| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
MS 0317 Received 11 November 1999; accepted after revision 23 March 2000.
 |
ABSTRACT |
|---|
 TopAbstract IntroductionMethodsResultsDiscussionReferences |
|---|
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The role of reflexes in the normal generation of walking has been debated for nearly a century. Sherrington (1910) suggested that walking could be produced entirely by a series of reflexes, but his contemporary, Brown (1911), provided the first evidence for a central pattern generator (CPG). Brown suggested that a CPG might produce the basic rhythmic movement, even in the absence of reflexes from sensory feedback. This was confirmed by later studies in which animals had been deafferented or paralysed with curare (reviewed in Rossignol, 1996). In the paralysed animal stimulation of a mid-brain centre, the mesencephalic locomotor region, produces a pattern of nerve activity similar to that seen in normal walking, even though no muscle activity is possible (often referred to as 'fictive locomotion', Shik et al. 1966). Although the spinal cord is sufficient to produce the basic rhythm in the absence of sensory feedback, feedback can elaborate this pattern, reset the rhythm, and generally adapt the output to the terrain over which an animal is walking. For example, even spinal walking preparations can adjust the rhythm to the speed of a treadmill belt over which they are suspended (Grillner, 1981). The relative importance of the CPG and reflex inputs in producing the EMG and force patterns during locomotion varies with the species of animal, the speed and type of locomotion (walking, running, swimming) and the part of the cycle considered. For example, the swing phase of walking (when the foot is in the air) varies less than the stance time (when the foot is on the ground), as the overall speed of locomotion changes (Rossignol, 1996). However, the swing phase can be modified dramatically when the foot hits an obstacle (reviewed in Zehr & Stein, 1999).
This paper will concentrate on EMG and force production in the stance phase. The force produced by the ankle extensor muscles, for example, will depend on three factors: (1) the muscle activation produced by the CPG, (2) the reflexes from muscles and other receptors and (3) the intrinsic properties of the muscle, such as the force-length and force-velocity relationships. The purpose of the paper is to quantify for the first time the relative importance of phasic reflexes evoked by length changes in force production during the stance phase.
The 'stance' phase can sometimes be confused with the 'extensor' phase defined by Phillipson (1905). During the last part of swing, the limb begins to extend passively and then actively (termed E1 by Phillipson, 1905). This period of extension ends when the foot hits the ground at the beginning of the stance phase. The weight of the body then flexes the limb even though extensor muscles are active (E2). Finally, the limb begins to extend again to push the body up and forward (E3) until the limb leaves the ground at the end of the stance phase.
Stretch reflexes from muscle spindles would be expected to be most important during the E2 phase when the extensor muscles are being stretched. Research on decerebrate cats also suggests that Golgi tendon organs can contribute to supporting the body weight during the stance phase (Pearson & Collins, 1993; McCrea et al. 1995). Loading the limb during the stance phase produces an excitation mediated, at least in part, by the Ib afferents from Golgi tendon organs. This represents a reversal of the well-known disynaptic inhibition that is produced by Ib afferents under static conditions. The magnitude of the effect in reduced preparations is substantial and can prolong the stance phase until the limb is eventually unloaded and swing can proceed. Hiebert & Pearson (1999) reported that completely unloading the limb by allowing it to go into a hole or deafferenting the limb reduces the EMG in ankle and knee extensors to less than half. Thus, removal of most or all length and force-related sensory information indicates that feedback may be responsible for the majority of activation in extensor motoneurones during the stance phase of walking. The specific question we wish to address is: How much of the force over the whole stance phase is due to phasic reflexes evoked by muscle length changes?
Previous studies have generally used stretches with a particular waveform, such as a pulse or ramp and made the dubious assumption that the effect of such a stretch could be scaled linearly to predict the responses to a larger, longer series of stretches and releases over the whole stance phase (e.g. Yang et al. 1991; Sinkjær et al. 1996). Furthermore, EMG, rather than torque, has usually been measured, but the relation between EMG and torque in moving muscle is complex. For example, Kearney et al. (1999) simulated the step cycle in supine human subjects by moving the ankle through a trajectory similar to that found in normal walking. They added small, brief perturbations at various times in the cycle and found surprisingly that the torque produced was very small at the time when the largest EMG was generated. This discrepancy resulted from force-velocity and other non-linear properties of the muscles.
