Coordination of medial gastrocnemius and soleus forces during cat locomotion
Faculty of Kinesiology, Human Performance Laboratory, University of Calgary, 2500 University Drive NW, Calgary, Alberta, Canada, T2N 1N4
* Author for correspondence (e-mail: walter{at}kin.ucalgary.ca)
Accepted 16 June 2003
![]() |
Summary |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Key words: muscle coordination, contractile condition, medial gastrocnemius muscle, soleus muscle, cat, uphill walking, downhill walking
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Although the force-sharing between cat MG and SOL has been investigated for
a variety of voluntary movements, little is known for steep uphill and
downhill walking. SOL forces from a single cat were found to decrease for
steep (37°) uphill walking compared to level walking, while MG forces
substantially increased under these conditions
(Gregor et al., 2001). It was
proposed that the decrease in SOL force with increasing uphill slope might be
caused by a resultant inhibition of SOL (i.e. smaller EMG activity) through MG
force-dependent pathways (Nichols,
1994
); however EMG activity was not measured to support this
claim. Therefore, in order to confirm the results from a single observation,
and to understand why MG forces increase and SOL forces decrease in steep
uphill walking, systematic measurements of in vivo muscle forces and
EMG activity must be made.
It is possible that the contractile conditions (length and rate of change
in length) of SOL might play an important role in the production of force,
since SOL contractile abilities may be limited under extreme conditions,
because of the dominant composition of type S fibers (Burke et al.,
1974,
1977
). The purposes of this
study were: (1) to investigate the effects of speed and intensity of
locomotion on the modulation of SOL EMG activity and force based on
statistical analysis and (2) to test the hypothesis that cat MG forces are
primarily associated with MG activation, while SOL forces are primarily
associated with the contractile conditions, rather than activation. In order
to gain novel insight into the mechanism of force-sharing between cat MG and
SOL, we performed a comprehensive study under a variety of locomotion
conditions, including steep uphill walking conditions (maximally 60°) and
galloping, while simultaneously measuring SOL and MG forces, activities and
contractile conditions, and the external kinematics and kinetics, so that the
external demands could be quantified accurately.
![]() |
Materials and methods |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
Muscle force and EMG measurement
Forces in SOL and MG were measured using E-shaped, stainless steel
tendon force transducers that were surgically implanted onto the separated
tendons of the SOL and MG. The EMG signals of these muscles were measured
using indwelling, bipolar fine wire electrodes placed into the mid-belly of
SOL and MG. The leads of all force transducers and EMG electrodes were routed
subcutaneously to a backpack connector from which all signals were transmitted
by telemetry to a custom-built amplifier. Signals were pre-amplified (gain=700
for EMG signals, and a variable, but appropriate, gain for each force) and
stored on a PC at 2000 Hz. After implantation of the force transducers and EMG
electrodes, training was continued one day following surgery. Measurement was
carried out when cats had completely recovered from surgery, such that
kinematics and kinetics (ground reaction force and resultant joint moment) of
the implanted hindlimb were similar to those obtained before surgery. This
typically happened 1 week following implantation of the force transducers and
EMG electrodes. For a detailed description of the surgical procedures, the
measurement of tendon forces and the recording of muscular EMGs, see Herzog et
al. (1993).
Kinematic and kinetic measurements
Five reflective markers (10 mm diameter) were placed over the hip, knee,
ankle, metatarsophalangeal (MP) joint and toe of the instrumented hindlimb, to
obtain knee, ankle and MP joint angles. For level, uphill and treadmill
walking, the three-dimensional positions of these markers were collected by a
motion analysis system (60 Hz; VP310, Motion Analysis Cooperation, Santa Rosa,
CA, USA). For downhill walking, movements were videotaped using a high-speed
camera (200 Hz; V-14B, NAC, Inc., Tokyo, Japan), and markers were manually
digitized using a custom-designed program written in MATLAB (Math Works, Inc.,
Natick, MA, USA). In order to avoid errors caused by skin marker movement, the
location of the knee joint center was calculated using an optimization
procedure, in which the estimated location of the knee joint center was
optimized to be closest to the measured location of the knee marker, with the
constraint that the distances from the estimated knee joint center to the
measured ankle and hip joint were the same as the measured shank and thigh
length, respectively. This optimization was performed using the MATLAB
function, `fmincon' (Math Works, Inc., Natick, MA, USA). In order to
synchronize the kinematic data with the muscle force and EMG data, a
synchronization pulse was sent from the motion analysis system to the computer
when data acquisition was started. For synchronization of the high-speed video
images with the muscle force and EMG data, a series of voltage pulses was sent
to the computer simultaneously with a light emitting diode (LED) pulse that
was recorded on the video images.
