Are skeletal muscles independent actuators? Force transmission from soleus muscle in the cat

Huub Maas, Thomas G. Sandercock


It is unclear if skeletal muscles act mechanically as independent actuators. The purpose of the present study was to investigate force transmission from soleus (SO) muscle for physiological lengths as well as relative positions in the intact cat hindlimb. We hypothesized that force transmission from SO fibers will be affected by length changes of its two-joint synergists. Ankle plantar flexor moment on excitation of the SO was measured for various knee angles (70–140°). This involved substantial length changes of gastrocnemius and plantaris muscles. Ankle angle was kept constant (80°-90°). However, SO ankle moment was not significantly affected by changes in knee angle; neither were half-relaxation time and the maximal rate of relaxation (P > 0.05). Following tenotomy, SO ankle moment decreased substantially (55 ± 16%) but did not reach zero, indicating force transmission via connective tissues to the Achilles tendon (i.e., epimuscular myofascial force transmission). During contraction SO muscle shortened to a much greater extent than in the intact case (16.0 ± 0.6 vs. 1.0 ± 0.1 mm), which resulted in a major position shift relative to its synergists. If the SO was moved back to its position corresponding to the intact condition, SO ankle moment approached zero, and most muscle force was exerted at the distal SO tendon. Our results also suggested that in vivo the lumped intact tissues linking SO to its synergists are slack or are operating on the toe region of the stress-strain curve. Thus, within the experimental conditions of the present study, the intact cat soleus muscle appears to act mechanically as an independent actuator.

  • connective tissue
  • tenotomy
  • myofascial
  • gastrocnemius muscle
  • plantaris muscle

are skeletal muscles independent actuators? Neural excitatory and inhibitory linkages among muscles acting at the same and distant joints in the cat (e.g., 10, 42) indicate the contrary. Such spinal circuitry is among others responsible for stretch-evoked reflexes between soleus (SO) and lateral gastrocnemius (LG) muscles (43). Mechanically, it is common to assume morphologically defined muscles to be independent force generators. This assumption is applied in most biomechanical models of the musculoskeletal system (e.g., 7), as well as in studying the in vivo properties of human muscle-tendon units (MTUs; e.g., 18). It has recently been shown in rat that the connective tissues linking synergistic muscles to each other and to nonmuscular surrounding structures are capable of transmitting muscle fiber force (22, 32). Mechanical interactions at the muscle belly interface between SO and whole gastrocnemius muscle (GAS) have also been noted in the cat as an experimental nuisance for the measurement of muscle properties (8). The presence of this so-called myofascial force transmission suggests that also the motors of human and animal movement cannot be considered as independent functional units.

However, the studies by Huijing and colleagues (for a review see 21) involved in several cases unphysiological muscle lengths and relative positions. For example, the MTU length of a single muscle was changed while the length of its synergists was kept constant. The imposed changes in MTU length were also beyond the operating length range during normal movements. In addition, summation of joint moments exerted by medial gastrocnemius muscle (MG) and LG-SO complex was reported to be nearly linear (<3%) in the cat (45). Their results suggest very minor mechanical interactions between adjacent muscles, but this was only tested for one MTU length. Therefore, the significance of myofascial force transmission for normal muscle function in vivo remains unclear.

The purpose of the present study was to investigate force transmission from SO muscle for physiological lengths and relative positions in the cat. SO is a one-joint muscle (5) exerting a plantar flexor moment about the ankle joint (31). The muscle belly of SO shares an interface with the two-joint LG and plantaris muscles (11). On the basis of the mechanical interactions between synergists reported in the rat (32) and the finding that the extent of myofascial force transmission is dependent on the relative position of muscle bellies (33), we hypothesized that force transmission from SO muscle fibers will be affected by length changes of its synergists. It should be noted that such intermuscular interaction can be mediated by both intermuscular as well as extramuscular connective tissues (35). The term epimuscular myofascial force transmission has been introduced to indicate transmission via all connective tissues linking a muscle to its immediate surroundings (20).

Ankle moment exerted by SO was measured at various knee angles and, thus, different MTU lengths of LG and plantaris. This was done in the nearly intact cat hindlimb, ensuring physiological muscle conditions. As the results did not confirm the hypothesis, we also investigated if there are any mechanical connections between SO and its surroundings that can bear force. For that purpose, force transmission from SO following tenotomy, which eliminated force transmission to its insertion on the calcaneus, was assessed.



The data were obtained from seven cats (male and female, 2.8–4.6 kg). All surgical and experimental procedures complied with the Guide for the Care and Use of Laboratory Animals (National Institutes of Health Publication No. 86-23, Revised 1985) and were approved by the Northwestern University Animal Care and Use Committee.