In the present study we have used the high decerebrate (mesencephalic) walking cat to estimate the relative percentages of ankle extensor torque produced by phasic reflexes and by other factors such as the CPG during the stance phase of the step cycle. Walking continues in these animals, even if one limb is largely denervated and the length of the ankle extensor muscles is controlled in various ways by a puller. In successive steps the ankle extensors were either held at a constant length (isometric) or moved (simulated walking) according to the length changes that are observed in normal walking (Goslow et al. 1973). The difference in the two conditions gives an estimate of the net force produced by muscle reflexes evoked by length changes in the walking animal for the first time. In the Discussion we will compare these values to those previously estimated for cats and humans.
|
METHODS |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Preparation
The experiments were performed using seven adult cats. The procedures were approved by the University of Alberta Health Sciences Laboratory Animal Welfare Committee. Under halothane anaesthesia (delivered in a mixture of 95 % O2 and 5 % CO2) the trachea of each animal was cannulated for the continued administration of anaesthetic. A carotid artery was cannulated to monitor blood pressure. A jugular vein was cannulated for administration of fluids and drugs. Both carotid arteries were ligated to reduce blood loss during decerebration. A heating pad maintained body temperature during the dissection.
In all animals the following nerves were transected in the right hindleg: femoral, obturator, sartorius, semimembranosus, semitendinosus, posterior biceps, common peroneal and caudal cutaneous sural. The tibial nerve distal to the branches serving triceps surae muscles was ligated and crushed, rather than transected. This reduced the tension on the triceps surae branches of the tibial nerve that can be produced with retraction of the tibial nerve following transection of the distal portion. The branch of the sciatic nerve that serves the anterior portion of biceps femoris (AB) was transected in five animals. In the remaining two animals this branch was left intact to allow recording of activity in AB. The tendons serving the lateral gastrocnemius, medial gastrocnemius and soleus muscles were identified as they inserted onto the calcaneum. All other tendons and connective tissue converging onto the calcaneum were transected. The calcaneum was severed close to the ankle joint. A heavy-gauge steel wire was passed through a hole in the detached fragment of the calcaneum, which in turn was attached to a stainless steel wire and secured to a muscle puller (Cambridge Instruments series 300B servomotor). Tension and length records were obtained from the instrumentation of the motor. The length of the triceps surae was specified relative to the length when the knee and ankle angles were 90deg. This reference length was taken as '0' mm.
The animal was then placed above a treadmill with its head secured in a stereotaxic holder. The right hindleg was immobilized with the use of knee and ankle clamps. The hip was secured in position with the use of a heavy wire that passed through both iliac crests and a steel flange placed over the sacrum. The flange was then secured to the mounting frame with a bolt. Each joint was fixed at approximately 90 deg. A temperature probe was placed subcutaneously alongside the triceps surae of the right leg. Radiant heat was applied to maintain a temperature of 37°C. Once the animal was secured over the treadmill, decerebration was performed by removing the cerebral hemispheres and then transecting the brainstem at a 50 deg angle, rostral to the superior colliculi. The anaesthetic was discontinued at this point. A bolus (2-5 ml) of volume expander (Dextran, Baxter) was routinely administered. Typically within 1 h of the decerebration the animals would produce spontaneous bouts of walking. In two animals walking was evoked with electrical stimulation of the mesencephalic locomotor region, as described by Shik et al. (1966).
Recording and stimulation
In all cats electromyographic (EMG) recordings were made from soleus (SOL), medial gastrocnemius (MG), lateral gastrocnemius (LG), and iliopsoas (IP) of the right leg, and medial gastrocnemius (coMG) of the left leg. The records from IP, coMG and AB (in two animals) were used as indicators of the general locomotor rhythm. The EMGs were recorded using a pair of stainless steel wires (Cooner Wire, AS632) implanted into the bellies of the muscles. The wires were Teflon-coated except for a 5 mm bared region. The wires of a recording pair were separated by a distance of about 5 mm within the muscle.
At the termination of each experiment a bipolar cuff electrode was placed around the sciatic nerve, just distal to the branching of the nerves to the hamstring muscles. The sciatic nerve was then crushed proximally. The nerve to AB was cut at this time, if necessary. The sciatic nerve was stimulated with 40 Hz trains of 0·2 ms square-wave pulses to activate the triceps surae. The rate of stimulation can affect not only force, but its dependence on length and velocity (Joyce & Rack, 1969; Joyce et al. 1969; Rack & Westbury, 1969). The selected rate provided a largely fused contraction without as much fatigue as at higher stimulus rates. The intensity was varied so as to elicit a range of contraction strengths and the duration was matched to the duration of the triceps surae contractions produced during the bouts of decerebrate locomotion. Trains were delivered at a rate of about 0·5 Hz.
Protocol
During periods of locomotor activity (occurring either spontaneously or in response to electrical stimulation of the mesencephalic locomotor region) the isolated triceps surae contracted rhythmically. On every fourth cycle the length of the muscle was varied to simulate the length changes that occur during extensor activity in an intact walking cat (Fig. 1), as published by Goslow et al. (1973). The EMG from the hip flexor (IP) was rectified and filtered. The onset of this burst was used to trigger the lengthening of the muscle to the length (+2 mm) found at the end of the flexion phase. The EMG burst of an ankle extensor muscle (typically soleus) was then used to trigger the length changes in soleus described by Goslow et al. (1973) for the extensor phase (see Fig. 1). In normal walking the triceps surae muscles begin to shorten passively, somewhat before the onset of EMG activity. However, the need for triggering prevented us from matching this feature of the normal pattern. As a result there was initially some extra EMG activity and force generation in the E1 part of the extensor phase. Normally, the triceps surae will shorten with no load until the foot hits the ground and begins to support the body weight (E2). Because of these methodological problems, we have analysed only the data for E2 and E3. Since there is normally no external force generation in E1 and the EMG is relatively flat over most of the extensor phase (Fig. 2), this should not affect our conclusions about the percentage of force generated during the stance phase of walking.
