertion (Fig. 1:b). This tendon attaches to the lateral intermuscular septum that separate the hamstring and quadriceps muscle groups and therefore attaches both to the shaft of the femur and the fascia lata. In

TY

' / i CRURAL - ' \ ' / FASCIA I,!

i )

Fig. 1. The anatomy and architecture of mm. biceps femoris (BFf and semitendinosus (ST) are shown in semi- diagramatic drawings in lateral (left) and posterior (right) views. Both BF and ST arise from the ischial tuberosity via an aponeurosis (a) that extends obliquely caudad and laterad. The anterior head of BF arises from its more lateral surface, the posterior head of BF and the proximal head of ST from its more medial surface. The anterior head of BF inserts into the deep surface of an extensive aponeurosis of insertion (bf that is continuous with the insertion of the posterior head and the crural

fascia. The proximal (STp) and distal (STdf heads of ST are separated by a tendonous inscription (cf. The distal head of ST inserts into a broad tendon (df that is continuous with the crural fascia. The sites for recording EMG activity used in this study are shown by the paired dots. The placement of dots does not necessarily reflect the position of the electrode tips relative to the muscle fas- ciles. In all cases EMG electrodes were surgically placed such that the electrode tips were oriented perpendicular to muscle fascicles, as described previously (English, '84).


the distal part of the thigh the tendon of insertion of BF attaches to the fascia lata, which is the superficial layer of deep fascia of the thigh, and also into the crural fascia, or superficial layer of deep fascia of the calf. The latter attachments to the crural fascia are mainly from the posterior head of BF.

Muscle fascicles of ST course obliquely distad and slightly mediad from the common tendon of origin and attach to the proximal surface of a tendinous inscription (Fig. l:c), which lies approximately parallel to the common tendon of origin of BF and ST. From the distal surface of this inscription, fascicles course distad and slightly mediad to insert onto the deep surface of a tendon of insertion (Fig. 1:d). This tendon attaches to the medial side of the fascia lata and the tendons of insertion of semimembranosus posterior, sar-

torius, and gracilis. It also has fairly extensive attachments into the medial part of the crural fascia. Thus, in addition to bony attachments to the thigh and leg, both BF and ST have extensive attachments into the crural fascia. The crural fascia envelops the leg posteriorly and is attached distally to the sides of the calcaneus.

Analysis of the gross fascicle architecture described above indicates that BF and ST are essentially groups of unipennate muscle masses that arise and insert mainly on tendinous surfaces. The masses in BF are arranged in-parallel, and those in ST are arranged serially. Examination of material prepared for demonstration of motor end plate sites (Fig. 2:mep) reveals a more complex organization of presumed muscle fibers. In both BF and ST the motor end plates are

Fig. 2. The motor end plate (mep) distribution on cat BF and ST is shown in this lateral view of a left hindlimb. The view is similar to that shown diagramatically in Figure 1 (left). Note that for both BF and ST, multiple, discontinuous rows of endplates are present. The tendonous inscription of ST is indicated (c). The dark color of the posterior head of BF is an artefact. After English (‘85), modified.


arranged in several discontinuous rows. The results of Ypey (‘78) in frogs and from studies of cat calf muscles from this lab (English and Letbetter, ’82a) indicate that motor end plates tend to be found in single rows near the center of muscle fibers. The results of microdissection of individual fascicles in both BF and ST reveal that the smallest dissecta- ble elements contain several AChE positive zones. No detailed digestion or impregnation techniques, such as employed by Richmond et al. (‘85), were used, so that whether this arrangement of end plates on BF and ST muscle fascicles reflects an in-series or tapered arrangement of fibers, or true multi- terminal innervation is not known. However, results of histological analysis (see below) are suggestive of the presence of tapered fibers, much as reported, for example, for cat splenius (Richmond et al., ’85).