Surgical preparations were done under gaseous anesthesia (1–4% isoflurane in a 3:1 mix of N2O and O2), which was then switched to pentobarbital sodium (10%, initial dose ∼10 ml iv via jugular vein) for data collection. Supplemental doses (1 ml a time) were administered to maintain a deep anesthetic state, as judged by complete absence of withdrawal reflexes and steady blood pressure. All cats were monitored for blood pressure, heart rate, respiration, and body temperature for the full length of the experiment. At the end of the experiments, the cats were euthanized without regaining consciousness with a lethal dose (100 mg/kg iv) of pentobarbital sodium and double-sided pneumothorax.

Experimental preparation.

The cat was mounted in a rigid frame securing the head and the dorsal process of lumbar vertebrae 3. Pelvic rotation was prevented by pins in the pelvis and a clamp on the ischium. The entire left foot was rigidly clamped to a 6 degree-of-freedom load cell (JR3) coupled to a 6 degree-of-freedom robotic manipulator (RX60, Staubli) using metal brackets placed over the calcaneous and metatarsals. This fixation did not allow any displacement of the ankle joint relative to the JR3 (illustrated in Fig. 1). The angle of ankle, knee, and hip joints in this multisegment rigid body consisting of the foot, shank, and thigh was thus determined by the position of the robot relative to the hip.

Fig. 1.

Schematic view of the experimental set-up. The entire left foot was rigidly attached to a 6 degree-of-freedom load cell (JR3) coupled to a 6 degree-of-freedom robotic manipulator (RX60, Staubli). The 1-joint soleus and the 2-joint synergists are indicated, as well as the hip fixation (see methods). In the 1st series of experiments, the soleus (SO) tendon was left attached to the calcaneus (see A). Following tenotomy, the distal SO tendon was attached to a force transducer (see C). With a constant angle of ankle and knee joints (80–90°), load cell data were collected at various muscle-tendon unit (MTU) lengths (l) of SO muscle. The latter was obtained by movements of the force transducer (Δl MTU).

The endpoint of the robotic arm moves with a precision of 20 μm with a very high stiffness (300,000 N/m). Mediolateral movements of the knee were restricted by blocking the medial and lateral aspect of the joint without any restriction of rotations in the sagittal plane (i.e., flexion-extension). The position of the rotation axes of ankle and knee joints relative to the endpoint of the robotic arm were measured with a fine ruler. This information was used by the position control software of the robotic arm (Adept Technology) to impose isolated sagittal plane rotations in each joint. The ankle joint was moved by rotations of the foot about the measured axis of rotation (i.e., through the medial and lateral malleoli perpendicular to the sagittal plane). The knee joint was moved by rotations of the foot-lower limb complex about the measured axis of rotation [i.e., through the distal end of the femur medially and midway between the origin of extensor digitorum longus (EDL) and the posterior fibular collateral ligament laterally; 13]. The position of the joint axes of rotation was assumed to be constant and perpendicular to the sagittal plane.

A longitudinal incision was made in the popliteal region of the left hindlimb to expose the tibial nerve. The LG-SO branch was identified proximal to where it enters the muscle compartment. Using a dissecting microscope the epineurium was then cut, and the individual nerve bundles supplying LG (typically 3 or 4) were dissected free from the nerve bundle supplying SO (see also 36). Each nerve bundle was identified using electrical stimulation via a bipolar hook electrode. In all but one experiment, which was excluded for further data analysis, a single nerve bundle that clearly excited only SO muscle fibers was successfully isolated. No SO muscle fibers were excited when stimulating the other nerve bundles, indicating that the nerve bundle included all SO motoneuron axons. The LG nerve bundles as well as the MG nerve branch, tibial, common peroneal, and sural nerves were severed. Muscle excitation via spinal reflexes was further eliminated by cutting the sciatic nerve as proximally as possible. The blood supply to the whole hindlimb was preserved, and care was taken not to disrupt the connective tissues of the ankle plantar flexor compartment. A cuff electrode was placed on the tibial or sciatic nerve with only the SO nerve bundle intact. Radiant heat was used to maintain hindlimb and core temperatures within physiological limits. Dehydration of muscles and nerves was prevented by regular irrigation with saline.

Experimental protocols.

SO muscle was excited by stimulation of the tibial or sciatic nerve using suprathreshold pulses delivered through a cuff electrode connected to a stimulator (Grass S88) and stimulus isolation unit (PSIU 6, constant-current stimulation 1–3 mA, pulse width 0.1 ms, pulse train 750–1,000 ms, frequency 100 Hz). One-minute rest periods were allowed between trials. Several control measurements were performed in different stages of the experiment to monitor any decrease in muscle force, but SO did not show evidence of fatigue (the decrease in SO ankle moment < 1.1%, 0.8 ± 0.3%).