Once the muscle reached its shortest length, it was again lengthened after a brief delay and held isometric for several cycles until the next period of simulated walking was initiated. For each sequence of data collection the isometric length was set to a value between -6 mm and +2 mm between the cycles of simulated walking. A similar protocol was followed when the sciatic nerve was electrically stimulated to evoke rhythmic contractions in the triceps surae, except that the sequence was triggered by the stimulation, rather than spontaneous bursts of EMG activity.
Data acquisition and analysis
The EMGs were amplified and filtered (30-10000 Hz bandpass, P511 amplifier, Grass Instruments) prior to storage to a magnetic tape (VHS, Vetter 4000A PCM recording unit). The length and tension signals from the motor were also stored. Data were replayed from the tape recorder, full-wave rectified and low-pass filtered. The processed data were sampled using Axotape software (Axon Instruments) at a frequency of either 500 or 200 Hz. The sampled data were analysed using a specially written program under Matlab (Math Works, Natick, MA, USA). In this program the data could be filtered further using the decimate function which reduced the effective sample rate to 100 or 200 Hz. Markers were then placed manually at the end of the IP burst to mark the transition from the flexor phase to the extensor phase of the rhythm. Markers were placed separately for data in which the triceps surae muscles were either held isometric or cycled in length to simulate normal walking.
For the isometric trials, we tried to select cycles where there were no length changes simulating walking over a period of 500 ms before or 800 ms after the marker, to minimize any effects of prior or later movements. In the simulated walking trials only EMG bursts were marked that were well aligned with the simulated walking. In many experiments this was true of nearly all walking cycles, but in others there were some erratic bursts or the triggering of the simulated walking was not accurate and cycles had to be excluded. Typically, there were 10-30 cycles of each type available for averaging. The two sets of cycles (isometric and walking) were then averaged separately and superimposed as shown in Fig. 2. Walking cycles are shown as continuous lines and the standard errors above and below the mean are indicated by dashed lines. The isometric cycles are shown as dotted lines. The standard errors for these cycles were similar to those for the walking cycles, but have been omitted for clarity.
To further analyse the averages, we subdivided the cycles into 150 ms segments. The first segment ended at the end of the E1 phase. This was determined manually by positioning a cursor on the appropriate point in the trace. There was some variability in the timing of the E1 phase, but this procedure was accurate to one or two time samples. EMG and force were compared during the simulated walking and isometric conditions at the same average length in each segment (Fig. 3). Over these segments the muscle length could change by 3 mm and force varied considerably as well. Shorter segments could have been used, but this would have increased the variability because of the differences in the time course from step to step.
|
RESULTS |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Figure 1 shows data from a mesencephalic cat walking spontaneously on a treadmill. Most muscles of the leg were denervated except the muscles of interest (see Methods). EMGs from a hip flexor muscle, iliopsoas (IP), and an ankle extensor muscle, medial gastrocnemius (MG), are illustrated. An alternating pattern is observed, as expected, with five bursts of MG separating six bursts of IP. In the three central step cycles, the length of the triceps surae muscles, including the MG muscle, was held isometric at a length of -3 mm with respect to the length at which the ankle and knee were at a right angle (90 deg). The leg was clamped at the knee and hip so that the entire leg was essentially isometric. The other legs were free to move and contact the moving treadmill belt over which the cat was suspended. In the other two steps the length was varied in a pattern that approximately matched that found during normal walking (Goslow et al. 1973). When the IP became active, the muscle was extended to +2 mm (Fig. 1, up arrow). During the phase when the MG was active, the muscle was first shortened (Fig. 1, down arrow). It was then lengthened, as occurs when stretched under the weight of the body during the normal step cycle, and finally shortened further to -8 mm. The peak MG EMG and force generated in these steps was markedly increased.