Innervation patterns Figure 3 is a diagram of the patterns of

branching of the hamstring nerves. Detailed dissections were performed on the limbs of 12 adult cats. Some of these limbs (six) were from AChE material, but other dissections were made in unembalmed specimens. The

A B

Fig. 3. The main pattern of branching of the hamstring nerves is shown (A) in diagramatic form. The view is of a left nerve dissection, with the sciatic nerve lying ventral to the hamstring nerves. In B the main variant of this pattern is shown. Only the ABMB branches differ so that they are the only branches illustrated in B. The main difference is related to the position of branching of a small blood vessel. In the principal pattern (A) the MB branch point is found deep to this vessel. In the main variant (B) the branching takes place superficial to the vessel. SM has two anterior branches (a) and two posterior branches (p).

main branching pattern described in Figure 3A is found in 75% (9 of 12) of the dissections. It differs from the main variant (Fig. 3B) only in the position of the branch to the middle portion of biceps femoris (MBjl. In the principal branch pattern, the MB branch leaves the nerve to biceps femoris proximal to a small perforating blood vessel and enters the connective tissue on the deep surface of the muscle near the vessel. The branch courses in the connective tissue for approximately 0.5 cm before dividing into branches that course to the posterior aspects of the anterior head of BF, which we describe here as BFm. In the main variant of this pattern, the MB branch leaves the main nerve trunk distal to the vessel and is bound to the deep surface of the muscle considerably less by connective tissue than noted above. Upon superficial dissection, this MB branch pattern is far more prominent than the more frequently encountered pattern, and is almost certainly the branch described as infrequently encountered by Eccles et al. (‘57). The observed terminal distribution of the MB nerve branch was similar in all of the dissections performed.

The distribution of the AB and PB branches of the hamstring nerve and the nerve to the two parts of semitendinosus (STp and STdj are essentially the same as described by Bot- terman et al. (‘83a,b). The AB branch contains branches that are distributed to the anterior portion of the anterior head of BF (BFa); the PB branch that are distributed to the posterior head of BF (BFp). The branches to semitendinosus each contain several branches that are distributed to the muscle fibers lying proximal (STp) or distal (STd) to the tendinous inscription (Fig. 1).

The details of these innervation patterns were examined using the method of glycogen depletion. Figures 4-6 show the results of glycogen depletion experiments for BF and ST. Figure 4 shows histological results from different cats in which single branches to BF and ST were stimulated. In A, a group of fibers from a region of the border between BFa and BFm is shown. Most of the fibers in BFm (the stimulated area) are more glycogen

‘The nomenclature used to describe BF nerve branches will essentially follow that used by Botterman et al. (‘83a,b) and is an extension of Sherrington’s (‘10) use of AB and PB for the nerves to the anterior and posterior heads. The territories sup plied hy these branches will be denoted by subscripts. Thus the AB nerve branch innervates the BFa muscular territory.


Fig. 4. Photomicrographs of material prepared from glycogen depletion experiments (PAS reaction) involving BF(A-C) and ST(D) are shown. In panels A and B, the MB nerve branch was stimulated, resulting in depletion of muscle fibers in BFm. In A the distinct border between these innervation territories can be seen, but the presence of glycogen-poor fibers, both in BFa and BFm, is indicated by asterisks (*I. In B, taken from the middle of the BFm muscular subvolume, the same type of glycogen-poor fibers can be seen. Panel C is from an experi-

ment in which the AB branch was stimulated, resulting in depletion of muscle fibers in the BFa territory, which shows a “patchy” type of depletion. Non-depleted fibers are scattered in the BFa subvolume. Panel D shows an area of ST that includes fibers from STp, the tendonous inscription (c) and STd in an experiment where the STZ, compartment was stimulated. Note that glycogen depleted fibers are found only in STp. In all four panels a population of very small fibers can be identified (arrows).


Fig. 5. The results of several glycogen depletion experiments involving different branches to BF are summarized. At the top is a drawing of BF in lateral view indicating the positions of the cross-sections shown below. Each section is oriented such that the lateral or superficial surface of the muscle is placed to the left and the deep or medial surface is to the right.

poor than the fibers of BFa. Both BFa and BFm contain fibers that are glycogen poor but not glycogen depleted (Fig. 4A:*). This was a consistent finding, whether the AB or MB branch was chosen for stimulation. The glycogen poor fibers are distributed throughout both the region of depleted fibers and the nondepleted regions. Whether these are par- tially depleted fibers or fibers with intrinsi- cally small quantities of stored glycogen is not known. In reconstructing the BFa and BFm territories, the areas used were those where the majority of the fibers were glycogen depleted even though not all of the fibers in the “depleted” area could be identified as depleted. A region of BFm from the same

experiment as in Figure 4A, but which lies in the center of the depleted region, is shown in Figure 4B. Example of glycogen-poor fibers are indicated by asterisks (*I.