For each experiment, the hip joint was kept at a constant angle (90–120°). Output from the JR3 [the force (Fx, Fy, Fz) and moment (Mx, My, Mz) in the x-, y-, and z-directions, respectively] was recorded at several combinations of ankle and knee angles before and during the tetanic contraction of SO muscle. Note that the joint angles were only changed in the sagittal plane. The data were sampled at 1 kHz and filtered using a finite impulse response filter (low pass cut-off 50 Hz). The moments were transformed from the geometrical center of the JR3 to the ankle joint to obtain the joint moments in the anatomic axes centered between the ankle malleoli (i.e., eversion-inversion, plantar flexion-dorsiflexion, and abduction-adduction; the first one mentioned was defined positive). As SO only exerts a substantial moment in the sagittal plane (31), the present study will focus on the plantar flexor moment.

In four cats, effects of the force exerted by SO muscle fibers were measured in the following two experimental conditions: 1) for different positions of the ankle joint (between 50° and 100°) while the knee was kept at a constant position (80–90°); and 2) for different positions of the knee joint (between 70° and 140°) while the ankle was kept at a constant position (80–90°). The target joint was moved to the desired angle by the robotic arm, and the muscle was excited. In between trials, the ankle joint was moved to such a position that corresponded with a low MTU length of the ankle extensors. Different joint angles were applied in randomized order. MTU length changes of LG and plantaris (the muscles that interface with the SO muscle belly, see 11) due to changes in knee joint angle were calculated using the geometric model presented by Goslow et al. (13), to assess relative muscle movements.

Ankle moment before (passive) and during the last 100 ms of the plateau phase of SO muscle contraction (total) were calculated (see Fig. 2). Passive moment was subtracted from the total moment to assess active ankle moment. In addition, the time of moment decrease from the highest value following the last nerve stimulus to half that value (half-relaxation time, HRT) and the maximal rate of relaxation were determined from the active joint moment-time curves (for representative waveforms, see Figs. 2 and 6). The relaxation rate was normalized to the moment level from which the muscle starts to relax (6).

Fig. 2.

Typical example of ankle moment waveform during SO muscle contraction. For this contraction, the ankle joint was held at 80° and the knee joint at 90°. Vertical dotted lines indicate beginning (t = 750 ms) and end (t = 1,500 ms) of SO nerve stimulation. The short vertical lines indicate the time windows (duration = 100 ms) that were used to calculate ankle moment before and during SO muscle contraction. The short horizontal line indicates where the ankle moment was decreased by half the highest value following the last nerve stimulus, which was used to calculate the half-relaxation time.

Tenotomy of SO muscle.

In four of the seven experiments, the ankle joint moment during SO muscle contraction was measured after tenotomy of the SO insertion onto the calcaneus. The hindlimb was kept at a constant position (i.e., 80–90° for ankle and knee joint). The posterior aspect of the crural fascia was sectioned from the popliteal region up to the calcaneus. With minimal disruption of the connective tissues at the muscle belly level, the distal tendon of SO was dissected free from the other tendons in the Achilles complex. For measurements of changes in SO relative position, one marker was placed on the LG tendon and one marker on the distal end of the muscle belly of SO. Marker movements in the sagittal plane were recorded using a digital camera (Nikon Coolpix 990, 15 frames/s, 320 × 240 pixels, resolution 1 pixel ∼0.25 mm). The SO distal tendon was then cut, which eliminated its distal myotendinous pathways for force transmission. Any ankle joint moment measured following SO excitation should thus be attributed to force transmission from SO muscle fibers via connective tissues onto adjacent ankle extensors (i.e., epimuscular myofascial force transmission). To be able to track the SO marker during muscle contraction, the lateral surface of the crural fascia was excised also. The SO nerve was stimulated by a range of frequencies (i.e., 10, 15, 20, 30, 50, and 100 Hz) to vary the force exerted by its muscle fibers. Simultaneously, movement of the SO marker relative to its surrounding structures (i.e., LG marker) was assessed as a measure of net tissue deformation.

Preliminary results indicated that the ankle moment exerted by SO muscle following tenotomy was absent if the muscle was held at its original position, as obtained by clamping the distal tendon and aligning the marker on SO with the marker on LG (see above). Therefore, this effect was studied in more detail in two cats. After measurements in the tenotomy condition, the distal tendon of SO was connected to a force transducer (model 31; Honeywell Sensotec, Columbus, OH) mounted on a linear puller (stiffness 250 N/mm, Copley ThrustTube TB3806; Copley Controls, Canton, MA). The puller was positioned in such a way that the SO tendon approached the original line of pull (i.e., to the calcaneus), but without contact to other tissues causing friction (Fig. 1C). All other conditions were equal to those described before.