 |
View larger version [in this window] [in a new window] |
|
|
The rectified and smoothed EMGs show alternation between a hip flexor muscle (IP) and an ankle extensor muscle (MG). In some step cycles the length was varied by a torque motor as occurs during normal walking (simulated walking), while in other steps the muscle was held at a constant length (isometric). Clear differences are seen in the force generated and the MG EMG. | ||
To examine the effect of the length changes in more detail, cycles were averaged with respect to the end of the burst of IP EMG, which is shown as time 0 in Fig. 2. Averages of isometric and simulated walking trials are superimposed for the parameters shown in Fig. 1 and some additional parameters. For ease of comparison each parameter has been normalized to the range shown in the simulated walking trials. The average EMGs for IP and the contralateral MG (coMG) muscles virtually superimpose with and without length changes. In other words, the length change did not modify the ongoing locomotor pattern generation in the cat's leg, as will be shown in more detail later. However, the three heads of the triceps surae muscles (MG, LG and SOL), that were being stretched, showed at least two periods of increased activity. Firstly, the muscles become active somewhat earlier and more forcibly at the longer length characteristic of the late swing phase of the walking pattern. This increased activity presumably results from the increased afferent activity at the longer length.
 |
View larger version [in this window] [in a new window] |
|
|
The dashed lines indicate the standard errors for the simulated walking trials (continuous lines). Note that the EMG is increased in only those muscles (MG, LG and SOL) that are being stretched and that the net force is increased, even though the muscles are shortening through much of the stance phase. Each trace has been normalized to the same vertical size for the purposes of display. The minimum and maximum values are shown for the force (in N) and the length (in mm). The EMG is in arbitrary units. E1-E3 represent the three parts of the extensor phase described by Phillipson (1905), as discussed further in the text. | ||
Secondly, an increased activity is seen when the muscles are being stretched (E2), as mentioned above. The mean and the standard error of the mean are shown for 26 simulated walking trials. The dotted trace is the mean of 26 trials in which the muscles were held isometric. The corresponding length and force records are shown in the lower two traces of the figure. Two periods of extra force generation are seen, resulting at least in part from the increased EMG. As the muscle shortens further (E3) the force declines to below isometric levels. The decline is due both to the force-length and force-velocity properties of the triceps surae muscles, since the EMG values are not reduced. However, the decrease is much less than the increase seen earlier, so the triceps surae muscles generated substantially more force overall during the period of EMG activity.
Estimating the effect of phasic reflexes at matched lengths
To quantify the effect of muscle length, we divided the data into four 150 ms segments as shown in Fig. 3. The first segment was chosen to finish at the end of the E1 phase (see Methods); the second segment corresponds to the E2 phase and the beginning of the E3 phase; the E3 phase extended through the third and fourth segments. The force generated under isometric conditions depends markedly on the length, as expected from the length-tension curve. However, the force during the simulated walking for these four data sets was largely unaffected, except in the first segment where there was a small residual effect from the isometric level at which the muscle had been held previously. Segment 1 corresponds to the end of the swing phase when the muscles are not normally generating net forces, so this segment will be ignored (see Protocol section of Methods).
 |
View larger version [in this window] [in a new window] |
|
|
The full period has been divided into four segments of 150 ms each (1-4), as described in the text, so that force production in each segment can be compared at the average length for the muscles during simulated walking. Length is measured relative to that at which the ankle and knee are at right angles. The different symbols are for lengths equal to 0 (+), -2 (
), -6 ( | ||
Before examining the mechanisms and the magnitude of the force and EMG effects in triceps surae, it should be noted that other muscles were not affected by the simulated walking or isometric trials. This is shown for the IP and the anterior biceps (AB) muscles in the same leg and the MG muscle in the contralateral leg in Fig. 4. A lack of effect was already noted qualitatively for two of the muscles at one length in Fig. 2, but Fig. 4 represents the average of six data sets for a variety of lengths for three muscles in a different animal. Similar results were found in all animals. AB is of particular interest, since the activity of AB motoneurones is enhanced by stimulation of triceps group I afferents during fictive locomotion in the decerebrate cat (Guertin et al. 1995).
 |
View larger version [in this window] [in a new window] |
|
|
The simulated walking ( | ||
The dependence of force and MG EMG on the isometric muscle length is shown more quantitatively in Fig. 5. In this experiment the muscle was held isometrically (
) at a variety of lengths from -6 mm to 0 mm in between the simulated walking trials (
). As the muscle is shortened, the force decreases for all segments shown. Straight lines have been fitted through the isometric data points. The slopes are similar and represent on average a 16 % decline in force for each millimetre shortening below 0 length in this cat. This was close to the value (16 ± 2 %; mean ± S.E.M.) for all the cats studied. The dependence was less for the MG EMG in this cat (7 % on average). Overall, the values were 5 ± 2 % (MG) and 4 ± 2 % (LG, not shown), which were significantly different from 0 (P < 0·05, using Student's t test). The SOL EMG had a similar trend (not shown), but was not statistically significant.