Figure 4 also demonstrates that the border between BFa and BFm is mainly very distinct. Figure 4C shows a view of a part of the border between BFa and BFm which was less distinct. In this case the AB branch was stimulated to produce glycogen depletion in BFa. Undepleted fibers of BFm are indicated in the lower right corner. Note that some areas of BFa contain truly undepleted fibers in addition to glycogen poor fibers. This “patchy” type of border was found in all BF depletion experiments but never accounted for more


G

Fig. 6. The results of glycogen depletion experiments involving different branches of ST are summarized. The format is similar to that of Figure 5 , Tn each cross-section thc more lateral surface of the muscle lies a t the top of the section and the medial surface lies toward the bottom.

than a very small proportion of the border between regions. In contrast, the border between BFm and BFp is uniformly distinct. No “patchy” regions are found. No evidence that the PB nerve branch innervated any fibers other than the posterior head of BF were obtained. Figure 4D shows the border between STp and STd regions. The position of the tendinous inscription is indicated. In no cases were glycogen depleted muscle fibers found on both sides of the inscription.

The panels of Figure 4 also demonstrate the differences in muscle fiber size noted in both BF and ST. In both muscles, a population of very small cells was noted (Figure 4: arrows). These fibers are fairly uniformly dis- persed among the larger fibers. The size and distribution of these fibers is very reminis- cent of the tapered muscle fibers of cat splenius muscle described by Richmond et al. (‘85).

Figures 5 and 6 are each a summary of several glycogen depletion experiments showing the distribution of each of the branches to BF and ST shown in Figure 3. These results essentially confirm those con- clusions made on the basis of dissections of nerve branches. The biceps femoris muscle is divisible into three parts, based on the described extramuscular nerve branches. The

portion known as the anterior head contains two such parts: BFa and BFm. The posterior head forms the third part BFp) (Fig. 5). The two parts of semitendinosus are innervated by different nerve branches (Fig. 6).

E lectromyography The activity of different parts of BF and ST

recorded during overground locomotion is shown in Figures 7 and 8. The data shown are from two different recording sessions from different cats, and reflect both the observed patterns of activation and the amount of variability in activation patterns. Figure 7 shows typical results of EMG recordings made from BFa, BFm, and BFp during nine different step cycles. Each trace begins and ends at the time of removal of the foot from the walkway, as determined from high-speed cinematography, and is denoted by upward pointing arrowheads. The time of foot placement in each trace is denoted by downward pointing arrowhead. The records contained in each trace represent rectified and averaged signals (see Methods, for technique). The three traces of each row were recorded simultaneously, i.e., they are from the same step cycle. Rows a-f are data from step cycles at moderate to fast walking speeds (0.9-1.3

170

A V A A

A. WM. ENGLISH AND 0.1. WEEKS

BFm - A V A - A V A

A V A -A V

A A V A - A V A

A V h

A V A

A d 'I A

P A V A

A V i -- A V A - A V A - A V A

A V

Fig. 7. The typical patterns of EMG activity recorded from different parts of BF during overground locomotion are shown. All data are from the same cat during a single recording session. Each row of traces shows the activity recorded from BFa, BFm and BFp during a single step cycle. The time of foot placement is indicated by an upward pointing arrow. The traces begin and end

msec-'), and rows g-i are from step cycles in which the animal was trotting (speeds 1.8- 2.2 msec-'). Data for galloping steps (not shown) show the same patterns as for walking and trotting steps.