Ankle moment and isometric force exerted at the distal tendon were simultaneously measured at various MTU lengths of SO muscle. The muscle was lengthened from the position it obtained following tenotomy (i.e., the tendon was slack) up to and beyond its original length. After each contraction the SO muscle was allowed to recover at low length for at least 1 min. Tendon force was assessed just before (passive) and during the plateau phase (total) of the tetanic muscle contraction, using the same 100-ms time windows as for ankle moment (see above). Active force was calculated by subtracting passive force from total force at equal MTU length.

At the end of each tenotomy experiment, SO was fully dissected free from surrounding tissues, leaving only the nerve and blood supply intact. Ankle moment on muscle stimulation was negligible in all cases, confirming that there was no experimental bias due to forces exerted by LG muscle fibers. This also excluded effects of lateral expansion on Achilles tendon force and, hence, ankle moment when SO muscle shortens following tenotomy.


An ANOVA for repeated measures was used to test for effects of knee and ankle angle on ankle moment generated by SO, HRT, and maximal rate of relaxation. Paired t-tests with Bonferroni correction were applied to test for effects of tendon dissection and tenotomy on ankle moment. P values <0.05 were considered statistically significant.


Ankle angle-moment characteristics of SO muscle.

Excitation of SO muscle resulted in substantial moments exerted in the sagittal plane (i.e., plantar flexion). In agreement with previous studies (30, 31), moments about the other two axes of rotation (i.e., eversion/inversion and abduction/adduction) were also found (data not shown). As in the present study ankle or knee joint angle was changed exclusively in the sagittal plane, we will focus on the plantar flexor moment exerted by SO muscle.

Passive ankle moment increased exponentially with ankle dorsiflexion, which corresponds to lengthening of the ankle plantar flexors (Fig. 3A). Note that the passive ankle moment originates not only from passive forces of SO, but also from passive forces of its synergists. This explains why the peak passive moment is higher than the maximum active moment of SO (Fig. 4).

Fig. 3.

Passive joint angle-ankle moment characteristics. A: passive ankle moment as a function of ankle angle (i.e., plantar flexion-dorsiflexion). Knee angle was kept constant at 80–90°. B: passive ankle moment as a function of knee angle in the sagittal plane (i.e., flexion-extension). Ankle angle was kept constant at 80–90°. Data of individual cats are shown (n = 4).

Fig. 4.

Ankle angle-ankle moment characteristics of SO muscle. Active ankle moment exerted by SO muscle plotted as a function of ankle angle in the sagittal plane (i.e., plantar flexion-dorsiflexion) for each cat (n = 4). A: absolute moments. B: moments as a percentage of the peak active moment. Knee angle was kept constant at 80–90°.

Measurements at various ankle angles yielded the ankle angle-ankle moment characteristics of SO muscle (Fig. 4). Changes in ankle moment below and beyond the optimum angle were small in the range of ankle angles tested (Fig. 4B). Maximal plantar flexor moment varied between 0.13 and 0.52 Nm. This variation may be explained by differences in physiological cross-sectional area of the SO muscles since our animals varied in body mass (46). Maximal ankle moment could also be affected by disruption of the SO nerve branch during the experimental procedures. Although dissection of the nerve was done meticulously, some damage cannot be excluded. It should be noted that while the muscle was in some cases only partially activated, the level of muscle activation was constant throughout the experiment (see methods).

Effects of knee angle on ankle moment generated by SO muscle.

SO plantar flexor moment was not significantly (P = 0.11) affected by changes in knee angle (Fig. 5), despite the fact that this involved a substantial change in MTU length of LG and plantaris muscles (i.e., on average 1.15 mm/10° knee angle, SD = 0.04). A monotonic increase in passive ankle moment from the most flexed knee position to the most extended one (Fig. 3B) confirmed that these muscles were not slack in any of the knee angles tested. Note also that LG and plantaris were not excited during SO muscle contraction, and ankle angle was kept constant (see methods). For three of the four cats, maximum ankle moment was found at the lowest two knee angles. In those cases, linear regression yielded a small negative slope, suggesting a trend of a decreasing active plantar flexor moment with knee extension (i.e., lengthening of LG and plantaris muscles). This yielded a maximal decrease in ankle moment to values of 87.4%, 94.6%, and 96.5 % for cats 4, 6, and 7, respectively (Fig. 5B).

Fig. 5.