 |
View larger version [in this window] [in a new window] |
|
|
The isometric length ( | ||
Also shown in Fig. 5 are the values of force and EMG in each segment for the simulated walking trials (). The values are plotted at a length corresponding to the average length of the muscles during that segment. Note that the values of force are generally higher than would be predicted from the isometric condition, because of the extra reflex activation associated with the length changes, as shown in Figs 1 and 2. Figure 6 shows the average ratios (simulated walking/isometric trials) for all the cats studied. The ratios for EMG (1·42-1·73) and force (1·35) during segment 2 were all above 1·0. Thus, about 50 % more EMG and force were observed during segment 2 for the simulated walking trials than would be predicted from isometric conditions at a matched muscle length. The values during segment 3 are around 1·0 (range between 0·79 and 1·20), except for the adjusted force. This adjustment will be explained in the next section. By segment 4 the values of force and EMG had declined and were more variable in amplitude and time course as the end of the burst approached. Therefore, they are not shown, but we have included values for the overall period covering segments 2-4, which is essentially the whole simulated stance phase. The ratios for MG, LG and force, but not SOL, are significantly greater than 1·0 (P < 0·05, Student's t test).
 |
View larger version [in this window] [in a new window] |
|
|
Ratios of values during simulated walking relative to isometric trials for the EMGs of the triceps surae muscles and force (first four bars of each group). The final bar shows the average force ratio after making an adjustment (adj) for the force-velocity and other intrinsic properties of muscle, as described in Fig. 7. The average ratios ± S.E.M. for all muscles are shown separately for segments 2 and 3 (as defined in Fig. 3) and for the whole stance phase (segments 2-4). | ||
Comparison to force production without reflexes
The ratio for force (simulated stepping/isometric trials) varied between the segments, but when averaged over segments 2-4 was 1·19 ± 0·08 (mean ± S.E.M.) in Fig. 6. This suggests that about 20 % more force was produced during the simulated stance phase, when the muscle was moving and producing phasic reflexes, compared with isometric trials at a matched average length. However, force will depend on other factors such as the force-velocity properties of the muscles. Since the muscles were shortening during the E3 phase of stance, the force could be considerably smaller with the same amount of activation. To test this possibility, we cut the sciatic nerve at the end of each experiment and stimulated the distal end at a constant rate (see Methods). The lower leg was denervated except for the triceps surae muscles, so only these muscles contracted. In Fig. 7 the sciatic nerve was stimulated at 40 Hz for 0·5 s at 1·5 × motor threshold which produced a similar force (No reflexes) to that seen in the same experiment during locomotion (With reflexes).
The stimulation was repeated for a variety of isometric lengths, but for clarity only -2 mm (
) and -4 mm (
) are shown in Fig. 7. Isometric stimulation was interposed between cycles in which the length changes simulating walking movements were applied (continuous lines in Fig. 7 show the averages of the corresponding cycles). Note that the force in the simulated E2 phase is much greater when reflexes are present, in comparison to the isometric force at a matched length (l = -2 mm). The force is also more prolonged, crossing the value for l = -4 mm near 450 ms, instead of near 350 ms. To adjust the force for these effects we did the full analysis shown in Fig. 5 to match lengths in each 150 ms segment. Then, we computed an adjusted force:
 |
View larger version [in this window] [in a new window] |
|
|
After recording the responses with reflexes in the walking cat, the sciatic nerve was cut and stimulated distally to produce levels of force equivalent to those generated during the previous walking trials. The muscle was either cycled through length changes to simulate walking or held at various isometric lengths. Only isometric lengths of -2 ( ) mm are shown here for clarity. The continuous lines are average forces for the trials in which the length changes were applied. E1-E3 refer to the extensor phases of Phillipson (1905). | ||
This analysis was done for all cats and the average force ratio in segment 2 was increased from 1·35 to 1·48. Interestingly, the adjusted force ratio in segment 3 was also 1·48, even though the EMG was not increased on average. This is due to the slow time course of muscle contraction, so that increased activation will produce a greater force over a period of time that spans more than one 150 ms segment. For the whole simulated stance phase the adjusted force ratio was 1·53 ± 0·13 which suggests that the phasic reflexes produced about 50 % more force on average than would occur under isometric conditions, rather than the initial estimate of 20 % above. Since the lengths are matched these effects arise from force-velocity and other properties of muscle. In other words, 0·53/1·53 or 35 % of the force over the whole stance phase is produced by reflexes due to length changes and 1/1·53 or 65 % by the CPG and other central and sensory inputs.
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The main finding of this paper is that the total EMG and force in the triceps muscles are substantially enhanced when the muscles move through the range of lengths found in normal walking, compared with when the muscles are held isometrically at an intermediate length. More specifically, muscle reflexes due to length changes during the stance phase produce over a third of the force of the ankle extensors. The balance comes from the spinal pattern generator, as well as other descending and sensory inputs that are still present in the isometric cycles. This result is rather surprising, based on the traditional view of velocity-dependent stretch and unloading reflexes, since the muscles are shortening through most of the extensor part of the step cycle (Fig. 1) and would be expected to produce less EMG and force. The largest increase in the EMG clearly does occur during the brief period of about 100 ms in Fig. 2, when the muscle is stretched by the weight of the body (the E2 phase in Phillipson's notation). As the muscle shortens in E3 the force generation is eventually reduced below isometric values (Fig. 2). However, this occurs when flexors such as IP are becoming active and the limb should be unloaded to initiate the swing phase.