In general, the EMG activity recorded from BFa began at or just subsequent to foot placement and continued until the time of foot liftoff. Although some variants were encoun-

A

A L b V A

A &. V A

A L d V A

A b e V A

A f A V A

A A

at the time of removal of the foot from the ground. The records are rectified and averaged signals. Rows a-f are from step c cles at moderate to fast walking speeds (0.9- 1.3 m-sec-. ) and rows g-i are from step cycles during which the cat was trotting (1.8-2.2 m-sec-'1. Scale bar=200 mseci0.5mv

Y

tered (e.g., Fig. 7:i), most intense activity was found early in the stance phase of the cycle. Occasional step cycles were found where BFa began activity before foot placement (Fig. 7:d), but the amount of this swing phase activity was never as prominent as noted in the ankle extensor muscles (e.g., Engberg and Lundberg, '69). Activity in BFm nearly al- ways showed a pattern that was similar to

CAT BICEPS FEMORIS AND SEMITENDINOSUS MUSCLES 17 1

that of BFa, except that its amplitude was consistently smaller. Variations were sometimes encountered (Fig. 7:e), but these were not common.

In contrast to BFa and BFm, the EMG activity recorded from BFp during stepping most consistently existed as two bursts. The largest burst occurred early in the cycle, usu- ally in the early swing phase. In some cats, this burst began at the time of foot liftoff, as shown in Figure 7. In other cats, this burst began rather consistently before foot liftoff. Examples of this latter pattern are shown in Figure 8. The second burst of activity was found around the time of foot placement. Sometimes it began slightly before placement and ended slightly after placement (Fig. 7: a,b,d), and other times it began and ended after the foot was on the ground (Fig. 7:c,e- i). In yet other instances it began and ended before the foot was placed (Fig. 8:a-f). The amplitude of this second burst was also vari- able. In some cycles it was much smaller

than the initial, swing phase burst (e.g., Fig. 7:b), but in others it was as large or larger in amplitude (e.g., Fig. 8:c). Thus the timing and amplitude of both bursts of activity in BFp during locomotion were less tightly tied to the kinematics of the step cycle than the activity of BFa or BFm.

Figure 8 shows examples of EMG activity recorded from the two heads of ST during overground locomotion. The format of the figure is the same as that of Figure 7, although the data are from a different cat. Also shown in Figure 8 is the activity recorded simultaneously from BFp. Traces a-e are of recordings made at fast walking speeds (0.8-1.0 msec-') while trace f is a recording made during a trotting step (1.8 msec-'). Activity in ST is generally found in two bursts, in a pattern similar to that described above for BFp. With a few exceptions (Fig. 8:a,d,e), the intensity of activity in the two heads of ST is similar, and is fairly similar to the intensity of activity recorded simultaneously in BFp.

Fig. 8. The patterns of EMG activity recorded from the two heads of ST and BFp during overground locomotion are shown in a format similar to that of Figure 7. All data are from a single recording session from a

different cat than Figure 7. Rows a-e show data from walking steps (0.8-1.0 m-sec-I), while row f is from a trotting step (1.8 m-sec- '1. Scale bar=200 msec/0.5mv


Although some variability in timing and relative amplitude is evident in the traces in Figure 8, the bursts of activity in ST and BFp are more closely tied to the kinematics of the step cycle than shown in Figure 7.

DISCUSSION

Recent studies (English and Letbetter, '82a, English and Weeks, '84; Letbetter, '74) have indicated that each of the ankle extensor synergists of cats is composed of smaller elements called neuromuscular compartments. Each such element is innervated by one of the naturally occurring first order muscular branches of its respective nerve and is composed of a unique aggregation of motor units. Studies of the central connections of muscle afferents and motoneurons in these compartments has indicated a central partitioning or "topographic weighting" of synaptic inputs, at least with respect to muscle spindle primary afferents (Lucas and Binder, '84; Lucas et al., '84; Vanden Noven et al., '86). Studies using the extramusuclar branches of the hamstring nerves have indicated a similar central partitioning of reflex inputs to BF and SM CBotterman et al., '83a; Hamm et al., '85) but not to ST (Botterman et al., '83b).

The present study has examined the anatomy and innervation territories of the same extramuscular nerve branches to two of the hamstring muscles: BF and ST. The results of microdissection and glycogen depletion analyses have indicated that innervation subvolumes exist that look like those described for the ankle extensors. The innervation volumes, termed BFa, BFm and BFp in biceps femoris and STp and STd in semitendinosus, are discrete aggregations of muscle fibers, with distinct borders and little or no suggestion of overlap. In BF these subvolumes are arranged in-parallel to the direction of pull of the muscle, but in ST they are arranged in-series, being connected through a tendonous inscription. However, despite this resemblance, a question remains as to whether or not these volumes are analogous to the neuromuscular compartments described for ankle extensors.

