JOURNALOF

NEUROPHYsOLOGY

Vol. 41, No. 5, September

1978. Printed

in U.S.A.

Forces Produced by Medial Gastrocnemius Soleus Muscles During Locomotion in Freely Moving Cats B. WALMSLEY,

J. A. HODGSON,

Laboratory of Neural Control, Disorders and Stroke, National

SUMMARY

AND

AND

and

R. E. BURKE

National Institute of Neurological and Communicative Institutes of Health, Bethesda, Maryland 20014

CONCLUSIONS

5. During the step cycle the peaks of both I. Force transducers designed for chronic MG and SOL force profiles occur before the end of the yield (E2) phase, as both muscles implantation were placed on the individual tendons of the medial gastrocnemius (MG) undergo active lengthening. In contrast, and soleus (SOL) muscles in the same hind- peak MG and SOL forces during vertical jumping coincide with the end of active limb of adult cats. Electromyographic (EMG) and kinesiological data were also lengthening and the onset of ankle extensor recorded. Force, EMG, and movement data shortening. 6. The results are consistent with the obtained from intact, freely moving cats illustrate the division of labor between the hypothesis that the inherent stiffness of mixed MG and its slow-twitch synergist active muscle during the step cycle is an important factor in the control of force outSOL during posture, treadmill locomotion, put from hindlimb extensor muscles in and jumping. 2. Throughout a wide range of speeds, locomotion. The division of labor between from slow walk (0.6 m/s) to fast run (3 m/s), MG and SOL and the absolute force levels the SOL develops approximately the same required from the MG during the full range peak force (between 1.6 and 2.0 kg wt, of hindlimb movements in posture, locomodepending on the animal) while the average tion, and jumping appear to be precisely MG force varies over a threefold range matched to the very different characteristics of the motor-unit populations composing (from about 0.6 to 2.0 kg wt). However, even during fast running, the SOL can these synergistic muscles. produce as much, or more, peak force as the MG. INTRODUCTION 3. In quadrupedal standing, the SOL can Studies of the neural control of locomoat times produce as much force as during tion necessarily involve some quantitation locomotion, but on the average, SOL output of motor output. Although this is usually is somewhat less (about 1.2- 1.4 kg wt). In done in terms of the resultant movement contrast, the MG usually produces no more itself or of patterns of electrical activity than 0.5 kg wt of force during quiet standing. in selected muscles, Manter (22) in 1938 4. Vertical jumping, up to a maximum measured the force vectors produced by height of 120 cm, involves peak MG forces hindlimbs of cats during locomotion. Using up to about 9 kg wt (average approximately these data, Grillner (18) made estimates of 8.0 kg wt), which can be lo-15% greater the forces developed by the triceps surae than the maximum isometric force developed muscles during walking and standing. bv the same muscle during fused tetanic Burke, Rymer, and Walsh (10) extrapolated contraction (stimulus rate X00 Hz) at Grillner’s results once more to arrive at optimum length. estimates of the force contributed by the Received for publication March 3, 1978. MG muscle alone during locomotion in 0022-3077/78/0000-0000$01.25

Copyright

0 1978 The

American

Physiological

Society

1203

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on August 7, 2018. Copyright © 1978 American Physiological Society. All rights reserved.

1204

WALMSLEY,

HODGSON,

order to compare these with the properties of MG motor units. In order to verify these indirect force estimates, we have developed a chronically implanted force transducer system to measure directly the force produced by individual ankle extensor muscles in unrestrained moving cats. The results illustrate the division of labor between the MG and its slow-twitch synergist, the SOL, during postural activity, walking, and running on a treadmill and vertical jumping. The SOL muscle of the cat often develops as much force during quiet standing as it does throughout the entire range of locomotion from slow walk to fast run. In contrast, MG generates only small forces during standing, but contributes a major fraction of the wide

A

FORCE

LEAD

TOP

WIRES

AND BURKE

dynamic range-of peak force output required during locomotion at different speeds, and particularly during jumping. METHODS

Force transducer manufacture and calibration The transducer substrate was cut from stainless steel sheet (316 alloy) and hand filed into the rounded letter E shape illustrated in Fig. 1A. This shape is a modification of the “belt buckle” transducer design suggested some years ago by Salmons (28). A semiconductor strain gauge (BLH type SPBl-12-35) was then bonded along one side of the substrate with light pressure, using epoxy cement (Devcon EK-20). Tefloninsulated stranded stainless steel wires (Bergen Wire Rope Co. BWR 3.48) were soldered to the

TRANSDUCER

+

VIEW

TENDON t

SIDE

STAINLESS

STEEL

~0~5-l~Omm

VIEW

B

BRIDGE OUTPUT (volts)

FORCE

(Kg.wt)

FIG. 1. A: photograph of the transducer assembly (en face view above), indicating placement of semiconductor strain element, lead wires, and position of tendon (shading). Lower diagram shows longitudinal section with tendon (stippled) threaded through the gauge. B: calibration curves for typical gauge showing bridge voltage output (ordinate) versus force (abscissa) during static test (weight on string) and during twitch and tetanus of MG muscle with same gauge in situ on MG tendon (see text).

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on August 7, 2018. Copyright © 1978 American Physiological Society. All rights reserved.