For downhill, level and uphill walking, ground reaction forces (GRFs) of the instrumented hindlimb were measured using two force platforms located in the center of a walkway (DRMC36, AMTI, Newton, MA, USA). GRFs were stored simultaneously with the muscle forces and EMG signals on a computer at 2000 Hz. All procedures were approved by the Life Sciences Animals Ethics Committee of the University of Calgary.
Joint moment, EMG and muscle length analysis
In order to identify the stance phase of locomotion, the instants of
paw-contact and paw-off were identified using the GRFs, when available, or the
video images. The resultant joint moments at ankle and knee were calculated
using the inverse dynamics approach
(Andrews, 1995) with hindlimb
kinematics and GRFs as input.
The amplified EMG signals were high-pass filtered with a cut-off frequency
of 15 Hz, and full-wave rectified. The EMG data were further processed to
determine the onset and offset of activation using a criterion based on a
minimum threshold of three standard deviations above the resting baseline for
each muscle (Neptune et al.,
1997). The magnitude of activation was quantified using the
average root mean square (RMS) values of the full-wave rectified EMG between
the onset and offset of activity, as recommended by Basmajian and De Luca
(1985
).
Muscle-tendon lengths of MG and SOL were calculated using the joint
kinematics, obtained during free locomotion and the tendon travel technique
(Grieve et al., 1978) after
all data collection was completed. Speeds of muscle shortening were calculated
as the first time derivative of muscle-tendon length using a quintic spline
function (GCVSPL; Woltring,
1986
).
Statistical analyses
In order to understand the effect of speed and intensity of locomotion on
peak muscle force, EMG activity, muscle-tendon length, and shortening speed
for MG and SOL, means of these variables from consecutive step cycles were
compared across walking conditions using a repeated-measures analysis of
variance (RM-ANOVA; SPSS Inc., Chicago, IL, USA). For analyses of the peak
muscle force and EMG activity, steps for which kinematic data were not
available were also included. The number of walking steps for the
determination of the peak muscle force and EMG activity was therefore greater
than that used for the determination of the muscle-tendon length and
shortening speed (Table 1). The
peak muscle force and the RMS value of EMG activity were normalized relative
to the mean value for slow treadmill walking (0.4-0.6 m s-1),
except for cat6, whose data were normalized relative to the mean value for
level walking on the walkway, since this animal did not perform treadmill
walking. SOL forces for cat3 and cat4 were not available, due to transducer
failure.
In order to distinguish between the effects of speed and intensity on EMG activity and muscle force, a RM-ANOVA was performed separately for the level treadmill walking conditions, for which speed was well controlled, and the sloped walking conditions, for which speed was controlled in a post-hoc manner through the stance times. For treadmill walking, not all animals walked at the same speeds (Table 1). Therefore, the number of animals varies for a given speed condition. The mean values for slow treadmill walking (0.4-0.6 m s-1; N=6) were compared to the mean values for fast treadmill walking (>0.6 m s-1; N=5). For level, uphill and downhill walking, data were obtained from six animals for all conditions, except that the steepest uphill condition (60°) was only performed by five animals (Table 1). The mean values of the target variables for 0.4-0.6 m s-1 level walking were compared with the downhill and uphill walking conditions (30, 45 and 60°).
In order to investigate whether peak SOL forces and EMG activities decrease
with increasing MG forces and EMG activities from level to uphill walking, as
suggested by Gregor et al.