In the case of ST, indirect evidence argues that each head is not equivalent to a compartment. Bodine et al. ('82) have described a different histochemical profile in different parts of each of the heads of ST (their "superficial" and "deep") that, considering the architecture of this muscle, might correspond to volumes of muscle that lie parallel to each other. If the close relationship between motor unit and muscle fiber types that has been

demonstrated in cat ankle extensors (e.g., Burke et al., '73) also exists in ST, then ST motor units are not evenly distributed in either head. This would suggest that the heads of ST are not the smallest volume of the ST muscle containing a collection of single motor units, as the compartments of the ankle extensors are said to be (English and Weeks, '84). Further, preliminary results (Letbetter and English, '81) indicate that each head of ST is supplied by many more motoneurons than the compartments of ankle extensors (Weeks and English, '85). Con- sidering that the in-series arrangement of ST is probably the result of the development of the heads from separate anlagen, as has been observed in the mouse (Lance-Jones, '791, the possibility that each head contains more than one compartment seems reasonable.

In the case of BF no published histochemical data exist for the three volumes described above, but preliminary results of studies using the retrograde transport of horseradish peroxidase (Letbetter and English, '81) indicate that much as noted for ST, the number of motoneurons supplying each subvolume of BF is quite large when compared to neuromuscular compartments of ankle extensors. Thus, the available indirect evidence suggests that the hamstring muscle subdivisions are each composed of more than one neuromuscular compartment.

If this tentative conclusion is supported by further experimental evidence, then, by com- paring the results of the present study to those of Botterman et al. ('83a,b), one might think of biceps femoris as comparable to three different muscles: BFa, BFm, and BFp; and semitendinosus as one muscle with two heads: STp and STd. However, the localization of synaptic inputs onto motoneurons of BFa, BFm, and BFp and the relationship of EPSP amplitudes within parts of the motor nucleus of a similar muscle, SM, by Hamm et al. ('85) all are suggestive of a central partitioning that is every bit as great as that noted for compartments of cat ankle extensors. This degree of central partitioning in BF may be related to architectural constraints imposed on its proprioceptive control (e.g., Sacks and Roy, '82). The length infor- mation contained in muscle spindle afferent discharges may be quite different in different parts of BF during movement. Such diversity might require very precise central partitioning (Loeb, '82). It might be speculated that examination of the partitioning of synaptic inputs to motoneurons from the different naturally occurring branches of the AB, MB, and PB nerves could reveal an even greater


localization. The lack of central partitioning between heads of ST can be accounted for on the basis of its peculiar architecture. Since each head is not histochemically uniform, and since, based on the AChE staining and appearance of very small fibers in cross-sections, the architecture of the fibers is almost certainly more complex than that of the fascicles, and since the distributed insertion of ST (and BFp) into the crural fascia may im- pose biomechanical constraints on proprioceptive control of different parts of the muscle in a manner similar to that of BFa and BFm, the differential weighting of inputs to the presumed different motor unit population in each head bears consideration.

These interpretation of acute physiological data in terms of the present anatomical results are generally consistent with the EMG findings reported above. Activity in BFa and BFm is consistent with their acting as synergists during hip extension. The pattern most often noted is roughly similar to that reported for ankle extensor synergists (e.g., Engberg and Lundberg, ’69). However, the timing of the onset of BFa activity differs from the ankle extensor synergists in that no swing phase (El) period of activity is noted. The most common activity patterns of BFp and both parts of ST are consistent with their action as synergistic knee flexors during the early swing phase of the step cycle. The gen- erality of simultaneous activation of both heads of ST has been reported elsewhere (Murphy et al., ’81) and is consistent with the mechanics of its serial fascicle architecture. The activation of BFp and ST near the swing- stance transition may reflect their ability to produce both hip extensor and ankle extensor torques, presumably via their attachments into the crurial fascia. Estimates of ankle joint torques produced by these muscles at the joint angles found at the swing- stance transition are relatively smaller than joints in the step cycle (Wilke and Zajac, ’81). However, these estimates indicate substan- tial torques (5-8 kg-cm) when compared to those produced by soleus (6-7.5 kg-cm) at the swing-stance transition. During the swing- stance transition bursts, BFp and ST could be acting as potent synergists with other extensor muscles, including BFa and BFm.