MUSCLE

FORCE

DURING

strain gauge, and a layer of epoxy was placed over the joints for strain relief. (The Teflon insulation was previously etched near the ends with fluorocarbon etchant to ensure that the epoxy would adhere to it.) The entire assembly was then given a conformal coating of Parylene C (15 pm thickness) to moisture-proof the gauge and to improve biocompatibility (2 1). A layer of Silastic rubber (Dow-Coming medical adhesive silicone type A) was then coated over the transducer and two slits cut in the side (opening of the E) to allow the gauge to be placed on the intact tendon. The purpose of the Silastic rubber coating (not shown in the figure) was primarily to keep the transducer in place on the tendon. The transducer operates by measuring the strain in the stainless steel substrate induced by tension in the tendon. The gauge was used in a simple DC Wheatstone bridge arrangement. (The nominal resistance of the gauges used was 350 a.) A static calibration for the transducer was obtained by placing the gauge on a string and suspending various weights. Figure 1B shows the static calibration curve for one transducer. Although semiconductor strain gauges are inherently nonlinear, the dimensions of the stainless steel substrate were chosen such that the transducers were reasonably linear over the force range of interest (up to 10 kg wt). Adaptation of the design for larger forces and different tendon dimensions should be possible. The linear least-squares regression line (of the form y = ax) on the graph in Fig. 1B shows that this particular gauge was quite linear up to 10 kg wt. It was found that the calibration curve for a given transducer depended to some extent on the thickness of the calibrating string. Each calibration was therefore checked in a terminal experiment under pentobarbital anesthesia with the gauges in situ on their tendons. To do this, the appropriate tendon was cut distally to the transducer and tied to a Grass strain gauge (type PT-10). The muscle nerve was then dissected free and stimulated. A calibration curve was obtained for both graded single twitches and tetani, as shown in Fig. 1B. In addition, a static calibration was obtained by suspending weights from the tendon. It was concluded that the static measurements produce an adequate calibration curve for each transducer. The effect of temperature on the transducer was also studied. Although the strain element resistance was quite temperature sensitive, the gauge factor varied negligibly over the range tested (lo-5OOC).

Implanting Force transducers and EMG electrodes were implanted in the left hindlimb under aseptic conditions. The SOL tendon is clearly separable

LOCOMOTION

1205

from those of LG and MG and the SOL transducer was slipped onto it, in series with SOL alone. The MG and LG tendons are formed from broad aponeuroses, but the tendon fibers that transmit force from each muscle can be identified at a point 3-4 cm proximal to the calcaneus. They were split from this point to the calcaneus and the MG transducer was slipped onto the MC portion, as in Fig. 1A. Bipolar EMG wires (Teflon insulated, stranded stainless steel) were inserted into the central portions of the MG and SOL muscles and kept in place with sutures. The MG electrodes were placed into regions of the muscle in which the distribution of histochemical muscle fiber types is representative of the whole muscle (9). All lead wires were then passed subcutaneously to a connector on the cat’s back, and a ribbon cable was used to carry the signals to the recording apparatus.

Recording The cats were trained to walk and run on a treadmill and to jump vertically from the floor onto a table in order to receive food rewards. The EMG and force data were recorded onto FM tape, and a strobed (60 Hz) videotape record was taken of the cat’s movements. The FM and videotapes were synchronized with the use of an IRIG-B code timer whose output was recorded onto both tapes. In order to determine joint angles and muscle lengths, reflective spots were placed over the iliac crest, great trochanter, lateral malleolus at the ankle, and the distal foot. Joint angles were measured using these spots and muscle lengths were then calculated from these measurements taken from successive single frames of the video record, using the trigonometric method of Goslow et al. (16). (Measurements of the distances between pivot points and the points of origin and insertion of the muscles were taken when the cat was sacrificed.) In later experiments, length gauges similar to those used by Prochazka et al. (26) were implanted along the MG muscle, from femoral condyle to calcaneous, to give a continuous electrical readout of muscle length. RESULTS

A total of seven cats was studied. Of these, three cats were implanted with separate force transducers and EMG wires for the MG and SOL muscles in the same hindlimb, four cats were implanted with force transducers and EMG wires for the MG muscle only. One of the latter group was also implanted with a length gauge along the MG muscle. The animals tolerated the

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on August 7, 2018. Copyright © 1978 American Physiological Society. All rights reserved.

1206

WALMSLEY,

tt MG FORCE

2

(Kg.wt)

o E

HODGSON,

tt

AND

tt

BURKE

tt4

t4

t4

14

tt

1

I

I

500

I

1

I

I

I

I

I

1

msec

FIG. 2. Records from 2%kg cat during slow (left set, treadmill speed 0.8 m/s) and fast (right set, speed 1.6 m/s) walking. MG and SOL force records from independent gauges in same leg. MG and SOL electrical activity (EMG) records show raw data (lower trace in each pair) and rectified-integrated traces (above). Integrator (Paynter) filter time constant was 0.05 s; EMG data prefiltered (high pass) at 30 Hz. Arrows at top show footfall (L) and foot lift (T) of instrumented left hindlimb from videotape records synchronized to analog tape data (see METHODS).

transducers well and were judged to walk normally about 1 wk after implantation. With the final transducer design, the tendons showed no evident damage even after 8 wk of implantation. MG and SOL forces

during locomotion

Figures 2, 3, and 4 show records taken from a 2.8-kg cat in which force transducers and EMG wires had been implanted for both the MG and SOL muscles of the left hindlimb. Figure 2 shows the forces produced by these muscles during a slow (left) and a fast (right) walk. The peak force produced by SOL was the same (1.8 kg wt) for the two different treadmill speeds. However, for the slow walk there was a plateau on the falling phase of the force record. A similar profile was found by Manter (22; his Fig. 12) for the calculated ankle torque required during the stance phase and by Grillner (18; his Fig. 5) who converted these torques to expected Achilles tendon force. This plateau was correlated with a prolongation of the electrical activity of the muscle, as can be seen in the traces of integrated EMG. During the slow walk the MG muscle produced, during the peaks, about 40% (0.7 kg wt) of the peak SOL force, whereas during the fast walk this proportion increased to about 80% (1.3 kg wt).