(2001), correlation
coefficients between the normalized peak SOL and MG force, and between the
normalized SOL and MG EMG activity were calculated for consecutive steps,
while animals walked on the level part of the walkway before walking up the
sloped walkway, and then as they walked on the level surface at the top of the
sloped part of the walkway. The numbers of walking steps for each
level-uphill-level walking condition (30, 45 and 60°) are summarized in
Table 2.
|
![]() |
Results |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
MG forces were greater for uphill walking than those for any of the other walking conditions (Fig. 1E). In contrast, SOL forces for the different animals were either similar or smaller for uphill than those for the other walking conditions (Fig. 1F). The ankle plantar flexor moments increased from level to uphill walking (Fig. 1G), thus MG forces increased with increasing movement demands, whereas SOL forces did not. However, not only the changes in magnitude, but also the MG force-time histories for the entire stance phase (Fig. 1E) were similar in shape to the ankle moment curves (Fig. 1G), whereas the SOL force-time curves (Fig. 1F) bore little resemblance to the ankle moment curves (Fig. 1G) but did not resemble the knee moment curves (Fig. 1H).
Peak MG forces increased significantly from slow walking (0.4-0.6 m s-1) to fast walking (0.8-1.2 m s-1) and galloping, and also increased significantly from downhill to level, and uphill walking (Fig. 2A). These increases in MG force were associated with increases in EMG activity (Fig. 2B). In contrast, the mean SOL peak forces did not change (Fig. 2C), while SOL EMG activity significantly increased with increasing speeds and intensities of locomotion (Fig. 2D). The increase in normalized peak force and EMG activity across all locomotor conditions was 8 and 6 times for MG (Fig. 2A,B) and approximately 1.4 and 1.4 times for SOL, respectively (Fig. 2C,D).
|
The relationships between the peak muscle forces and the corresponding RMS values of the EMG for MG and SOL are shown in Fig. 3 for downhill, level+treadmill (0.4-1.2 m s-1), and uphill walking, as well as galloping. MG peak forces and EMG activities had a strong positive correlation for all animals (r2=0.66-0.90), while SOL peak forces and activities were not correlated in a systematic way (positive and negative regression slopes, r2=0.03-0.17).
|
Coordination of MG/SOL force and activation between level and uphill
walking
In consecutive step cycles of uphill walking, SOL forces were small when MG
forces were great, and vice versa
(Fig. 4). Systematic analysis
revealed that the peak MG and SOL forces were negatively correlated
(r=-0.50 for 30°; r=-0.61 for 45°; and
r=-0.56 for 60°; Fig.
5 and Table 2),
while the corresponding EMG activities were positively correlated
(r=0.24 for 30°; r=0.58 for 45°; and r=0.43
for 60°; Fig. 5 and
Table 2), indicating that
greater activation of MG, and greater force in MG, were associated with
increased SOL activation, but decreased SOL force.
|
|
Changes in contractile conditions
Muscle stretching from paw contact to the instant of peak muscle force
(stretching is negative; Fig.
6) was significantly greater for downhill than for level and
uphill walking. This difference in muscle excursion was associated with
significant differences in the instantaneous velocity of shortening at the
instant of peak force production (stretching is negative;
Fig. 7). For all cats, MG and
SOL were stretched at the instant of peak force production for downhill
walking, they were about isometric for level walking, and were shortening for
all animals for the uphill walking conditions
(Fig. 7).
|
|
![]() |
Discussion |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Effect of speed and intensity of locomotion on the modulation of SOL
force and EMG activity
SOL EMG activity significantly increased with increasing speeds and
intensities of locomotion (Fig.
2D). Hence, the net change in SOL activity (excitation and
inhibition) was excitatory for fast walking/trotting, galloping and uphill
walking compared to slow walking. In contrast, SOL peak forces were found to
remain constant for the conditions tested here. The force results are
consistent with previous studies, which suggested that SOL peak forces remain
nearly constant for a wide range of walking and trotting speeds
(Walmsley et al., 1978;
Hodgson, 1983
). This result
has been interpreted as a saturation of SOL activation at low speeds of
locomotion (Walmsley et al.,
1978
; Pierotti et al.,
1989
). However, we found a statistically significant increase in
SOL activation with increasing speeds and intensities of locomotion,
suggesting that SOL activation is not maximal (or saturated) at slow speeds of
walking. Similarly, Walmsley et al.
(1978
) showed a dramatically
increased SOL EMG during the take-off phase of jumping compared to all
locomotor conditions (their fig. 8), thereby providing strong evidence against
their own argument that SOL activation was maximal at slow speeds of walking.