Effects of knee angle on SO ankle moment. Active ankle moment exerted by SO muscle plotted as a function of knee angle in the sagittal plane (flexion-extension) for each cat (n = 4). A: absolute moments. B: moments as a percentage of the peak active moment. Ankle angle was kept constant at 80–90°.

On the basis of the relatively wide plateau of the ankle angle-moment characteristics (Fig. 4), it could be argued that only small effects of knee angle on ankle moment as a result of SO muscle contraction can be expected. It was hypothesized that the contraction dynamics are more sensitive to potential mechanical influences of adjacent muscles than the steady-state output of a tetanic muscle contraction. It was found that the relaxation phase of SO muscle contraction varied strongly with ankle angle, while the phase of moment development was much less ankle angle dependent (Fig. 6A). Similar results have been reported for SO forces in rats (55). Therefore, the effects of knee angle on HRT and the maximal rate of relaxation were assessed and compared to their variation with ankle angle. Both parameters were significantly (P < 0.01) affected by ankle angle (Fig. 7, A and B), but no such dependency was found for changes in knee angle (Figs. 6B and 7, C and D).

Fig. 6.

Normalized ankle moment waveforms to indicate relaxation rate. Representative data from one cat (cat 7) of normalized active ankle moment plotted as a function of time (ms). Note that the passive moment was subtracted from all waveforms so that the contribution from SO contraction could be compared between conditions. A: for a range of ankle angles measured with constant knee angle (80°) and B: for a range of knee angles measured with constant ankle angle (80°). This figure clearly indicates that the relaxation phase is affected by changes in ankle angle but not by changes in knee angle (see also Fig. 7).

Fig. 7.

Effects of ankle and knee angle on rate of relaxation of active moment. Half-relaxation time (HRT; ms) and maximum rate of relaxation divided by the active moment level from which muscle started to relax (1/s). Effects of ankle angle (A and B) and knee angle (C and D) are shown for each cat (n = 4). Either the knee or the ankle joint was kept constant at 80–90° (see methods).

These results indicate that changing the MTU length of the passive two-joint LG and plantaris muscles and, consequently, their position relative to the one-joint SO, does not affect force transmission from SO muscle fibers to its insertion.

Force transmission from SO muscle after tenotomy.

Dissecting the SO tendon free from the Achilles tendon complex did not significantly alter ankle moment generated by SO muscle fibers (Fig. 8A). Following tenotomy, with the SO active this moment decreased substantially (55 ± 16%) but did not reach zero (Fig. 8), despite the fact that myotendinous force transmission via the SO distal tendon to the calcaneus was removed. During contraction SO muscle shortened 16.0 ± 0.6 mm more than in the intact-tendon condition, which was only 1.0 ± 0.1 mm. Thus the smaller ankle moment is at least partly caused by a decrease in muscle fiber length. These results indicate force transmission from SO muscle fibers to the Achilles tendon, likely mediated by connective tissues linking the SO muscle belly to LG and plantaris (i.e., epimuscular myofascial force transmission).

Fig. 8.

Effects of tenotomy on active ankle moment of SO. Active ankle moment of SO before and after tenotomy of the distal tendon (n = 4). A: mean ± SD before any disruption of connective tissues (intact), following dissection of the distal tendon (tendon dissected), and after tenotomy. Asterisk denotes a significant difference. B: Pre- and posttenotomy values for each cat. Ankle moment decreased by 55%, 50%, 78%, and 40% for cat 6, 7, 8, and 9, respectively. The ankle and knee joints were kept at 80–90°.

To approximate the mechanical characteristics of the lumped intact tissues linking SO to its synergists, changes in position of the marker on the distal end of the muscle belly were measured at various ankle moment levels. The moment was changed by altering the stimulation frequency of the SO nerve. Plantar flexor moment was considered as a measure of net force borne by these tissues. Changes in marker position relative to its surrounding structures were considered as a measure of net tissue deformation. The higher the ankle moment was, the more relative movement of the SO marker in proximal direction was observed (Fig. 9). This relationship was fairly linear in the range of moments measured. The mean slope of the linear regression lines was 0.045 Nm/mm (SD = 0.020). Using an Achilles tendon moment arm of 16 mm (3), the apparent stiffness of the lumped myofascial linkages of SO muscle was 2.8 N/mm.

Fig. 9.

Ankle moment-relative movement of SO. Displacement of the marker on the distal end of the SO muscle belly relative to the original (∼intact) rest position (0) during contractions of various force levels (see methods) following tenotomy (n = 4). The slope (S) of the linear regression lines, which is an indirect measure of the stiffness of the lumped connective tissues linking SO to its synergists, is indicated for each cat. In the intact condition, contraction of SO muscle resulted in a relative marker movement of only 1.0 ± 0.1 mm. The ankle and knee joints were kept at 80–90°.