These experiments are the first to provide an estimate of the reflex contribution from length changes during walking to force generation by direct force measurements. Several technical issues that could affect the calculated values have already been raised in the Methods, but there are physiological issues that need to be discussed here. First is the obvious caveat that the measurements are for a decerebrate animal and the values may be different in an intact cat. Decerebrate cats are well known to have strong stretch reflexes that produce the classic extensor rigidity. However, this applies more to the intercollicular decerebrate animal than the high decerebrate used here. If the reflexes were very different in the decerebrate walking animal from normal, they should produce abnormal movements, since they contribute substantially to the force generation. In fact, the kinematics are quite similar to normal (reviewed by Rossignol, 1996), which suggests that the reflexes are also similar.
Another issue is that the leg was denervated except for a few muscles of interest. Thus, excitatory inputs were lacking from other synergistic muscles as well as inhibitory inputs, including the classic disynaptic inhibition from antagonistic muscles such as the ankle flexors. Presynaptic inhibition from other muscle groups such as the quadriceps will also be lacking (e.g. Misiaszek et al. 1997). Since feedback of force from Golgi tendon organs in the walking animals reverses from inhibition to excitation (Pearson & Collins, 1993; Guertin et al. 1995; Hiebert & Pearson, 1999), the lack of force in the many denervated muscles of the leg could also have underestimated the effect of phasic reflexes. The net effect of these excitatory and inhibitory influences is uncertain.
The force will also depend on the time course (history) of previous forces and lengths experienced by the muscle. These factors affect the intrafusal as well as the extrafusal muscle fibres (Proske et al. 1992, 1993). Therefore, they affect the muscle spindles and the stretch reflex (Gregory et al. 1998). However, Fig. 3 shows that the effect of different preceding lengths on the force generated has virtually disappeared by 100 ms after the end of the flexor burst, so the forces measured in the stance phase (segments 2-4) were independent of the history.
Finally, the force generation during simulated walking was compared with isometric trials in which the muscle was held at a constant length equal to the average value during each segment of the extension phase. Thus, we were only assessing the importance of muscle reflexes due to the length changes during walking. As mentioned in Introduction, Hiebert & Pearson (1999) found even greater reductions (70 % in MG EMG) when the muscle was allowed to shorten in an unloaded fashion in decerebrate walking animals. To compare our results, we assessed the effects of holding the muscle isometrically at different lengths (Fig. 5). On average there was a 5 % reduction in MG EMG for each millimetre reduction in length. If the MG muscle shortened by 6-7 mm in Hiebert & Pearson's experiments and produced another 30-35 % reduction, our results would be in complete agreement with theirs. The decline in force is even steeper (16 % per mm), because of the effects of the length-tension curve. Again from Fig. 5, a 6 mm shortening can almost abolish force generation (the muscles remain nearly slack). Thus, although reflexes from length changes contribute about 35 % to the normal force generation during walking, the reflex effects will be even more profound from extra shortening to lengths where the muscles can generate little force. This may be appropriate if the foot goes into a hole and force production would delay lifting the foot out of the hole.
Our results for reflex contributions to EMG are also in good agreement with values obtained from earlier EMG estimates in the human (30-60 % in the soleus muscle during early stance phase, Yang et al. 1991; Sinkjær et al. 1996). In humans the calf muscles are being stretched during most of the stance phase, so more reflex force generation might be expected, but experiments similar to these have not been possible. The human experiments to date have used relatively standard stretches or releases and assumed that the values would scale linearly. They also had to assume that the EMG would translate into force or torque production, which may not be true (Kearney et al. 1999). Our results (Fig. 6) show that the increases in force and EMG are similar in segment 2, but that the force continues at a higher than expected level in segment 3, presumably because of the long time course of muscle contraction. Clearly, phasic reflexes in cats and humans can increase or decrease EMG and force generation dramatically when required during stretch or unloading (Sinkjær, 1997).
Functional consequences
The relative contributions of reflexes and CPG are adjusted according to conditions. Because of task-dependent reflex modulation (Edamura et al. 1991; Zehr & Stein, 1999) the importance of reflexes is different in running than in walking and in walking uphill or on different surfaces (sand, ice etc.), rather than on a flat, predictable surface. The full range of reflex modulation on force generation remains to be explored quantitatively, but reflexes appear to produce a substantial fraction of the force needed to support the body and accelerate it upward during stance in both cat and humans. An obvious functional advantage of having strong phasic reflexes is that the force can be quickly adjusted up or down as required in a given step.