Therefore, we conclude that SOL is not fully activated at slow speeds of
walking. Of course, our results do not exclude the possibility of SOL
inhibition through speed- (Hodgson,
1983
) or force-dependent pathways
(Nichols, 1994
;
Gregor et al., 2001
). However,
any such inhibition would be offset by an even greater excitation associated
with increased movement demands. Our results also do not exclude the
possibility that a net speed-dependent inhibition might exist for movements
performed at greater speeds than galloping, for example, during a paw shake or
a scratch response; i.e. movements that are performed at frequencies of 5-10
Hz (Smith et al., 1980
;
Abraham and Loeb, 1985
;
Fowler et al., 1988
; M. Kaya,
T. Leonard and W. Herzog, unpublished observations).
It has been suggested that SOL activity is strongly inhibited by increasing
MG force in decerebrate cats (Nichols,
1994). This force-dependent inhibition of SOL has been used by
Gregor et al. (2001
) to
explain the decrease in SOL force with increasing MG force for uphill walking
compared to level or downhill walking. Our results did not show a
statistically significant decrease in SOL force from level (0.4-0.6 m
s-1) to uphill walking, although three out of the four animals for
which force recordings were available in our study, showed substantial
decreases in the peak SOL forces, as did the single animal tested by Gregor et
al. (2001
). Also, peak SOL and
MG forces were negatively correlated for consecutive step cycles of uphill
walking (Fig. 5). Thus, our
results are consistent with those of Gregor et al.
(2001
), and they suggest that
increasing MG forces are typically associated with decreasing SOL forces.
However, SOL activity increased significantly with increasing MG activity
(Fig. 5) and MG force (not
shown) for all uphill walking conditions, suggesting that SOL forces did not
remain constant (one animal) or decrease (three animals) because of a decrease
in activation, but despite a net increase. Again, our measurements do not
distinguish between inhibitory and excitatory signals to SOL. Thus, a strong
MG force-dependent inhibition (Nichols,
1994
) may have been present in SOL for the uphill walking
conditions, but this inhibition (if present) was offset by an even greater
increase in excitation. Unfortunately, the only other force data for uphill
walking (Gregor et al., 2001
)
cannot be used for resolving this issue, as no EMG data were collected in that
work.
MG force depends primarily on MG activation, while SOL force
primarily depends on SOL contractile conditions
The instantaneous force in any muscle depends on the degree of activation
(e.g. Rack and Westbury, 1969)
and the contractile conditions (e.g. length,
Gordon et al., 1966
; velocity,
Hill, 1938
; history,
Abbott and Aubert, 1952
). In
this study, we found that peak MG forces and EMG activities increased with
increasing speeds and intensities of movement
(Fig. 2), and that they were
strongly correlated for all animals (Fig.
3). These results suggest that MG forces are primarily associated
with MG activation, and are not greatly influenced by the contractile
conditions (at least not for the movements tested here). In contrast, SOL
forces were not systematically correlated with EMG activity (Figs
2 and
3). This tendency is
exemplified in Fig. 5, where
peak SOL forces decreased with increasing SOL activity. These results suggest
that SOL forces are influenced to a great degree by the contractile
conditions, and that the increases in SOL activity with increasing movement
demands keep SOL forces constant in the presence of less favorable contractile
conditions (i.e. increasing speeds of shortening for uphill compared to level
and downhill walking; Fig.
7).
MG and SOL are stretched from paw contact to the instant of peak muscle
force for downhill and level walking, while they shorten for uphill walking
(Fig. 6). This active
lengthening of SOL for downhill and level walking enhances SOL force
production because of the active lengthening of the muscle and the
force-velocity properties (Hill,
1938). Thus, the substantial decrease in peak SOL force, despite
an increase in activation, for uphill walking compared to level walking found
in three animals (Fig. 2) is
probably associated with the increase in SOL shortening velocity for the
uphill compared to the level walking conditions.
Modulation of MG and SOL coordination
Changes in the amplitude of peak force and EMG activity for a range of
movements were different between MG and SOL. MG peak force and EMG activity
increased 2-8 times from slow walking (0.4-0.6 m s-1) to fast
walking, trotting, uphill walking, or galloping. Changes in the peak SOL force
and EMG activity were within about ±50% of that for slow walking
(Fig. 2). MG is primarily
composed of fast-twitch fibers (Burke et
al., 1977) and has a maximum isometric force capacity of about 100
N (Spector et al., 1980
; M.