Motor units in BFa and BFm, are quite clearly activated according to a pattern which is similar to that described as a typical “extensor” burst, but without the small swing- phase component. Motor units in BFp and both heads of ST are activated during locomotion according to a different pattern, which has been termed “bifunctional” (e.g.,

Engberg and Lunberg, ’69). We have obtained little substantive evidence that the activation of motor units in BF is any more or less independent than in the different heads of ST. This is not meant to imply that the finding of reflex partitioning in BF and the lack of such localization in the ST motor nucleus has no correlate during volitional movements. Examination of the activation of motor units at the single cell level may provide a better opportunity for examination of such correlations than the use of EMG data.

On the other hand, a clear difference was found in the relationship between the activation of BF and ST motoneurons and the kinematics of the step cycle. In recordings made from both BFa and BFm, a very consistent relationship was found between the onset, duration and relative intensity of EMG activity and the placement and removal of the hindlimb. In at least some of the recordings obtained from BFp and ST, considerably more variability in the time of onset and relative intensity of activity was observed. This was especially true with regard to the burst of activity found near the swing-stance transition. ‘l’he remarkable feature of this variability was its consistency across muscles. If activity in BFp began after foot placement, then activity in both heads of ST was also delayed.

One possible explanation for this finding is that the timing of EMG activation in “bifunctional’’ muscles such as BFp and ST during locomotion may be much more heavily influenced by sensory feedback from the hindlimb than that of “extensor” muscles such as BFa and BFm. Perret and Cabelguen (’76, ’80) have reported a variability in activation timing of ST during locomotion in decorticate cats, and have suggested that the timing and amplitude of the swing-stance transition burst may be strongly influenced by sensory feedback. Perret (‘83) found that the timing of ST discharge was greatly influenced by tonic exteroceptive stimulation during “fictive” locomotion in paralyzed decorticate cats, but that the timing of bursts of activity to “extensor” and “flexor” muscles was only very subtly affected. Thus, it is possible that during locomotion in intact cats, sensory input to the spinal cord may have profoundly different influences on the pattern of activation of different species of motoneurons.

This difference could arise from differences in the drive to the motoneurons by the central pattern generator for locomotion, as suggested by Perret (’831, whereby the drive to BFp and ST might be more heavily affected


by sensory feedback than the drive to BFa and BFm. Alternatively, the difference could be attributed to a marked difference in the nature of the sensory feedback itself. The data on firing patterns of muscle spindle afferents in a “bifunctional” muscle, sartorius, from Loeb and his colleagues (’85) are consistent with this speculation. Muscle spindle group Ia afferents show a wide range of different firing patterns during locomotion that are not particularly well correlated with spindle location in the muscle, and therefore with their perturbation during the step cycle. In contrast, similar data recorded from “extensors” or “flexors” are much more consistent (Loeb and Duysens, ’79; Loeb et al., ’85; Prochazka et al, ’77). However, whether the nature of the sensory feedback, the output of the central pattern generators for stepping, or some combination of the two is responsible for the observed variations in activation of BFp and ST remains to be tested.

ACKNOWLEDGMENTS

The authors are grateful to W.D. Letbetter who participated in some of the early experiments in this study, and whose influence almost certainly shaped our thinking. In addition B.R. Botterman, G.E. Loeb, F.J.R. Richmond, and especially D.G. Stuart all provided useful input during the gestation of this paper. This work was completed while supported by grants AM 19916, NS 15452, and NS 07127 from the USPHS and NIH core research grant PR 00165 to the Yerkes Re- gional Primate Research Center.

LITERATURE CITED

Bodine, S.C., R.R. Roy, D.A. Meadows, R.F. Zernicke, R.D. Sacks, M. Fournier, and V.R. Edgerton (1982) Architectural, histochemical, and contractile charac- teristics of a unique biarticular muscle: The cat semitendinosus. J. Neurophysiol. 48:192-201.