Figure 3 illustrates in greater detail two individual step cycles taken from a fast run at 3.0 m/s. The records on the left show a typical step cycle at this speed in which MG produced more peak force (1 lo%, or about 2.0 kg wt) than SOL. The EMG activity in both muscles appeared about 50 ms before foot contact, as described by Engberg and Lundberg (14). Foot contact coincided with an abrupt increase in muscle force. The length changes calculated from joint angles for MG and SOL during the step cycle were somewhat different, since the MG muscle spans both the ankle and knee joints, whereas SOL spans only the ankle. However, the peak forces in both muscles occurred before the end of the yield (E2) phase (the period of increasing muscle length after foot contact). Thus, the magnitude and timing of peak MG and SOL forces are not determined by the onset of muscle shortening. Figure 3, right, shows an unusual step cycle during the same run, which appeared on the videotape to be an attempt by the cat to convert to a gallop. The swing phase began as usual (left) and at the time of the first extension phase (E,), a brief burst of EMG was seen in the SOL muscle. However, the foot did not extend to contact the

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on August 7, 2018. Copyright © 1978 American Physiological Society. All rights reserved.

MUSCLE

4

MG EMG

w

e

DURING

1207

LOCOMOTION

4

> g

140

s

! ooo pa.

00 0°0

O”O a

zs

5=->

FORCE

0

70

0 O.0

O

0

KNEE ANKLE

0 0

0

O

l o l 00o8

IOOmsec FIG. 3. Records of two step cycles during fast run (treadmill speed 3 m/s) in same cat as Fig. 2. At the bottom are graphs of knee and ankle joint angles, measured from videotape record (frame rate 60 Hz), and estimated corresponding lengths of MG and SOL, calculated trigonometrically (see METHODS). Step cycle on left was typical of this speed of running; that on the right was atypical (see text).

treadmill, but remained in a flexed position while the contralateral (right) hindlimb continued in contact. During this time another short EMG burst was seen in SOL. The final EMG burst in MG started as usual about 50 ms before left foot contact and there was a large increase in the peak force produced by MG (3.4 kg wt). In contrast, the EMG of SOL was absent prior to foot contact, with a concomitant delay in SOL force production as compared to MG. The muscle lengths during stance in this step were similar to the other steps in the sequence (left records) except that the amount of yield was somewhat greater for both MG and SOL. The implanted transducers detected virtually no passive force in either MG or SOL during externally imposed movements of knee and ankle through the full physiological range when the animals were under anesthesia. In Fig. 3, when the foot accelerated

forward during the swing phase (from the start of each trace to the time of foot touchdown), little passive force was evident in the records from the electrically inactive MG or SOL. Thus, the muscles themselves did not appear to contribute significantly to the passive forces about the ankle joint, which can be quite large near the limits of joint movement in the cat (Fig. 4 in Ref. 18). MG and SOL forces during vertical jumping Figure 4 illustrates a vertical jump made by the cat from the floor onto a table 120 cm high. Preparation for the jump began with the cat crouching, as indicated by the decreasing ankle and knee angles. During this period the MG and SOL muscles underwent an active lengthening contraction. The force in MG then rose rapidly during the launching phase to a peak of 9.0 kg wt, which was in

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on August 7, 2018. Copyright © 1978 American Physiological Society. All rights reserved.

WALMSLEY,

1208

MG FORCE (Kg.wt)

SOL FORCE (Kg.wt)

EMG

bi 2

5

0 2 , 0 E

r I -

IO

-



MG A SOL A

5-z A

I AII

II

u

HODGSON,

I

I

I

200msec

I KNEE ANKLE

l

0

.

AND

BURKE

is changed in opposite ways by flexion or extension at either joint. Further, the MG not only contributes its own force to the calcaneus but also can transmit forces generated by the knee extensors by acting as a stiff linkage between femur and calcaneus. The use of the video system and the trigonometric method for determining muscle length was found to be inadequate for studying jumps since the animal was not as well visualized as during treadmill locomotion. Therefore, one cat was implanted with a length gauge (see METHODS) along the MG muscle, which gave a continuous electrical readout of muscle length, as shown in the lower traces of Fig. 5. In addition, the force transducer used in this cat was linear up to nearly 12 kg wt. Figure 5 shows the results for this cat jumping onto a table at three different heights. The MG muscle underwent active lengthening during the launching phase of all jumps but unlike the yield phase of locomotion (Fig. 3), the peak of the force trace coincided with the peak A.

4. table 120 3. Format out of the could not FIG.

Records from a vertical jump from floor to cm high in same cat as shown in Figs. 2 and as in Fig. 3. Just after liftoff the cat moved camera field and subsequent limb positions be defined.

fact greater (15%) than the 7.8 kg wt maximal isometric tetanus developed by the same muscle in the terminal experiment. The EMG activity decreased abruptly about 20 ms prior to liftoff, as found by Zomlefer et al. (34). The peak force developed by SOL during the launching phase of the jump was 1.5 kg wt, which was somewhat less than that developed during locomotion (1.8 kg wt; Figs. 2 and 3). The SOL force decreased before that of MC, probably because of the rapid shortening of SOL, and dropped to zero approximately 40 ms before liftoff despite continued electrical activity in this muscle. Using the videotape records of joint angles, it appeared that the MC muscle length during the launching phase was almost constant until about 30 ms before liftoff. This again reflects the difference between the two-jointed MG muscle and the single-jointed SOL muscle. Since it snans both knee and ankle. the MG length

45cm

k MG FORCE (Kg-W 1

1

I

I

I

I

500

msec

FIG. 5. Records from another 2%kg cat showing MG force and EMG (rectified and integrated) plus readout of a length gauge implanted along the MG IllUSCk (See METHODS). Three different jumps to heights indicated above each set of records are illustrated. Length scale calibrated in situ in a terminal experiment; zero point corresponds to MG length with knee and ankle both 90”.