Kaya et al., unpublished results). The 15-20 N of MG force during slow walking
is only about 15-20% of the isometric force capacity, therefore MG's force and
activation is only a small percentage of its full potential. In contrast, SOL
is primarily composed of slow-twitch fibers
(Ariano et al., 1973
;
Burke et al., 1977
) and has a
maximum isometric force capacity of about 30 N
(Spector et al., 1980
;
Herzog et al., 1992
;
Scott et al., 1996
).
Therefore, the 15-20 N of SOL force during slow walking constitutes a
substantial percentage (
50%) of SOL's potential. Therefore, a substantial
increase in SOL force above that obtained for slow walking is only possible if
activation goes towards maximum, or if the muscle is stretched at full
activation. Although the size principle of motor unit recruitment is not
applicable across muscles, but only within a given muscle (see paw-shake
response in cat SOL and MG; Smith et al.,
1980
; Abraham and Loeb,
1985
), it is feasible to assume that the 95-100% slow SOL is
recruited to a much greater extent for standing and slow walking than the
70-80% fast MG. Consequently, there is less room for SOL force and activity
modulation from the slow walking baseline, compared to MG force and activity
modulation.
Mechanical interpretation of MG and SOL
As shown in Fig. 1, the MG
force-time curves were similar in shape to the resultant ankle joint moment
curves, whereas the SOL force-time curves showed little resemblance to them.
These results suggest that MG force production is closely related to the ankle
extensor moments, whereas SOL force production appears to be largely
independent of these moments. Gregor et al.
(2001) suggested that uphill
walking constitutes an ideal situation for the two-joint MG, because the
initial phase of uphill walking is associated with a knee flexor and an ankle
extensor moment (Fig. 1G,H).
However, careful analysis reveals that MG forces and ankle extensor moments
increase systematically with increasing slopes of uphill walking, but knee
flexor moments do not (Fig.
1E,H). Therefore, our results are consistent with the idea that MG
acts like an ankle extensor with little regard to satisfying movement demands
at the knee.
![]() |
Conclusion |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
Acknowledgments |
---|
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Abraham, L. D. and Loeb, G. E. (1985). The distal hindlimb musculature of the cat (patterns of normal use). Exp. Brain Res. 58,580 -593.
Andrews, J. G. (1995). Euler's and Lagrange's equations for linked rigid-body models of three-dimensional human motion. In Three-Dimensional Analysis of Human Movement (ed. P. Allard, I. A. F. Stokes and J. P. Blanchi), pp.145 -175. Champaign, IL: Human Kinetics.
Ariano, M. A, Armstrong, R. B. and Edgerton, V. R. (1973). Hindlimb muscle fiber populations of five mammals. J. Histochem. Cytochem. 21, 51-55.[Medline]
Abbott, B. C. and Aubert, X. M. (1952). The force exerted by active striated muscle during and after change of length. J. Physiol. 116,77 -86.
Basmajian, J. V. and De Luca, C. J. (1985). Description and analysis of the EMG signal. In Muscle Alive. Their Function Revealed by Electromyography. 5th edn, pp.65 -100. Baltimore, MD: Williams and Wilkins.
Burke, R. E., Ten Bruggencate, G. and Jankowska, E. (1970). A comparison of peripheral and rubrospinal input to slow and fast motor units of triceps surae. J. Physiol. 207,709 -732.[Medline]
Burke, R. E., Levine, D. N., Salcman, M. and Tsairis, P. (1974). Motor units in cat soleus muscle: physiological, histochemical and morphological characteristics. J. Physiol. 238,503 -514.[Medline]
Burke, R. E., Strick, P. L., Kanda, K., Kim, C. C. and Walmsley,
B. (1977). Anatomy of medial gastrocnemius and soleus motor
nuclei in cat spinal cord. J. Neurophysiol.
40,667
-680.