Botterman, B.R., T.M. Hamm, R.M. Reinking, and D.G. Stuart (1983a) Localization of monosynaptic Ia excitatory post-synaptic potentials in the motor nucleus of the cat biceps femoris muscle. J. Physiol. 388:355-377.

Botterman, B.R., T.M. Hamm, R.M. Reinking, and D.G. Stuart (1983b) Distribution of monosynaptic Ia excitatory post-synaptic potentials in the motor nucleus of cat semitendinosus muscle. J. Physiol. 338:379-393.

Burke, R.E. and P. Tsairis (1973) Anatomy and innervation ratios in motor units of cat gastrocnemius. J. Physiol. 234r749-765.

Burke, R.E., D.N. Levine, P. Tsairis, and F.E. Zajac (1973) Physiologic types and histochemical profiles in motor units of the cat gastrocnemius. J. Physiol. 234:723-748.

Cameron, W.E., M.D. Binder, B.R. Botterman, R.M. Reinking, and D.G. Stuart (1981) “Sensory partitioning” of cat medial gastrocnemius muscle by its muscle spindles and tendon organs. J . Neurophysiol. 46t32- 47.

Carlson, H. (1978) Histochemical fiber consumption of lumbar back muscles in the cat. Acta Physiol. Scand. 103:198-209.

Eccles. J.C.. R.M. Eccles, and A. Lundbertr (1957) The convergence of monosynpatic excitatory afferents onto many different species of a-motoneurons. J. Physiol. 137:22-50.

Edstrom, L. and E. Kugelberg (1968) Histochemical composition, distribution of fibres, and fatiguability of single motor units. J . Neurol. Neurosurg. Psychiatr. 31:424-433.

Engberg, I. and A. Lundberg (1969) An electromyographic analysis of muscular activity in the hind limb of the cat during unrestrained locomotion. Acta Phys- iol. Scand. 75:614-630.

English, A.W. (1983) Locomotor activity of single motor units in compartments of cat lateral gastrocnemius muscle. Abstr. SOC. Neurosci. 9:360.

English, A.W. (1984) An electromyographic analysis of compartments of cat lateral gastrocnemius muscle during unrestrained locomotion. J. Neurophysiol. 52114- 125.

English, A.W. (1985) Limbs vs. jaws: Can they be compared? Amer. Zool. 25:351-363.

English, A.W. and W.D. Letbetter (1981) Intramuscular “compartmentalization” of the cat biceps femoris and semitendinosus muscles: Anatomy and EMG patterns. Abstr. Soc. Neurosci. 7.557.

English, A.W. and W.D. Letbetter (1982a) Anatomy and innervation patterns of cat lateral gastrocnemius and plantaris muscles. Am. J. Anat. 164:67-77.

English, A.W. and W.D. Letbetter (198210) A histochemical analysis of identified compartments in cat lateral gastrocnemius muscle. Anat. Rec. 204:123-130.

English, A.W. and 0.1. Weeks (19841 Compartmentaliza- tion of single muscle units in cat lateral gastrocnemius. Exp. Brain Res. 56361-368.

English, A.W. and 0.1. Weeks (1985) Reliability of electromyographic recording: Can we (cross) talk? Anat. Rec. 211:58A.

Gonyea, W.J. and G.C. Ericson (1977) Morphological and histochemical organization of the flexor carpi radialis muscle in the cat. Am. J. Anat. 148:329-344.

Hamm, T.M., W. Koehler, D.G. Stuart, and S. Vanden Noven (1985) Partitioning of monosynaptic Ia excitatory post-synaptic potentials in the motor nucleus of the cat semimembranosus muscle. J. Physiol. 369r379- 398.

Kugelberg, E. (1973) Histochemical composition, contraction speed, and fatiguability of rat soleus motor units. J. Neurol. Sci. 20:177-198.

Lance-Jones, C. (1979) The morphogenesis of the thigh of the mouse with special reference to tetrapod muscle homologies. J . Moi-phol. 162:275-310.

Letbetter, W.D. (1974) Influence of intramuscular nerve branching on motor unit organization in medial gastrocnemius muscle. Anat. Rec. 178:402.