When the ground is uneven and the foot is partially unloaded, much of the force can be removed after a short delay, if no reflexes are activated. Such a brisk unloading response has been observed by Gorrasini et al. (1994) in the 'foot in hole' experiment. In their experiment an intact cat walked on a treadmill and occasionally a trap-door opened and the foot went through a hole, rather than being supported by the treadmill surface. If the output were produced wholly by a CPG, such a rapid, unloading response would not be possible. Similarly, an added, unexpected load will lead to an enhanced response after a short latency. After nearly a century of controversy about whether motor patterns for walking are centrally or peripherally generated, we conclude that both are quantitatively important for force generation during the stance phase. The CPG produces the basic alternation and initial activation. To the extent required, reflexes can adjust the force up or down depending on the loading or unloading. If spinal reflexes are inadequate, longer latency reflexes and eventually voluntary corrections can be produced. All interact to generate graceful, coordinated locomotion over varied terrain.
|
REFERENCES |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
| Brown, T. G. (1911). The intrinsic factors in the act of progression in the mammal. Proceedings of the Royal Society B 84, 308-319. | |
| Edamura, M., Yang, J. F. & Stein, R. B. (1991). Factors that determine the magnitude and time course of human H-reflexes in locomotion. Journal of Neuroscience 11, 420-427 | [Abstract] |
| Gorrasini, M. A., Prochazka, A. K., Hiebert, G. W. & Gauthier, M. J. A. (1994). Corrective responses to loss of ground support during walking. I. Intact cats. Journal of Neurophysiology 71, 603-610 | [Medline] |
| Goslow, G. E., Reinking, R. M. & Stuart, D. G. (1973). The cat step cycle: hind limb joint angles and muscle lengths during unrestrained locomotion. Journal of Morphology 141, 1-42 | [Medline] |
| Gregory, J. E., Wise, A. K., Wood, S. A., Prochazka, A. & Proske, U. (1998). Muscle history, fusimotor activity and the human stretch reflex. The Journal of Physiology 513, 927-934 | [Abstract/Full Text] |
| Grillner, S. (1981). Control of locomotion in bipeds, tetrapods and fish. In Handbook of Physiology, section 1, The Nervous System, vol. II, Motor Control, part 2, ed. Brooks, V. B., pp. 1179-1236. American Physiological Society, Bethesda, MD, USA. | |
| Guertin, P., Angel, M. J., Perreault, M.-C. & McCrea, D. (1995). Ankle extensor group I afferents excite extensors through the hind limb during fictive locomotion in the cat. The Journal of Physiology 487, 197-207 | [Abstract] |
| Hiebert, G. W. & Pearson, K. G. (1999). Contribution of sensory feedback to the generation of extensor activity during walking in the decerebrate cat. Journal of Neurophysiology 81, 758-770 | [Abstract/Full Text] |
| Joyce, G. C. & Rack, P. M. H. (1969). Isotonic lengthening and shortening movements of cat soleus muscle. The Journal of Physiology 204, 475-491 | [Medline] |
| Joyce, G. C., Rack, P. M. H. & Westbury, D. R. (1969). The mechanical properties of cat soleus muscle during controlled lengthening and shortening movements. The Journal of Physiology 204, 461-474 | [Medline] |
| Kearney, R. E., Lortie, M. & Stein, R. B. (1999). Modulation of stretch reflexes during human walking: mechanisms and functional implications. Journal of Neurophysiology 81, 2893-2902 | [Abstract/Full Text] |
| McCrea, D. A., Shefchyk, S. J., Stephens, M. J. & Pearson, K. G. (1995). Disynaptic group I excitation of synergist ankle extensor motoneurones during fictive locomotion in the cat. The Journal of Physiology 487, 527-539 | [Abstract] |
| Misiaszek, J. E. & Pearson, K. G. (1997). Stretching quadriceps inhibits the soleus H reflex during locomotion in decerebrate cats. Journal of Neurophysiology 78, 2975-2984 | [Abstract/Full Text] |
| Pearson, K. G. & Collins, D. F. (1993). Reversal of the influence of group 1b afferents from plantaris on activity in medial gastrocnemius muscle during locomotor activity. Journal of Neurophysiology 70, 1009-1017 | [Medline] |
| Phillipson, M. (1905). L'autonomie et la centralization dans le systeme nerveux des animaux. Travails du Laboratoire de la Physiologie, Institut Solvey 7, 1-208. | |
| Proske, U., Morgan, D. L. & Gregory, J. E. (1992). Muscle history dependence of responses to stretch of primary and secondary endings of cat soleus muscle spindles. The Journal of Physiology 445, 81-95 | [Abstract] |
| Proske, U., Morgan, D. L. & Gregory, J. E. (1993). Thixotropy in skeletal muscle and in muscle spindles: a review. Progress in Neurobiology 41, 705-721 | [Medline] |
| Rack, P. M. H. & Westbury, D. R. (1969). The effects of length and stimulus rate on tension in the isometric cat soleus muscle. The Journal of Physiology 204, 443-460 | [Medline] |