Fowler, E. G., Gregor, R. J., Hodgson, J. A. and Roy, R. R. (1993). Relationship between ankle muscle and joint kinetics during the stance phase of locomotion in the cat. J. Biomech. 26,465 -483.[Medline]
Fowler, E. G., Gregor, R. J. and Roy, R. R. (1988). Differential kinetics of fast and slow ankle extensors during the paw-shake in the cat. Exp. Neurol. 99,219 -224.[Medline]
Gordon, A. M., Huxley, A. F. and Julian, F. J. (1966). The variation in isometric tension with sarcomere length in vertebrate muscles. J. Physiol. 184,170 -192.[Medline]
Gregor, R. J., Roy, R. R., Whiting, W. C., Lovely, R. G., Hodgson, J. A. and Edgerton, V. R. (1988). Mechanical output of the cat soleus during treadmill locomotion: In vivo vs in situ characteristics. J. Biomech. 21,721 -732.[Medline]
Gregor, R. J., Smith, J. L., Smith, D. W., Oliver, A. and Prilutsky, B. I. (2001). Hindlimb kinetics and neural control during slope walking in the cat: Unexpected findings. J. Appl. Biomech. 17,277 -286.
Grieve, D. W., Pheasant, S. and Cavanagh, P. R. (1978). Prediction of gastrocnemius length from knee and ankle joint posture. In Biomechanics VI-A (ed. E. Asmussen and K. Jorgensen), pp. 405-418. Baltimore, MD: University Park Press.
Herzog, W., Leonard, T. R., Renaud, J. M., Wallace, J., Chaki G. and Bornemisza, S. (1992). Force-length properties and functional demands of cat gastrocnemius, soleus, and plantaris muscles. J. Biomech. 25,1329 -1335.[Medline]
Herzog, W., Leonard, T. R. and Guimaraes, A. C. S. (1993). Forces in gastrocnemius, soleus, and plantaris tendons of the freely moving cat. J. Biomech. 26,945 -953.[Medline]
Hill, A. V. (1938). The heat of shortening and the dynamic constants of muscle. Proc. R. Soc. Lond. B 126,136 -195.
Hodgson, J. A. (1983). The relationship between soleus and gastrocnemius muscle activity in conscious cats - a model for motor unit recruitment? J. Physiol. 337,553 -562.[Abstract]
Neptune, R. R., Kautz, S. A. and Hull, M. L. (1997). The effect of pedaling rate on coordination in cycling. J. Biomech. 30,1051 -1058.[CrossRef][Medline]
Nichols, T. R. (1994). A biomechanical perspective on spinal mechanisms of coordinated muscular action: An architecture principle. Acta Anat. 151, 1-13.[Medline]
Pierotti, D. J., Roy, R. R., Gregor, R. J. and Edgerton, V. R. (1989). Electromyographic activity of cat hindlimb flexors and extensors during locomotion at varying speeds and inclines. Brain Res. 481,57 -66.[CrossRef][Medline]
Rack, P. M. and Westbury, D. R. (1969). The effects of length and stimulus rate on tension in the isometric cat soleus muscle. J. Physiol. 204,443 -460.[Medline]
Spector, S. A., Gardiner, P. F., Zernicke, R. F., Roy, R. R. and
Edgerton, V. R. (1980). Muscle architecture and
force-velocity characteristics of cat soleus and medial gastrocnemius:
Implications for motor control. J. Neurophysiol.
44,951
-960.
Scott, S. H., Brown, I. E. and Loeb, G. E. (1996). Mechanics of feline soleus: I. Effect of fascicle length and velocity on force output. J. Muscle Res. Cell Motil. 17,207 -219.[Medline]
Smith, J. L., Betts, B., Edgerton, V. R. and Zernicke, R. F. (1980). Rapid ankle extension during paw shakes: Selective recruitment of fast ankle extensors. J. Neurophysiol. 432,612 -620.
Walmsley, B., Hodgson, J. A. and Burke, R. E.
(1978). Forces produced by medial gastrocnemius and soleus
muscles during locomotion in freely moving cats. J.
Neurophysiol. 41,1203
-1216.
Whiting, W. C., Gregor, R. J., Roy, R. R. and Edgerton, V. R. (1984). A technique for estimating mechanical work of individual muscles in the cat during treadmill locomotion. J. Biomech. 17,685 -694.[Medline]
Woltring, H. J. (1986). A FORTRAN package for generalized, cross-validatory spline smoothing and differentiation. Adv. Eng. Software 8,104 -113.