These results do not answer the question of whether SO muscle fibers transmit force to their synergists when the muscle belly is at its original position. Note that even for the lowest ankle moment levels the relative marker displacement was substantially higher than observed on excitation of SO before tenotomy (Fig. 9). Therefore, an additional set of experiments was performed. The distal SO tendon was connected to a force transducer, and the MTU was lengthened from the length it obtained after tenotomy up to and beyond its original length, corresponding to the intact condition. It was found that moving the SO tendon in distal direction (i.e., toward its original position) decreased the plantar flexor moment at the ankle, but increased the force exerted at the distal tendon (Fig. 10A). Note that at high MTU lengths of SO, beyond the original position, a small dorsiflexor moment (∼0.01–0.02) was found. The most likely explanation for this is that the ankle joint is not a perfect pin joint. The force exerted by SO on the tibia can cause small posterior movements of the tibia relative to the talus, thereby rocking the foot into dorsiflexion.

Fig. 10.

Simultaneously measured ankle moment and force exerted at the distal tendon of SO. With the foot still attached to the robotic arm, the distal tendon of SO was connected to a force transducer (see Fig. 1C). Ankle moment and distal tendon force were simultaneously measured at various MTU lengths of SO muscle. The ankle and knee joints were kept at 90° (for more details, see methods). Representative data for 1 cat (cat 8) are shown. A: active tendon force (□) and ankle moment (▴) plotted as a function of the position of the marker on the distal end of the SO muscle belly. The latter is expressed as the deviation (Δ) from its original position (0). B: relationship between active tendon force and active ankle moment. C: passive tendon force and change in passive ankle moment plotted as a function of the SO marker position. D: relationship between passive tendon force and the change in passive ankle moment. In C and D, passive ankle moment is expressed as the deviation (Δ) from the value measured at the most proximal marker position. Linear regression line and r2 are indicated in B. Note that the left (ankle moment) and right (SO force) y-axes in A and C have different scaling factors.

If plotted in one figure, a linear relationship was found between ankle joint moment and tendon force (Fig. 10B). This indicates a partitioning of muscle fiber force between two pathways: 1) via the distal tendon of SO and 2) via the connective tissues linking SO muscle to the Achilles tendon. From ∼6 mm below up to 8 mm beyond the original position, ankle moment was close to zero (Fig. 10A), which suggests that all SO force was transmitted to its distal tendon. Lengthening SO muscle also increased the passive force measured before muscle contraction (Fig. 10C). However, substantial forces were found only if the muscle was lengthened beyond its original position. Such an increase in passive SO tendon force was accompanied by a decrease in passive ankle moment (Fig. 10C). In contrast to the active condition, the change in passive ankle moment was not linearly correlated with passive tendon force (Fig. 10D). Note that in addition to the exemplar data shown in Fig. 10, similar results were obtained in two other animals.

These results indicate that the connective tissues between SO and its synergists have a high force-bearing capacity but that the steep portion of the lumped stress-strain curve is not within the in vivo range of SO muscle positions.


The major finding of the present study is that despite strong connective tissue linkages, cat SO muscle appears to be mechanically independent if acting at physiological lengths and relative positions. Specifically, it was found that 1) force transmission from SO is not affected by physiological length changes of adjacent synergists (i.e., obtained by changing knee angle), 2) SO muscle fibers produced a substantial ankle moment after force transmission via its distal tendon was eliminated by tenotomy, and 3) most force is transmitted to the distal tendon if the SO muscle belly is moved back to its original position corresponding to the in vivo condition.

In addition, a first attempt was made to assess the material properties of the lumped intact tissues linking SO to its synergists (Fig. 9). The stress-strain curve was found to be linear over the range of ankle moments tested, and a stiffness of 2.8 N/mm was calculated. For comparison, the common elastic stiffness (i.e., any in-series compliance shared between muscle fibers) of SO ranged between ∼10 N/mm and ∼25 N/mm as estimated by the common elasticity model (51). Similar results were found for the stiffness of the series elastic component of SO using the alpha method (16.9 N/mm, 41). The spindle null technique yielded an in-series stiffness that increased ∼2 N/mm for each newton increase of SO force up to 25 N/mm at 11 N (48). This indicates that on average the myofascial stiffness is lower than the stiffness of myotendinous pathways.