| Rossignol, S. (1996). Neural control of stereotypic limb movements. In Handbook of Physiology, Section 12, Exercise: Regulation and Integration of Multiple Systems, ed. Rowell, L. B. & Shepherd, J. T., pp. 173-216. Oxford, New York. | |
| Sherrington, C. S. (1910). Flexion-reflex of the limb, crossed extension reflex and reflex stepping and standing. The Journal of Physiology 40, 28-121. | |
| Shik, M. L., Orlovskii, G. N. & Severin, F. V. (1966). Organization of locomotor synergism. Biophysics 11, 1011-1029 (translated from Biofizika 11, 879-886). | |
| Sinkjær, T. (1997). Muscle, reflex and central components in the control of the ankle joint in healthy and spastic man. Acta Neurologica Scandinavica 96, Suppl., 170. | |
| Sinkjær, T., Andersen, J. B. & Larsen, B. (1996). Soleus stretch reflex modulation during gait in humans. Journal of Neurophysiology 76, 1112-1120 | [Medline] |
| Yang, J. F., Stein, R. B. & James, K. B. (1991). Contribution of peripheral afferents to the activation of the soleus muscle during walking in humans. Experimental Brain Research 87, 679-687 | [Medline] |
| Zehr, E. P. & Stein, R. B. (1999). What functions do reflexes serve during human locomotion? Progress in Neurobiology 58, 185-205 | [Medline] |
This research was supported by grants from the Medical Research Council of Canada. J.E.M. received fellowship support from the Natural Sciences and Engineering Research Council. We thank Dr P. R. Murphy for helpful comments on the manuscript.
Corresponding author
R. B. Stein: Division of Neuroscience, University of Alberta, Edmonton, AB, Canada T6G 2S2.
Email: richard.stein{at}ualberta.ca
Author's present address
J. E. Misiaszek: Department of Occupational Therapy, University of Alberta, Edmonton, AB, Canada.
This article has been cited by other articles:
|
|
 |
|
  Y. Chen, X. Y. Chen, L. B. Jakeman, L. Chen, B. T. Stokes, and J. R. Wolpaw Operant Conditioning of H-Reflex Can Correct a Locomotor Abnormality after Spinal Cord Injury in Rats J. Neurosci., November 29, 2006; 26(48): 12537 - 12543. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. F Israel, D. D Campbell, J. H Kahn, and T G. Hornby Metabolic Costs and Muscle Activity Patterns During Robotic- and Therapist-Assisted Treadmill Walking in Individuals With Incomplete Spinal Cord Injury Physical Therapy, November 1, 2006; 86(11): 1466 - 1478. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Rossignol, R. Dubuc, and J.-P. Gossard Dynamic Sensorimotor Interactions in Locomotion Physiol Rev, January 1, 2006; 86(1): 89 - 154. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Stecina, J. Quevedo, and D. A. McCrea Parallel reflex pathways from flexor muscle afferents evoking resetting and flexion enhancement during fictive locomotion and scratch in the cat J. Physiol., November 15, 2005; 569(1): 275 - 290. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Chen, X. Y. Chen, L. B. Jakeman, G. Schalk, B. T. Stokes, and J. R. Wolpaw The Interaction of a New Motor Skill and an Old One: H-Reflex Conditioning and Locomotion in Rats J. Neurosci., July 20, 2005; 25(29): 6898 - 6906. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. M. Donelan and K. G. Pearson Contribution of Force Feedback to Ankle Extensor Activity in Decerebrate Walking Cats J Neurophysiol, October 1, 2004; 92(4): 2093 - 2104. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. Dietz and S. J. Harkema Locomotor activity in spinal cord-injured persons J Appl Physiol, May 1, 2004; 96(5): 1954 - 1960. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L.J.G. Bouyer and S. Rossignol Contribution of Cutaneous Inputs From the Hindpaw to the Control of Locomotion. I. Intact Cats J Neurophysiol, December 1, 2003; 90(6): 3625 - 3639. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. M.J.I. Huyghues-Despointes, T. C. Cope, and T. R. Nichols Intrinsic Properties and Reflex Compensation in Reinnervated Triceps Surae Muscles of the Cat: Effect of Activation Level J Neurophysiol, September 1, 2003; 90(3): 1537 - 1546. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M.-P. Cote, A. Menard, and J.-P. Gossard Spinal Cats on the Treadmill: Changes in Load Pathways J. Neurosci., April 1, 2003; 23(7): 2789 - 2796. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. K. Timoszyk, R. D. de Leon, N. London, R. R. Roy, V. R. Edgerton, and D. J. Reinkensmeyer The Rat Lumbosacral Spinal Cord Adapts to Robotic Loading Applied During Stance J Neurophysiol, December 1, 2002; 88(6): 3108 - 3117. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. E. Misiaszek and K. G. Pearson Adaptive Changes in Locomotor Activity Following Botulinum Toxin Injection in Ankle Extensor Muscles of Cats J Neurophysiol, January 1, 2002; 87(1): 229 - 239. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