Letbetter, W.D. and A.W. English 11981) Th:<relationship between peripheral intramuscular compartments” and spatial arrangement of biceps femoris and semitendinosus motor nuclei in the cat lumbar spinal cord. Abstr. SOC. Neurosci. 7:557.

Loeb, G.E. (1982) Task groups - a proposed functional unit for motor control. Abstr. SOC. Neurosci. 8:947.

Loeb, G.E. and Duysens, J. (1979) Activity patterns in individual hindlimb primary and secondary muscle spindle afferents during normal movements in unrestrained cats. J. Neurophysiol. 42420-440.

Loeb, G.E., J.A. Hotter and C.A. Pratt (1985) Activity of spindle afferents from cat anterior thigh muscles. I. Identification and patterns during normal locomotion. J. Neurophysiol. 54.549-565.

Lucas, S.M. and M.D. Binder (1984) Topographic factors


in the distribution of homonymous group Ia afferent input to cat medial gastrocnemius motoneurons. J. Neurophysiol. 5150-63.

Lucas, S.M., T.C. Cope and M.D. Binder (1984) Analysis of individual Ia EPSPs in an homonymous motoneuron pool with respect to muscle topography. J. Neurophys- iol. 5:64-74.

Murphy, K., R.R. Roy, and S.C. Bodine (1981) Recruit- ment of the proximal and distal portions of the cat semitendinosus during running and jumping. Med. Sci. Sports Exercise 13:127-128.

Perret, C. (1983) Centrally generated pattern of motoneuron activity during locomotion in the cat. pp. 405- 422 IN: Neural Origin of Rhythmic Movements, A. Roberts and B. Roberts, eds., Sac. Exp. Biol.

Perret, C. and J.M. Cabelguen (1976) Central and reflex participation in the timing of locomotor activations of a bifunctional muscle, the semitendinosus, in the cat. Brain Res. 106:390-395.

Perret, C. and J.-M. Cahelguen (1980) Main characteris- tics of the hindlimb locomotor cycle in the decorticate cat with special reference to bifunctional muscles. Brain Res. 187:333-352.

Rochazka, A., R.A. Westerman and S.P. Ziccone (1977) Ia afferent activity during a variety of voluntary movements in the cat. J. Physiol. 268:423-428.

Richmond, F.J.R. and V.C. Abrahams (1975) Morphology and enzyme histochemistry of dorsal muscles of the cat neck. J. Neurophysiol. 38:1312-1321.

Richmond, F.J.R., D.R.R. MacGillis and D.A. Scott (1985) Muscle-fiber compartmentalization in cat splenius

muscles. J . Neurophysiol. 53:868-885. Russell, C.J., D.B. Dunbar, D.S. Rushmer, J.M. Mac-

Pherson, and J.O. Phillips (1982) Differential activity of innervation suhcompartments of cat lateral gastrocnemius during natural movements. Ahstr. Sac. Neu- rosci. 8:948.

Sacks, R.D. and R.R. Roy (1982) Architecture of the hindlimb muscles of cats: functional significance. J . Mor- phol. 173:185-195.

Sherrington, C.G. (1910) Flexion-reflex of the limb, crossed extension reflex and reflex stepping and stand- ing. J . Physiol. 40:28-121.

Vanden Noven, S., T.M. Ham and D.G. Stuart (1986) Partitioning o f monosynaptic Ia excitatory post-synaptic potentials in the motor nucleus of the cat lateral gastrocnemius muscle. J. Neurophysiol. 55.569-587.

Weeks, 0.1. and A.W. English (1985) Compartmentaliza- tion of the cat lateral gastrocnemius motor nucleus. J. Camp. Neural. 235.255-267.

Wilke, R.W. and F.E. Zajac (19811 Isometric torque produced by the cat hamstring muscles about the ankle as a function of hindlimb position. Abstr. Sac. Neu- rosci. 7;684.

Wineski, L.E. and S.W. Herring (1983) Innervation and activity patterns in the masseter and zygomatico-man- dibularis muscles of the pig. Anat. Rec. 205:217A.

Ypey, D.L. (1978) A topographical study of the distribution of endplates in the cutaneous pectoris, sartorius, and gastrocnemius muscles of the frog. J. Morphol. 155:327-348.

Documents

An anatomical and functional analysis of cat biceps femoris and semitendinosus muscles