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on August 7, 2018. Copyright © 1978 American Physiological Society. All rights reserved.

MUSCLE

FORCE

DURING

Stand

1

1

I

I

I

I

I

500msec

FIG. 6. Records of MG and SOL force and EMG (raw data plus rectified and integrated) from another cat (2.9 kg) during a sequence of movements during preparation for (step), and landing after (land-stand), a vertical jump to a table 70 cm above the floor. Stand portion during quadrupedal standing while eating food reward.

of the length trace (i.e., when the muscle was at its greatest length during the launching phase) for all three jumps. The peak force developed during the highest jump was 8.7 kg wt, which was 10% greater than the maximal isometric tetanus at optimum length for this muscle (7.9 kg wt) during the POSTURAL

WALK

LOCOMOTION

1209

terminal experiment. The optimum muscle length for the latter corresponded to a muscle length of about +5 mm on the scale used in Fig. 5. The changes in MG length in this cat during jumping were greater than those estimated from videotape records in Fig. 4. Whether this is a true difference between animals or simply represents more accurate estimates with the length gauge is unknown, although the latter seems more likely. The division of labor between MG and SOL is dramatically illustrated in a sequence of movements taken from yet another cat and shown in Fig. 6. The peak force generated by SOL was virtually the same (about 1.9 kg wt) for all four movements; stepping, jumping, landing, and standing. The MG muscle, on the other hand, demonstrated a wide range of peak force outputs during the same movements, from 300 to 400 g during standing to nearly 8 kg wt in the jump. The traces of integrated EMG show a concomitant difference in the dynamic ranges of the two muscles. The peaks of integrated EMG in SOL were similar for all four activities (although somewhat smaller during standing), whereas those for MG varied over a much wider range. In order to compare quantitatively the forces produced by MG and SOL for different movements, the plots shown in Fig. 7 were constructed using data from the same GALLOP

RUN

JUMP

2 SOL FORCE ( Kg.Wt)

0 0

0 00

0

8 MG FORCE FIG. 7. Plot of same cat shown apparently steady muscle. The single

SOL (ordinate) versus in Figs. 2-4. Postural quadrupedal standing. gallop point was taken

(Kg.wt)

MG (abscissa) forces during standing, locomotion, and jumping in points (A) denote instantaneous forces during several periods of Points during locomotion and jumping (0) are peak forces in each from the step cycle illustrated in Fig. 3, right.

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on August 7, 2018. Copyright © 1978 American Physiological Society. All rights reserved.

1210

WALMSLEY,

HODGSON,

IO

0

r

SOL

.. . .. .. .. .. ..

MG

FORCE ( Kg.wt

1

POSTURE

WALK

RUN

JUMP

FIG. 8. Mean values of MC and SOL force and rectified-filtered EMG during full output range of hindlimb movements. Postural points during quadrupedal standing averaged for over 40 measurements; jump averaged for five ju .mps to 120 cm. Points during locomotion averaged from 20 to 65 step cycles. Same cat as used for Fig. 7 but different set of data points. Integrated EMG means plotted on arbitrary linear scale matched to show correspondence between mean force and EMG during locomotion. Note, however, divergence between EMG and force means during standing and jumping. For these measurements, EMG records were rectified and low pass (RC) filtered at 20 Hz (Krohn-Hite filter, model 3322R).

cat used for Figs. 2,3, and 4. Figure 7 shows the peak forces produced by SOL (ordinate) plotted against the forces produced by MG (abscissa) at the same time, for a wide range of activities. The “postural” points were selected during periods of steady force production in both muscles. All other points represent the peak forces produced during walking and running at various speeds, and jumping to 120 cm. There was a wide scatter in the postural points, but in a majority of examples SOL produced considerable force at the same time that MG produced little or none. There were, however, also some periods during quadrupedal standing in which MG produced more force than SOL. In progressing from walking to running, MG took an increasingly dominant role in peak force production, while the force produced by SOL during the jumps was slightly less than that during walking and running, whereas that produced by MG was up to 4 times greater. The apparent fall off in SOL

AND

BURKE

force output during jumping probably reflects the rapid shortening of this singlejoint muscle at a time when the MG is either virtually isometric (see Fig. 4) or still lengthening (Fig. 5) due to its origin above the knee. The division of labor between these two muscles is summarized by Fig. 8 in which the mean peak force developed by each muscle (solid symbols, continuous lines) and the associated peak of integrated EMG records (open symbols, dashed lines) are plotted for a variety of activities arranged in order of increasing effort (from standing to jumping). These data show that, overall, the mean peak force output and the mean values for peaks of integrated EMG from the SOL muscle hardly change as locomotion speed increases from slow walk to fast run. Both indexes of activity in MG, however, change over a twofold range during locomotion at increasing speeds. Clearly, the SOL contributes the major fraction of force during quasi-isometric standing although, on the average, its force production and integrated EMG are not as large as during locomotion. It is useful to note that although peak SOL force changes little as locomotion speed increases, nevertheless the SOL muscle produces more peak force than MG, up to fast speeds of running (3 m/s). Although there is relatively good correspondence between the average force and integrated EMG values, it cannot be concluded that integrated EMG records accurately predict force production because the relation between instantaneous (and even peak) EMG and force was highly variable (see, e.g., Figs. 2 and 6). Clearly, a detailed correlation between EMG and force developed by a muscle must also take account of length changes occurring in that muscle. DISCUSSION

The present results demonstrate that when the forces from two contrasting triceps heads are individually measured, there is a distinct difference in the division of labor between the slow-twitch SOL and the mixed (nominally “fast twitch”) MG. Quadrupedal standing can at times result in as much SOL force (up to 1.8 kg wt in a 3-kg cat) as that muscle produces under any

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on August 7, 2018. Copyright © 1978 American Physiological Society. All rights reserved.

MUSCLE

FORCE

DURING

conditions, while the MG typically produces only a few hundred grams during quiet standing (Figs. 6, 7, and 8). Progressively increasing speeds of locomotion require concomitantly greater total ankle extensor forces, and this is evidently supplied by MG as well as by the lateral gastrocnemius (LG), with some possible contribution also from the plantaris (see Refs. 17, 31). The “fast” triceps heads, MG and LG, are both mixed muscles composed of similar motor-unit populations (as judged by both physiological and histochemical studies; Ref. 8), and they presumably act similarly during locomotion and jumping. The observed range of MG force output modulation, from quadrupedal standing to maximum vertical jumping, is greater than lo-fold. The LG dynamic range is presumably similar. In contrast, for the same range of movements, average force output of SOL varies only over a less than twofold range. In an unpublished thesis, Yager (32) described measurement of forces produced by the combined ankle extensor muscles in the cat using an implanted belt-buckle transducer placed on the whole Achilles tendon (see also Refs. 2,28). He found that the forces recorded from the Achilles tendon during stepping were more or less linearly related to the EMG activity recorded over the surface of the triceps surae muscles, but made no attempt to record force production from individual heads of the triceps surae and no absolute force data were reported. The present observations on the contrasting force output patterns of MG and SOL are entirely in accord with recent EMG studies of Smith and co-workers (29), who examined quantitative electrical activity in SOL and LG during a wide range of movements in unrestrained cats. In that study, the EMG records from LG were obtained from the lateral edge of the muscle where the muscle fiber population contained a higher percentage (90%) of type FG fibers (typical of motor units with large force output, rapid fatigue and, probably, high functional threshold, the type FF units; see Ref. 8) than that of the LG muscle as a whole (66% FG fibers; Ref. 1). This apparent bias in the EMG results probably accounts for the wider (fourfold) range in peak LG EMG found by Smith et al. (29) for locomotion between 0.5 and 3.0 m/s, as compared to the

LOCOMOTION

1211

present range (threefold) over the same speeds (Fig. 8). Smith et al. (29) found a narrow (< 1.5-fold) range of average SOL EMG modulation from standing to jumping, as was true in the present study (Fig. 8). Records of rectified and integrated EMG represent some function of muscle fiber activation and have been used (e.g., Ref. 29) to infer variations in muscle force output. The relation between EMG and force output may be reasonably linear over a restricted range of movements (e.g., treadmill walking at various speeds, Fig. 8; see also Ref. 32), but there is considerable scatter for individual measurements. The methods for EMG signal processing are not standardized and the signal itself depends on electrode configuration and on the characteristics of the local population of muscle fibers “seen” by the electrode, which may or may not be representative of the overall fiber population in the muscle of interest. Finally, the EMG alone provides no reference points for estimating absolute force. For all these reasons, direct force measurements from individual muscles seem advantageous for making inferences about muscle and motor pool usage during normal movement. However, the interpretation of the force records is also not simple. At a given muscle length, active muscle force is graded by recruitment and derecruitment of motor units and modulation of firing frequency of the active units. The well-known dependence of muscle force on isometric length is also important in the interpretation of records during postural (i.e., quasi-isometric) actions, and this effect may play a considerable role in modulating SOL forces over the range of observed ankle angles (Fig. 3 of Ref. 27; see also Ref. 18). However, in interpreting the forces developed during dynamic movements such as walking, running, and jumping, care must be taken in drawing conclusions based on the isometric force-length relationship for particular muscles. The force records obtained during walking and running indicate that the peak MG and SOL forces occur during the yield (E2) phase of the step cycle when the muscles are still lengthening (16). It is known that active lengthening may produce marked enhancement of force output from muscles (e.g., Ref. 1l), but this

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on August 7, 2018. Copyright © 1978 American Physiological Society. All rights reserved.

1212

WALMSLEY,

HODGSON,

effect (at least in the cat SOL) is also dependent on activation frequency (20). Even if the muscle is operating over a length range equivalent to the relatively flat portion of the isometric tetanus force-length curve (30), stretch of the active muscle, such as occurs during the yield phase of the walk, may produce a large increase in force over the isometric level (see, e.g., Fig. 10 of Ref. 25). SOL forces during standing and locomotion The present results show that the SOL muscle can generate as much as 1.8 kg of force during quasi-isometric standing (e.g., Fig. 6), whereas the same muscle produced, in maximum fused tetanic contraction, 2.8 kg wt at optimum length (100/s stimulus rate). When driven by synaptic inputs, SOL motoneurons in the cat probably fire at relatively low frequencies (19) and the isometric force output from each SOL motor unit would probably be no more than about 75% of maximum (at optimum length; see Ref. 9), or about 2.1 kg wt for 100% recruitment. Given the fact that the isometric optimum length of the cat SOL muscle occurs at ankle angles around 60” for tetani (at 10-30/s stimulation; Ref. 26), and greater angles (corresponding to shorter SOL lengths) are observed during standing, 100% recruitment should produce less than 2 kg wt. Thus, quiet standing in the cat probably requires, on some occasions at least, recruitment of the entire SOL motor-unit population, with each unit firing near its maximum physiological frequency. Even though the average SOL forces recorded during standing are lower (approximately 1.0 kg wt; Fig. S), these may also at times involve nearcomplete recruitment of the SOL motor-unit pool, given the steep isometric forcelength relationship for this muscle at ankle angles around 90” (18, 27). It should be noted, however, that SOL forces are quite small in some examples of standing (Fig. 7), so that postural actions clearly do not always involve complete SOL recruitment. It is also interesting in this regard that there is a certain degree of independence between SOL and MG action in standing (Fig. 7). For reasons mentioned in the preceding section, interpretation of the SOL forces produced during treadmill locomotion at various speeds i:, complicated. Stepping

AND

BURKE

requires motor-unit firing in relatively short bursts, precisely timed with respect to foot fall. During a run at 3.0 m/s (Fig. 3), SOL force reached a peak approximately 40 ms after foot contact, and returned to base line less than 100 ms later. EMG activity in SOL could be detected 50 ms before foot contact and lasted throughout the period of force production. For comparison, the average single twitch time to peak for SOL motor units is approximately 80-90 ms with the muscle isometric (7, 23). Since SOL muscle units contract relatively slowly, the time at which a unit begins firing with respect to foot contact will determine the contribution that that unit can make to the peak of SOL force. Even allowing for the effect of active lengthening, it would appear that those SOL units that begin firing before foot contact should contribute relatively more to the peak of SOL force than those that fire after foot contact, at least during fast runs. This timing effect may be less critical during slow walks where the force lasts over several hundred milliseconds (Fig. 2). The mechanical stiffness of the active SOL motor units during the yield phase, when the muscle is being stretched, is probably responsible for the rapid rise of the SOL force during the yield phase of the run (e.g., see Ref. 18). This conclusion is further supported by the fact that the rate of rise of SOL force is much greater during fast treadmill locomotion than during slow walks (see Figs. 2 and 3), presumably because the rate and extent of muscle lengthening during the yield are also much greater during fast runs as compared to slow walks (16). It seems paradoxical that the average peak SOL force was found to be the same over the whole range of treadmill locomotion studied (0.6-3.0 m/s), in view of the greater rate and extent of stretch during fast as compared to slow locomotion. However, as discussed above, the period of force production is so brief during fast runs that it may not allow all of the recruited SOL units to contribute fully to the peak of SOL force. The data obtained in the present study suggest that this negative effect balances to some extent the force enhancement due to increased active lengthening in the faster runs, resulting in an apparent “saturation” of peak SOL force irrespective of locomotion speed.

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on August 7, 2018. Copyright © 1978 American Physiological Society. All rights reserved.

MUSCLE

FORCE

DURING

LOCOMOTION

1213

The presumed full recruitment of SOL motor units at relatively low levels of output demand fits well with its motor-unit composition. The cat SOL is made up entirely of type S, slow-twitch motor units with a rather narrow range of properties (7, 23). SOL motoneurons are all powerfully excited by primary, or group Ia, muscle spindle afferents (3, 13) and, if one accepts the view (10) that the strength of group Ia synaptic connectivity is a reasonable index of susceptibility to recruitment (as it is in the stretch reflex (4, 12)), then the evident readiness of SOL recruitment and the limited dynamic range over which it is modulated are not surprising. 25

Correlation of dynamic unit properties in MG

range with motor-

PERCENT FIG.

9.

Recruitment

50

of

POOL

model

75

100

RECRUITED

for MG motor-unit

pool

of MG motor-unit types and of Recently, Burke, Rymer, and Walsh (10) based on distributions showed that the strength of group Ia maximum tetanic tension for individual units (10). The heavy solid curve indicates estimated percentage of connectivity to MG motoneurons is closely maximum (fused tetanus) MG force (left ordinate) as a related to motor-unit type. The average function of percentage MG pool recruited (abscissa), amplitude of MG Ia EPSPs was greatest in based on set of assumptions noted in text. Intensity of shading beneath the curve denotes relative fatigue motoneurons of the slowly contracting, resistance of the motor-unit groups in MG (i.e., fatigue-resistant type S units, somewhat less types S > FR > F(int) > FF). The labeled circles and in the faster contracting but fatigue-resistant brackets indicate means t SD MG forces during type FR and F(int) units, and least among standing, locomotion at various speeds, and 120 cm jumps, all referred to right ordinate scaled in kilothe large force, fast-twitch but fatigue-sensigram weight. See text for discussion. tive type FF units. Based on the assumption that the strength of group Ia input, and of other functionally relevant input systems force (in percent of total; left ordinate) similarly organized, is the key factor that developed at the MG tendon if: 1) recruitgoverns MG motor-unit recruitment order ment of the MG motor-unit pool occurs (see Ref. 5 for a discussion of this issue), a according to the above recruitment rule, recruitment model of the MG motor-unit 2) the first unit recruited produces its pool was constructed (Fig. 8 of Ref. 10). The maximum isometric tetanic force before the estimates made for absolute force outputs next unit becomes active, and 3) each from the cat MG during locomotion, which subsequent unit adds its full tetanic force were based on the fragmentary indirect linearly to that of the already active units. evidence then available, have been essen- The nonlinear character of the relation tially corroborated by the present results. between percent cumulative force output The use of group Ia EPSP amplitude as a (left ordinate) and percent of the motor unit “recruitment rule” for the cat MG pool pool recruited (abscissa) is due to the strong negative correlation between average force suggests that MG motor units are recruited in the following general sequence: type output per unit and motor-unit type (8; see S + type FR -+ type F(int) -+ type FF. also Fig. 3 of Ref. 10). We have reexamined the estimated force The above set of assumptions is overin that recruitment is almost production from MG motor units, now using simplified this sequence as the recruitment rule rather certainly not so strictly tied to the unit-type motor units when than Ia EPSP amplitude. For purposes of groups. Furthermore, argument we will assume that all type S recruited probably exhibit some range of units are recruited fully before any type FR firing-frequency modulation as other units unit, and so on through the MG unit pool. become active, as occurs in human forearm The curve in Fig. 9 shows the cumulative muscles (see Refs. 15, 24). Mitigating such

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on August 7, 2018. Copyright © 1978 American Physiological Society. All rights reserved.

1214

WALMSLEY,

HODGSON,

objections is the fact that considerable indirect evidence suggests that the general order of motor-unit recruitment is in fact related closely, if not precisely, to both motor-unit type and to unit force output in the pattern assumed for the present model (for references, see Ref. 6). MG motor units can produce forces approaching fused isometric tetanic force, even at low average firing frequencies, if one or more short intervals (“doublets”) occur as they are activated (9). Doublet firing has been found to occur in a large proportion of motor units active during fictive locomotion in mesencephalic cat preparations (33). The active lengthening of MG during the step cycle presumably enhances force output from MG motor units as it appears to do in SOL. The effects of doublet firing and of active lengthening during locomotion both tend to increase motor-unit forces above those expected for a given rate of motoneuron firing under isometric conditions. In the absence of quantitative data with which to construct a more accurate model, we suggest that the curve shown in Fig. 9, although based on isometric force data, nevertheless represents a reasonable first estimate of the recruitment-force output relation characteristic of the MG motor-unit population during locomotion. The right-hand ordinate of Fig. 9, scaled in absolute force, is based on data from a 2.8-kg cat in which fused isometric tetanus of the MG muscle (100% recruitment) produced 8 kg wt of force. The open circles and brackets show the means and standard deviations of MG forces measured in the same cat during standing, locomotion at four different treadmill speeds, and during 120-cm vertical jumps. Extrapolating these forces onto the curve of force versus recruitment suggests that quadrupedal standing in the cat may involve, primarily if not exclusively, activation of the type S motor-unit complement. The model further suggests that the increasing output from MG during locomotion must involve recruitment of the fatigue-resistant fast-twitch (type FR) units. However, it is surprising that even fast runs (at 3 m/s) may require predominant recruitment of only that half of the MG motor unit pool that is fatigue resistant (summing the outnuts of S. FR. and Flint)

AND

BURKE

units together). On the average, even fast running demands less than 25% of the maximum isometric force capability of the MG muscle. The MG forces developed during vertical jumps to about 120 cm are very close to those produced by the fully tetanized isometric MG muscle. If we make allowance for the competing effects of less-than-fused firing frequency versus doublet firing and active lengthening, the correspondence between the observed jump forces and the estimated forces generated by 100% MG recruitment suggests that the MG motorunit pool is essentially fully recruited during the intense burst that occurs during such jumps (Figs. 5 and 6). The fact that some vertical jumps involve forces greater (by lo- 15%) than maximum fused tetani can be accounted for by the factor of active lengthening during the launch phase and by the fact that the fully active MG muscle must possess great inherent stiffness enabling it, by virtue of its two-joint disposition, to transmit forces from the powerful knee extensors to the calcaneus. The present results, together with previous evidence, suggest that the type FF units, which represent about 45% of the MG motor-unit pool and which produce about 75% of its total force output, may be largely unused even during rather fast running. They are, however, essential for brief but very large force demands such as occur, presumably, in galloping and which certainly occur in vertical jumps. The relative infrequency of such demands during the normal activities of cats fits very well with the extreme susceptibility of the type FF units to mechanical fatigue during even modest levels of repetitive activity (8). The contrasting terms tonic and phasic are sometimes used to describe the supposed functional characteristics of slowtwitch versus fast-twitch muscles, motor units, and even motoneurons. The present results show, however, that the nominally tonic SOL can produce as much or more peak force during locomotion as the MG, and the rates of rise in SOL and MG forces during stance are also comparable despite their very different motor-unit populations. The observations can only be explained by taking account of the effect of active

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on August 7, 2018. Copyright © 1978 American Physiological Society. All rights reserved.

MUSCLE

FORCE

DURING

lengthening in addition to factors of motorunit properties and activity patterns. The SOL also clearly acts phasically during vertical jumping. The adjectives tonic and phasic cannot really describe the range of action of either SOL or MG, and it may be

LOCOMOTION

1215

best to abandon these terms as descriptors of muscles or their motor units.

Present address Monash University,

of B. Walmsley: Dept. Clayton, Victoria,

of Physiology, Australia.

REFERENCES 1. ARIANO, M. A., ARMSTRONG, R. B., AND EDGERTON, V. R. Hindlimb muscle fiber populations of five mammals. J. Histochem. Cytochem. 21: 51-55, 1973. 2. BARNES, G. R. G. AND PINDER, D. N. In vivo tendon tension and bone strain measurement and correlation. J. Biomech. 7: 35-42, 1974. 3. BURKE, R. E. Group Ia synaptic input to fast and slow twitch motor units of cat triceps surae. J. Physiol. London 196: 605-630, 1968. 4. BURKE, R. E. Firing patterns of gastrocnemius motor units in the decerebrate cat. J. Physiol. London 196: 631-654, 1968. 5. BURKE, R. E. On the central nervous system control of fast and slow twitch motor units. In: New Developments in Electromyography and Clinical Neurophysiology , edited by J. E. Desmedt. Basel: Karger, 1973, p. 69-94. 6. BURKE, R. E. AND EDGERTON, V. R. Motor unit properties and selective involvement in movement. In: Exercise and Sport Sciences Reviews, vol. 3, edited by J. H. Wilmore and J. F. Keogh. New York: Academic, 1975, p. 31-81. 7. BURKE, R. E., LEVINE, D. N., SALCMAN, M., AND TSAIRIS, P. Motor units in cat soleus muscle: physiological, histochemical and morphological characteristics. J. Physiol. London 238: 503-514, 1974. 8. BURKE, R. E., LEVINE, D. N., TSAIRIS, P., AND ZAJAC, F. E. Physiological types and histochemical profiles in motor units of the cat gastrocnemius. J. Physiol. London 234: 723748, 1973. 9. BURKE, R. E., RUDOMIN, P., AND ZAJAC, F. E. The effect of activation history on tension production by individual muscle units. Brain Res. 109: 515-529, 1976. 10. BURKE, R. E., RYMER, W. Z., AND WALSH, J. V., JR. Relative strength of synaptic input from shortlatency pathways to motor units of defined type in cat medial gastrocnemius. J. Neurophysiol. 39: 447-458, 1976. 11. CAVAGNA, G. A. AND CITTERIO, G. Effect of stretching on the elastic characteristics and the contractile component of frog striated muscle. J. Physiol. London 239: I-14, 1974. 12. DENNY-BROWN, D. On the nature of postural reflexes. Proc. Roy. Sot. London Ser. B 104: 252-301, 1929. 13. ECCLES, J. C., ECCLES, R. M., AND LUNDBERG, A. The convergence of monosynaptic excitatory afferents on to many different species of alpha motoneurones. J. Physiol. London 137: 22-50, 1957. 14. ENGBERG, I. AND LUNDBERG, A. An electromyo-

15.

16.

17.

18.

19.

20.

21.

22. 23.

24.

25.

26.

27.

graphic analysis of muscular activity in the hind limb of the cat during unrestrained locomotion. Acta Physiol. Stand. 75: 614-630, 1969. FREUND, H. J., B~~DINGEN, H. J., AND DIETZ, V. Activity of single motor units from human forearm muscles during voluntary isometric contractions. J. Neurophysiol. 38: 933-946, 1975. GOSLOW, G. E., REINKING, R. M., AND STUART, D. G. The cat step cycle: hind limb joint angles and muscle lengths during unrestrained locomotion. J. Morphol. 141: l-42, 1973. GOSLOW, G. E., STAUFFER, E. K., NEMETH, W. C., AND STUART, D. G. Digit flexor muscles in the cat: their actions and motor units. J. Morphol. 137: 335-352, 1972. GRILLNER, S. The role of muscle stiffness in meeting the changing postural and locomotor requirements for force development by the ankle extensors. Acta Physiol. &and. 86: 92-108, 1972. GRILLNER, S. AND UDO, M. Motor unit activity and stiffness of the contracting muscle fibers in the tonic stretch reflex. Acta Physiol. Stand. 81: 422-424, 1971. JOYCE, G. C., RACK, P. M. H., AND WESTBURY, D. R. The mechanical properties of cat soleus muscle during controlled lengthening and shortening movements. J. Physiol. London 204: 461-474, 1969. LOEB, G. E., SALCMAN, M., AND SCHMIDT, E. M. Parylene as a chronically stable, reproducible microelectrode insulator. IEEE Trans. Biomed. Eng. 24: 121-128, 1977. MANTER, J. T. The dynamics of quadrupedal walking. J. Exptl. Bipl. 15: 522-540, 1938. MCPHEDRAN, A. M., WUERKER, R. B., AND HENNEMAN, E. Properties of motor units in a homogeneous red muscle (soleus) of the cat. J. Neurophysiol. 28: 71-84, 1965. MILNER-BROWN, H. S., STEIN, R. B., AND YEMM, R. The orderly recruitment of human motor units during voluntary isometric contractions. J. Physiol. London 230: 359-370, 1973. NICHOLS, T. R. AND HOUK, J. C. Improvement in linearity and regulation of stiffness that results from actions of stretch reflex. J. Neurophysiol. 39: 119-142, 1976. PROCHAZKA, A., WESTERMAN, R. A., AND ZICCONE, S. P. Remote monitoring of muscle length and emg in unrestrained cats. ElectroClin. Neurophysiol. 37: 649-653, encephalog. 1974. RACK, P. M. H. AND WESTBURY, D. R. The effects of length and stimulus rate on tension in the isometric cat soleus muscle. J. Physiol. London 204: 443-460, 1969.

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on August 7, 2018. Copyright © 1978 American Physiological Society. All rights reserved.

1216

WALMSLEY,

HODGSON,

28. SALMONS, S. Report on the 8th international conference in medical and biological engineering. Biomed. Eng. 4: 467-474, 1969. 29. SMITH, J. L., EDGERTON, V. R., BETTS, B., AND COLLATOS, T. C. EMG of slow and fast ankle extensors of cat during posture, locomotion, and jumping. J. Neurophysiol. 40: 503-513, 1977. 30. STEPHENS, J. A., REINKING, R. M., AND STUART, D. G. The motor units of cat medial gastrocnemius: electrical and mechanical properties as a function of muscle length. J. Morphol. 146: 495-512, 1975. 31. WETZEL, M. C., GERLACH, R. L., STERN, L. Z., AND HANNAPEL, L. K. Behavior and histochemistry of functionally isolated cat ankle extensors. Exptl. Neurol. 39: 223-233, 1973.

AND

BURKE

32. YAGER, J. G. The Electromyogram as a Predictor of Muscle Mechanical Response in Locomotion (Doctoral Thesis). Memphis: University of Tennessee, 1972. 33. ZAJAC, F. E. AND YOUNG, J. L. Discharge patterns of motor units during cat locomotion and their relation to muscle performance. In: Neural Control’of Locomotion, edited by R. M. Herman, S. Grillner, P. S. G. Stein, and D. G. Stuart. New York: Plenum, 1976, p. 789-793. 34. ZOMLEFER, M. R., ZAJAC, F. E., AND LEVINE, W. S. Kinematics and muscular activity of cats during maximum height jumps. Brain Res. 126: 563-566, 1977.

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on August 7, 2018. Copyright © 1978 American Physiological Society. All rights reserved.

Forces produced by medial gastrocnemius and soleus muscles during locomotion in freely moving cats.

JOURNALOF NEUROPHYsOLOGY Vol. 41, No. 5, September 1978. Printed in U.S.A. Forces Produced by Medial Gastrocnemius Soleus Muscles During Locomoti...
2MB Sizes 0 Downloads 0 Views