Thus changes in knee angle did not affect SO moment (Fig. 5). Ankle angle movements caused substantially higher, but still rather small (17–30% maximal change) changes in SO moment (Fig. 4). Over the maximum range of ankle angles tested (50–100°) the MTU of SO changes by only 12.9 mm, as calculated using the geometric model of Goslow et al. (13). Due to the in-series compliance, this will yield even smaller changes in fiber length. These results are in agreement with earlier studies that also reported only slight changes in SO force for ankle angles between 50° and 100° (16, 47). This can be explained by the relatively long fiber length of SO muscle (49). Note also that the moment arm is fairly constant for that range of joint angles (56).

Force transmission between muscles.

Mechanical interactions between cat SO muscle and its surrounding tissues have not been studied before. Our results to some extent agree with previous studies of muscular force transmission using a rat model, but in other aspects they lead to different conclusions. In the present study, force exerted by SO muscle could only be measured at the tendon of insertion. Therefore, a potential difference with force exerted at the origin, such as reported for rat extensor digitorum longus (EDL) muscle (e.g., 22), could not be confirmed here. Recently, it was shown that the difference in force between the proximal and distal tendon of EDL can be substantial when its length is changed simultaneously with the (40) synergists.

In two other studies on rat EDL, it was found that one of the four distal tendons could be cut or considerably shortened with minimal effects on force measured at the proximal tendon (24, 34). These results showed that force can be transmitted from the tenotomized muscle fibers to the tendon via the endomysial-perimysial network. In similar fashion, the present results demonstrate that force can be transmitted from the tenotomized SO muscle fibers to the Achilles tendon, most likely via connective tissues at the interface with LG and plantaris. Recent observations indicate that myofascial force transmission is not limited to synergistic muscles, but force can also be transmitted between antagonists (25). Force transmission from SO to its antagonists cannot be confirmed with the present results but could potentially have contributed to the decrease in ankle moment following tenotomy.

Several previous studies reported mechanical interactions between synergistic muscles (23, 32, 33). MTU length or position changes of an agonist altered forces exerted at the tendons of its synergists, despite the fact that the synergists were kept at a constant MTU length. In our study, such interactions were only found following tenotomy (Fig. 8) and at muscle positions proximal or distal to the original position of SO (Fig. 10), but not in the intact condition (Fig. 57). As adjacent muscles sometimes have different moment arms (e.g., 3), changes in relative position between synergistic muscles with single joint movements can also occur in vivo. Nevertheless, the length changes of a single muscle imposed in the experiments by Huijing and colleagues (e.g., 23, 32) were clearly beyond the physiological range. It should be kept in mind that the main purpose of those studies was to investigate the presence and capacity of mechanical connections between muscles. More substantial changes in relative muscle position are expected in vivo if a synergistic group consists of both single-joint and two-joint muscles, such as the ankle plantar flexors in the cat. However, length changes of GAS and plantaris muscles, as obtained by movements of the knee joint, did not significantly affect the SO ankle moment (Fig. 5). This is in agreement with the nearly linear summation of joint moments exerted by MG and LG-SO as reported previously in the cat (45). Also for adjacent muscles in the human arm, it was recently found that myofascial force transmission was minimal (57); an upper limit of 6% for force transmission between the flexor pollicis longus and the index finger compartment of the flexor digitorum profundus was calculated.

It should be noted that in the present study only the muscle of interest was activated, which is a major difference with the simultaneous activation of all synergists and some antagonists used in the above-described experiments in rat. In awake, freely moving cats, selective activity of either SO during standing or LG during paw shakes, as well as coactivation of both muscles during locomotion, has been reported (e.g., 15, 53, 54). Coactivation of synergistic muscles may preload and thus remove some slack in the connective tissues and, hence, facilitate force transmission between muscles. With the experimental procedures used here, it was not possible to keep the nerve branches of LG intact, and thus this hypothesis could not be tested.

Alternatively, the connective tissue linkages between SO and adjacent muscles within the intact cat could be slack or operating on the toe region of their lumped stress-strain curve (for a schematic representation, see Fig. 11, A and B). Such an explanation for the independent behavior of SO muscle is in agreement with our finding that substantial displacement of the distal end of the SO muscle belly and, consequently, changes in length and orientation of myofascial connections (see Fig. 11C) were needed before SO force was transmitted via myofascial pathways to the Achilles tendon, generating an ankle moment (Figs. 9 and 10). Effects of myofascial force transmission may become more prevalent following surgical interventions that involve changes in muscle relative position, such as agonist-to-antagonist tendon transfers (1, 52).

Fig. 11.

Schematic drawings to illustrate length changes of intermuscular linkages. SO muscle is connected to gastrocnemius (GAS) and plantaris (PL) muscles via connective tissue linkages. Changing knee angle from a flexed (A) to a more extended position (B) resulted in reorientation as well as unfolding of those linkages. Such unfolding or straightening of macroscopic crimp in collagen fibrils is correlated to the toe region of the stress-strain curve (9). Note that none of the linkages were fully unfolded with knee movements, indicating minimal force transmission between SO and synergistic muscles. Following tenotomy, SO muscle shortened substantially, straightening as well as extending the more distal linkages (C), now reaching the steep portion of their stress-strain curve.

Measurements of muscle properties in humans using noninvasive techniques.

Besides some exceptions (e.g., 14, 27), forces exerted at the tendons of muscles in humans are measured indirectly from joint moments and tendon moment arms (e.g., 38, 39). These methods are often combined with ultrasonography to measure tendon and aponeurosis strain, as well as changes in muscle fascicle length and pennation angle. In such experiments, it is implicitly or explicitly assumed that muscles are mechanically independent of one another. We encountered only a limited number of studies from which the validity of this assumption may be derived. As forces of individual muscles cannot be measured directly in humans and imaging techniques cannot yield information about force transmission, only indirect evidence exists. Several groups have measured ankle joint moment during voluntary contractions of the ankle plantar flexors for a range ankle and knee angles (17, 26, 50). Such measurements have been used to assess the passive and active length-force characteristics of human two-joint GAS muscle (17, 18, 26). Theoretically, the contribution of the one-joint SO to the ankle moment exerted at various knee angles is the same, but this was not always the case (26). Such inconsistency may be attributed to effects of myofascial force transmission (for a more detailed discussion, see 37). It should be noted that also in humans, the SO muscle belly shares an interface with the gastrocnemius (19). In contrast to cats, the plantaris muscle in humans is very small.

In a recent study, a twitch contraction of MG muscle at a fixed angle of the ankle and knee joints elicited a decrease of fascicle length not only in MG but also in SO (44), suggesting a mechanical connection between these muscles. This observation was used to explain the rather similar contraction dynamics of MG and SO muscle fascicles and tendinous tissues during isometric (44), as well as concentric and eccentric plantar flexor contractions at the ankle (4). In contrast, but in agreement with the results of the present study, the same group has reported that SO fascicle length was not affected by changes in knee angle, as measured in both passive and maximally active conditions of the ankle plantar flexors (29). In addition, differential displacement between SO and GAS aponeuroses was recently observed during voluntary isometric contractions (2). In that same study, human cadaver investigation revealed that the aponeuroses of SO and GAS are separate structures allowing for relative muscle movements. However, connective tissues at other locations of the muscle bellies were not mentioned. The above indicates that it is not possible to draw conclusions about myofascial force transmission based on the presence or absence of relative muscle movements. Therefore, more controlled experiments mimicking in vivo muscle lengths and relative positions, as in the present study, are needed to elucidate if skeletal muscles are independent actuators.

Concluding remarks.

The data do not support our hypothesis that force transmission from SO muscle fibers is affected by length changes of its synergists. Within the experimental conditions of the present study, the intact cat soleus muscle acted mechanically as an independent actuator. This suggests that the central nervous system can control the force exerted at the tendons of individual muscles and take full advantage of their unique three-dimensional action (31), assuming that muscles can be independently recruited. Because for each synergistic group the muscle-connective tissue architecture and composition is different, generalizing the results for SO to the whole musculoskeletal system should be done with caution.

Strong mechanical connections between SO and synergistic muscles were found. However, those connective tissue linkages are slack or on the toe region of the stress-strain curve for physiological muscle conditions. We hypothesize here that those connective tissues may function as a safety barrier for traumatic events in muscle or tendon. Adhering to adjacent structures following damage of the muscle components providing the myotendinous route of force transmission (e.g., muscle fibers, aponeurosis, and tendon) may prevent further trauma and facilitate the recovery process. Such adhesions have been reported in chronically tenotomized muscle (12). At the macromolecular level, the results of a very elegant study indicated that the integrin-vinculin-mediated connections between the subsarcolemmal cytoskeleton and extracellular matrix are temporarily reinforced in ruptured muscle fibers (28). This likely reduced to load on the injured site, allowing repair with a reduced chance of rerupture. Further studies that specifically address this potential function are needed.


This work was supported by National Institute of Arthritis and Musculoskeletal and Skin Diseases Grant AR-041531 as well as by NIDRR Advanced Rehabilitation Research Training Grant H133PO40007.


We thank Dr. C. J. Heckman for comments on the manuscript; Dr. Matthew Tresch for feedback on the experimental methods; Dr. Eric Perreault for help on the revision; and Michael Johnson and Lei Cui for technical support.

Present address for H. Maas: Research Institute MOVE, Faculty of Human Movement Sciences, VU Univ. Amsterdam, Van der Boechorststraat 9, 1081 BT Amsterdam, The Netherlands.


  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract