Peripheral control of the cat's step cycle. II. Entrainment of the central pattern generators for locomotion by sinusoidal hip movements during "fictive locomotion."

Acute low spinal and curarized cats injected with noradrenergic agonists i.v. can elicit an efferent burst pattern which can be recorded in muscle nerve filaments and can be referred to as "fictive locomotion". This study investigates the effect that feedback, arising from movements in the hip joint, can exert on the central network generating fictive locomotion. The central network is uncoupled from generating any active movements by curarization. The motor pattern could be entrained by applying sinusoidal hip movements, even when a very extensive denervation of the leg had been performed leaving only some of the muscles around the hip and the hip joint innervated. During flexion movements, efferents to different flexor muscles became active and during movements in the reverse direction (extension), efferents to extensors were active. With an increasing movement frequency the onsets of both flexor and extensor bursts were delayed in the movement cycle. The duration of the extensor bursts varied markedly with the movement cycle, whereas pure flexors changed less in burst duration. The frequency within which the efferent burst activity was entrained in a strict 1:1 relation to the movement varied between 5 to 70% above and below the resting burst period. In preparations with a narrow 1:1 range, a "relative coordination" was encountered outside this range. The flexor burst duration was in these cases dependent on where in the hip movement cycle the bursts appeared.     

Peripheral control of the cat's step cycle

Acute low spinal and curarized cats injected with noradrenergic agonists i.v. can elicit an efferent burst pattern which can be recorded in muscle nerve filaments and can be referred to as “fictive locomotion”. This study investigates the effect that feedback, arising from movements in the hip joint, can exert on the central network generating fictive locomotion. The central network is uncoupled from generating any active movements by curarization. The motor pattern could be entrained by applying sinusoidal hip movements, even when a very extensive denervation of the leg had been performed leaving only some of the muscles around the hip and the hip joint innervated. During flexion movements, efferents to different flexor muscles became active and during movements in the reverse direction (extension), efferents to extensors were active. With an increasing movement frequency the onsets of both flexor and extensor bursts were delayed in the movement cycle. The duration of the extensor bursts varied markedly with the movement cycle, whereas pure flexors changed less in burst duration. The frequency range within which the efferent burst activity was entrained in a strict 1:1 relation to the movement varied between 5 to 70% above and below the resting burst period. In preparations with a narrow 1:1 range, a “relative coordination” was encountered outside this range. The flexor burst duration was in these cases dependent on where in the hip movement cycle the bursts appeared.     

Interlimb coordination during fictive locomotion in the thalamic cat

Summary Efferent discharges in muscle nerves of the four limbs were recorded simultaneously during spontaneous fictive locomotion in thalamic cats with the goal of understanding how the central nervous system controls interlimb coordination during stepping. The onset of the bursts of activity in the nerve of a selected flexor muscle in each limb allowed the temporal and the phase relationships between the fictive step cycle of a pair of limbs to be determined. Our main results are the following: 1) the fictive step cycles of the two forelimbs are always strictly alternated whereas the phasing of the step cycles of either the two hindlimbs or pairs of homolateral or diagonal limbs is more variable; 2) the time interval between the onsets of the flexor bursts of one of the two pairs of diagonal limbs is independent of the step cycle duration; 3) distinct patterns of interlimb coordination exist during fictive locomotion; a small number of patterns of coordination involving all four limbs, which correspond to the walking and the trotting gaits in the intact cat, occur very frequently. The results demonstrate that the central nervous system deprived of phasic afferent inputs from the periphery has the capacity to generate most of the patterns of interlimb coordination which occur during real locomotion. They further support the view that the central pattern of interlimb coordination essentially results from diagonal interactions between a forelimb generator for locomotion and a hindlimb one.Abbreviations CD step cycle duration - LF left forelimb - LH left hindlimb - m slope of correlation curve - N number of step cycles - r correlation coefficient - RF right forelimb - RH right hindlimb - Ti time interval     

Plasticity of reflexes from the foot during locomotion after denervating ankle extensors in intact cats

Although sensory feedback is important in regulating the timing and magnitude of muscle activity during locomotion few studies have evaluated how it changes after peripheral nerve lesions. To assess this, reflexes evoked by stimulating a nerve before and after denervating other nerves can be quantified to determine changes. The aim of this study was to investigate consequences of denervating ankle extensor muscles, the lateral gastrocnemius, and soleus (LGS) on reflexes from the plantar foot surface evoked by stimulating the tibialis (Tib) nerve. Three cats (n = 3) were trained to walk on a treadmill and chronically implanted with electrodes in 14 hindlimb muscles bilaterally to record EMG activity. A stimulating cuff electrode was placed around the left Tib nerve (Tib) nerve at the ankle to evoke reflexes. Several control values of EMGs, limb kinematics, and Tib nerve reflexes were obtained during locomotion for at least 3 wk before the left LGS nerve was cut. We found that the locomotor EMG bursts of several muscles was altered, with a large increase in amplitude in the early days postneurectomy followed by a gradual decrease toward intact values later on. There were changes in the stimulated locomotor EMG bursts (Tib nerve reflexes) of ipsilateral flexors and extensors and of contralateral ankle extensors, which dissociated from changes in baseline locomotor EMG (e.g., nonstimulated bursts during reflex trials). The functional significance of these changes in muscle activity and reflex pathways on the recovery of locomotion after denervating ankle extensors is discussed.     

Rhythmic phrenic,intercostal and sympathetic activity in relation to limb and trunk motor activity in spinal cats

During L-DOPA-induced fictive spinal locomotion rhythmic activities in nerves to internal intercostal and external oblique abdominal muscles and in phrenic and sympathetic nerves were observed which were always coordinated with locomotor activity in forelimb and hindlimb muscle nerves. A periodicity with longer lasting tonic phases could be induced by cutaneous nerve stimulation or asphyxia. This activity was observed in limb motor nerves as well as in respiratory motor and sympathetic nerves. A slow independent activity of the phrenic and intercostal nerves or the sympathetic nerves, which could be related to a normal respiratory rhythm or independent sympathetic rhythms was not observed. The findings indicate that during fictive spinal locomotion the activity of spinal rhythm generators for locomotion also projects onto respiratory and sympathetic spinal neurones.     

Electrically evoked fictive swimming in the low-spinal immobilized turtle

Fictive swimming was elicited in low-spinal immobilized turtles by electrically stimulating the contralateral dorsolateral funiculus (cDLF) at the level of the third postcervical segment (D(3)). Fictive hindlimb motor output was recorded as electroneurograms (ENGs) from up to five peripheral nerves on the right side, including three knee extensors (KE; iliotibialis [IT]-KE, ambiens [AM]-KE, and femorotibialis [FT]-KE), a hip flexor (HF), and a hip extensor (HE). Quantitative analyses of burst amplitude, duty cycle and phase were used to demonstrate the close similarity of these cDLF-evoked fictive motor patterns with previous myographic recordings obtained from the corresponding hindlimb muscles during actual swimming. Fictive rostral scratching was elicited in the same animals by cutaneous stimulation of the shell bridge, anterior to the hindlimb. Fictive swim and rostral scratch motor patterns displayed similar phasing in hip and knee motor pools but differed in the relative amplitudes and durations of ENG bursts. Both motor patterns exhibited alternating HF and HE discharge, with monoarticular knee extensor (FT-KE) discharge during the late HF phase. The two motor patterns differed principally in the relative amplitudes and durations of HF and HE bursts. Swim cycles were dominated by large-amplitude, long-duration HE bursts, whereas rostral scratch cycles were dominated by large-amplitude, long-duration HF discharge. Small but significant differences were also observed during the two behaviors in the onset phase of biarticular knee extensor bursts (IT-KE and AM-KE) within each hip cycle. Finally, interactions between swim and scratch motor networks were investigated. Brief activation of the rostral scratch during an ongoing fictive swim episode could insert one or more scratch cycles into the swim motor pattern and permanently reset the burst rhythm. Similarly, brief swim stimulation could interrupt and reset an ongoing fictive rostral scratch. This shows that there are strong central interactions between swim and scratch neural networks and suggests that they may share key neural elements.     

The effect of dorsal root transection on the efferent motor pattern in the cat's hindlimb during locomotion

Mesencephalic cats can walk on a treadmill if the midbrain locomotor region is stimulated. The motor pattern of different hindlimb muscles is similar to that of th intact cat. The present experiments in the mesencephalic preparation test if the complex motor pattern in one hindlimb is causally dependent on the afferent signals arising in the same limb during walking. The electromyographical activity and the movement pattern during locomotion were compared before and after transecting all dorsal root fibres originating from one hindlimb. Flexor and extensor muscles at different joints may retain their general pattern after the dorsal root transection. This applies also to muscles such as the knee flexors, which have a short and early flexor burst and a second burst during the extension phase, and the short toe dorsiflexor , which has an early burst in the transition between flexor and extensor activity. After the dorsal root transection the pattern of activity may become more variable and it can even break down altogether. The present results demonstrate that the central nervous system devoid of phasic afferent inflow from one hindlimb can produce a complex motor output to this limb rather than a motor pattern degraded to a simple alternation between flexors and extensors.     

An ascending spinal pathway transmitting a central rhythmic pattern to the magnocellular red nucleus in the cat

The activity of cells in the magnocellular red nucleus (RNm) was recorded extra and intracellularly in the curarized thalamic cat performing fictive locomotion. The locomoter episodes were detected from the rhythmic activity recorded in the motor nerves of the contralateral hindlimb. It was confirmed that, during fictive locomotion, a large proportion of the rubrospinal cells (56% in our sample) exhibit a rhythmic pattern of activity which is synchronized with the efferent spinal motor nerve activity. On the basis of the intracellular recordings it was established that phases of intense synaptic activity with mixed excitatory postsynaptic potentials (EPSPs) and inhibitory postsynaptic potentials (IPSPs) are involved in this rhythmicity. After eliminating the cerebellar input to the RNm, it was observed that the cells still received intense excitatory and inhibitory inputs, resulting in a continuous modulation of their membrane potential, due to the occurrence of EPSPs and IPSPs. During fictive locomotor like activity and after elimination of the cerebellar afferents to the RNm, it was observed that the spontaneous PSPs in RNm cells (in the case of 45% of the cells) were organized in repetitive subthreshold bursts occurring in phase relationships with the activity recorded in the motor nerves. Some extracellularly recorded cells (12%) showed a rhythmic firing pattern. It is generally recognized that, in the thalamic cat preparation, the locomotor pattern observed in efferent nerves originates from the central pattern generator (CPG) of the spinal cord. It therefore seems likely that the rhythmicity observed here in the RNm may originate from the spinal CPG and be transmitted through the spino rubral pathway ascending in the ventral part of the cord. It is concluded that the spino rubral pathway may transmit both somatosensory information and corollary discharges relating to the activity of the spinal CPG for locomotion.     

Stumbling corrective reaction during fictive locomotion in the cat

An obstacle contacting the dorsal surface of a cat's hind foot during the swing phase of locomotion evokes a reflex (the stumbling corrective reaction) that lifts the foot and extends the ankle to avoid falling. We show that the same sequence of ipsilateral hindlimb motoneuron activity can be evoked in decerebrate cats during fictive locomotion. As recorded in the peripheral nerves, twice threshold intensity stimulation of the cutaneous superficial peroneal (SP) nerve during the flexion phase produced a very brief excitation of ankle flexors (e.g., tibialis anterior and peroneus longus) that was followed by an inhibition for the duration of the stimulus train (10-25 shocks, 200 Hz). Extensor digitorum longus was always, and hip flexor (sartorius) activity was sometimes, inhibited during SP stimulation. At the same time, knee flexor and the normally quiescent ankle extensor motoneurons were recruited (mean latencies 4 and 16 ms) with SP stimulation during fictive stumbling correction. After the stimulus train, ankle extensor activity fell silent, and there was an excitation of hip, knee, and ankle flexors. The ongoing flexion phase was often prolonged. Hip extensors were also recruited in some fictive stumbling trials. Only the SP nerve was effective in evoking stumbling correction. Delivered during extension, SP stimulus trains increased ongoing extensor motoneuron activity as well as increasing ipsilateral hip, knee, and ankle hindlimb flexor activity in the subsequent step cycle. The fictive stumbling corrective reflex seems functionally similar to that evoked in intact, awake animals and involves a fixed pattern of short-latency reflexes as well as actions evoked through the lumbar circuitry responsible for the generation of rhythmic alternating locomotion.     

Short-latency crossed inhibitory responses in extensor muscles during locomotion in the cat

During locomotion, contacting an obstacle generates a coordinated response involving flexion of the stimulated leg and activation of extensors contralaterally to ensure adequate support and forward progression. Activation of motoneurons innervating contralateral muscles (i.e., crossed extensor reflex) has always been described as an excitation, but the present paper shows that excitatory responses during locomotion are almost always preceded by a short period of inhibition. Data from seven cats chronically implanted with bipolar electrodes to record electromyography (EMG) of several hindlimb muscles bilaterally were used. A stimulating cuff electrode placed around the left tibial and left superficial peroneal nerves at the level of the ankle in five and two cats, respectively, evoked cutaneous reflexes during locomotion. During locomotion, short-latency ( approximately 13 ms) inhibitory responses were frequently observed in extensors of the right leg (i.e., contralateral to the stimulation), such as gluteus medius and triceps surae muscles, which were followed by excitatory responses ( approximately 25 ms). Burst durations of the left sartorius (Srt), a hip flexor, and ankle extensors of the right leg increased concomitantly in the mid- to late-flexion phases of locomotion with nerve stimulation. Moreover, the onset and offset of Srt and ankle extensor bursts bilaterally were altered in specific phases of the step cycle. Short-latency crossed inhibition in ankle extensors appears to be an integral component of cutaneous reflex pathways in intact cats during locomotion, which could be important in synchronizing EMG bursts in muscles of both legs.     

Muscle activity during forelimb stepping in decerebrate cats.

In decerebrate cats with the lower thoracic cord transected, electromyographic activities were analyzed in up to 41 forelimb muscles, almost all muscles involved in forelimb stepping (intrinsic hand muscles were not included). From the active period in the step cycle, muscles were classified into three groups: extensors, of which activity is like that of elbow extensors; flexors, activity like that of elbow flexors; others, including dorsiflexors of the wrist, pronators, and supinator. The results were well consistent with those from conscious animals as well as efferent pattern of fictive locomotion in elbow and distal muscles. Nevertheless, in some proximal muscles discrepancies were noted, suggesting their changeability depending on environmental conditions. Recording from almost all muscles allowed to estimate rhythmic change of the overall output of the forelimb central pattern generator.     

Intracellular analysis of reflex pathways underlying the stumbling corrective reaction during fictive locomotion in the cat

In cat and humans, contact between an obstacle and the dorsum of the foot evokes the stumbling corrective reaction (reflex) that lifts the foot to avoid falling. This reflex can also be evoked by short trains of stimuli to the cutaneous superficial peroneal (SP) nerve in decerebrate cats during the flexion phase of fictive locomotion. Here we examine intracellular events in hindlimb motoneurons accompanying stumbling correction. SP stimulation delivered during the flexion phase excites knee flexor motoneurons at short latency [minimum excitatory postsynaptic potential (EPSP) latency 1.8 ms; mean 2.7 ms]. Although a similar short latency excitation occurs in ankle extensors (mean latency, 2.8 ms), recruitment is delayed until successive shocks in the stimulus train overcome the locomotor-related hyperpolarization of ankle extensors. In ankle flexor motoneurons, SP stimulation evokes an inhibition (mean latency, 2.7 ms) that briefly reduces or stops their firing during the flexion phase. There is a phase-dependent modulation of SP-evoked EPSP amplitude as well as latency during locomotion. However, the more obvious change in SP reflex pathways with the onset of fictive locomotion is the reduced inhibition of ankle extensor motoneurons and the increased inhibition of ankle flexors. These results show that the characteristic pattern of hindlimb motoneuron activation during SP nerve-evoked stumbling correction results from 1) di- and trisynaptic excitation of knee flexor and ankle extensor motoneurons; 2) increased inhibitory postsynaptic potentials in ankle flexors and a suppression of inhibition in extensors, 3) sculpting of the short-latency SP postsynaptic effects by motoneuron membrane potential, and 4) longer latency excitatory effects that are likely evoked by lumbar interneurons involved in the generation of fictive locomotion.     

Muscle Spindle Control during Locomotor Movements Generated by the Deafferented Spinal Cord

In acute spinalized (Th 12) cats, treated with DOPA, curarized or with all lumbosacral dorsal roots transected, stepping patterns were elicited in the hindlimbs by continuous electrical stimulation of the dorsal roots. Single primary endings in dorsal root filaments or single γ and α efferents in muscle nerve filaments from flexors and extensors were recorded together with the efferent nerve activity to various hindlimb muscles. The experiments showed a clear coactivation of the intra- and extrafusal system to the same muscle. The α efferent discharges typically started with a very short first interspace interval. It is concluded that there exists a central spinal α-γ-Iinkage for locomotion in the hindlimbs.     

Muscle spindle control during locomotor movements generated by the deafferented spinal cord.

In acute spinalized (Th 12) cats, treated with DOPA, curarized or with all lumbosacral dorsal roots transected, stepping patterns were elicited in the hindlimbs by continuous electrical stimulation of the dorsal roots. Single primary endings in dorsal root filaments or single gamma and alpha efferents in muscle nerve filaments from flexors and extensors were recorded together with the efferent nerve activity to various hindlimb muscles. The experiments showed a clear coactivation of the intra- and extrafusal system to the same muscle. The alpha efferent discharges typically started with a very short first interspike interval. It is concluded that there exists a central spinal alpha-gamma-linkage for locomotion in the hindlimbs.     

Motoneuronal and muscle synergies involved in cat hindlimb control during fictive and real locomotion: a comparison study

We compared the activity profiles and synergies of spinal motoneurons recorded during fictive locomotion evoked in immobilized decerebrate cat preparations by midbrain stimulation to the activity profiles and synergies of the corresponding hindlimb muscles obtained during forward level walking in cats. The fictive locomotion data were collected in the Spinal Cord Research Centre, University of Manitoba, and provided by Dr. David McCrea; the real locomotion data were obtained in the laboratories of M. A. Lemay and B. I. Prilutsky. Scatterplot representation and minimum spanning tree clustering algorithm were used to identify the possible motoneuronal and muscle synergies operating during both fictive and real locomotion. We found a close similarity between the activity profiles and synergies of motoneurons innervating one-joint muscles during fictive locomotion and the profiles and synergies of the corresponding muscles during real locomotion. However, the activity patterns of proximal nerves controlling two-joint muscles, such as posterior biceps and semitendinosus (PBSt) and rectus femoris (RF), were not uniform in fictive locomotion preparations and differed from the activity profiles of the corresponding two-joint muscles recorded during forward level walking. Moreover, the activity profiles of these nerves and the corresponding muscles were unique and could not be included in the synergies identified in fictive and real locomotion. We suggest that afferent feedback is involved in the regulation of locomotion via motoneuronal synergies controlled by the spinal central pattern generator (CPG) but may also directly affect the activity of motoneuronal pools serving two-joint muscles (e.g., PBSt and RF). These findings provide important insights into the organization of the spinal CPG in mammals, the motoneuronal and muscle synergies engaged during locomotion, and their afferent control.     

The distal hindlimb musculature of the cat

In order to better understand the organization of the locomotor control system, we examined the temporal patterns of distal hindlimb muscle responses to brief electrical stimulation of cutaneous nerves during walking on a treadmill. Electromyographic recordings were made from twelve muscles; stimuli were applied individually to three nerves at random times throughout the step cycle. A new graphical technique was developed to assist detailed examination of the time course and gating of complex reflex patterns. The short latency reflexes were of two primary types: inhibition of extensors and excitation of flexors; these responses were only evident during locomotor phases in which the respective motoneuron pools were active. Longer-latency response components were gated in a similar but not identical manner, suggesting some independence from the basic locomotory influence on the motoneuronal pool. The phase-dependent gating of reflexes appeared to be consistent with a functional role for reflex responses during locomotion. The reflex responses of muscles with complex anatomical actions were often correspondingly complex.     

Deletions of rhythmic motoneuron activity during fictive locomotion and scratch provide clues to the organization of the mammalian central pattern generator

We examined the features of spontaneous deletions of bursts of motoneuron activity that can occur within otherwise rhythmic alternating flexor and extensor activity during fictive locomotion and scratch in adult decerebrate cats. Deletions of activity were observed both in hindlimb flexor and extensor motoneuron pools during brain stem-stimulation-evoked fictive locomotion but only in extensors during fictive scratch. Paired intracellular motoneuron recordings showed that deletions reduced the depolarization of homonymous motoneurons in qualitatively similar ways. Differences occurred in the extent to which activity in synergist motoneuron pools operating at other joints within the limb was reduced during deletions. The timing of the rhythmic activity that followed a deletion was often at an integer multiple of the preexisting locomotor or scratch cycle period. This maintenance of cycle period was also seen during deletions in which there was a complete failure of motoneuron depolarization. The activity of antagonist motoneurons was usually sustained during deletions with some rhythmic modulation at intervals of the preexisting cycle period. We discuss an organization of the central pattern generator for locomotion and scratch that functions as a single rhythm generator with separate and multiple pattern formation modules for controlling the hyper- and depolarization of subsets of motoneurons within the limb.     

Fictive motor activity in rat after 14 days of hindlimb unloading

The goal of the present study was to examine the effects of chronic hindlimb unloading on fictive motor patterns which can be developed in hindlimb nerves of adult rats. The animals were divided into two groups. The first group was submitted to hindlimb unloading for 2 weeks by tail suspension. The second group served as controls. After this initial phase, the animals of both groups were acutely decorticated, paralysed and electroneurographic efferent activity was recorded from hindlimb muscle nerves under conditions of "fictive locomotion" in order to evaluate variations in central locomotor command. Fictive rhythmic motor episodes were either spontaneous or evoked by electrical stimulation of the mesencephalic locomotor region. Only the second ones were recognised as locomotor-like activities. The motor pattern was not fundamentally affected by unloading except that, after the unloading period, extensor muscle nerves were significantly more frequently activated and their burst durations were increased compared to activity in control animals, despite the fact that the phasic sensory afferent inputs were suppressed. This suggests that unloading induces plastic modifications of the central networks of neurons implicated in the locomotor command. The origin of this extensor hyperactivity is discussed. It is proposed that it could be the consequence of either changes in motoneuronal properties or of an increase in afferent input to motoneurones. Electronic Publication     

Proprioceptive input resets central locomotor rhythm in the spinal cat

Summary The reflex regulation of stepping is an important factor in adapting the step cycle to changes in the environment. The present experiments have examined the influence of muscle proprioceptors on centrally generated rhythmic locomotor activity in decerebrate unanesthetized cats with a spinal transection at Th12. Fictive locomotion, recorded as alternating activity in hindlimb flexor and extensor nerves, was induced by administration of nialamide (a monoamine oxidase inhibitor) and L-DOPA. Brief electrical stimulation of group I afferents from knee and ankle extensors were effective in resetting fictive locomotion in a coordinated fashion. An extensor group I volley delivered during a flexor burst would abruptly terminate the flexor activity and initiate an extensor burst. The same stimulus given during an extensor burst prolonged the extensor activity while delaying the appearance of the following flexor burst. Intracellular recordings from motoneurones revealed that these actions were mediated at premotoneuronal levels resulting from a distribution of inhibition to centres generating flexor bursts and excitation of centres generating extensor bursts. These results indicate that extensor group I afferents have access to central rhythm generators and suggest that this may be of importance in the reflex regulation of stepping. Experiments utilizing natural stimulation of muscle receptors demonstrate that the group I input to the rhythm generators arises mainly from Golgi tendon organ Ib afferents. Thus an increased load of limb extensors during the stance phase would enhance and prolong extensor activity while simultaneously delaying the transition to the swing phase of the step cycle.     

Synaptic connections from large muscle afferents to motoneurones in human lower limbs

OBJECTIVES: Our study was undertaken to investigate reciprocal inhibition in humans both from ankle flexors to extensors and from ankle extensors to flexors. METHODES: Changes in the firing probability of single motor units in response to electrical stimulation of muscle nerves (the peristimulus time histogram technique) were used to derive the reciprocal projections of muscle spindle Ia afferents to the motoneurones of ankle muscles. Discharges of units in ankle flexors (the tibialis anterior muscle [TA]) and extensors (soleus [SOL] and medial gastrocnemius [MG] muscles) were investigated respectively after stimulation of the posterior tibial (PTN) and common peroneal (CPN) nerves (predominantly on the deep branch). In eight normal subjects aged 24 to 40 years, one motor unit per each muscle was studied. RESULTS: CPN stimulation produced reciprocal Ia inhibition in the SOL of 5 of 7 of them and in the MG of 3 of 5, whereas PTN stimulation produced reciprocal Ia inhibition in the TA of only 2 of 6 subjects. CONCLUSIONS: These findings suggest that at low level contraction reciprocal Ia inhibition from ankle flexors to extensors may be stronger than that from ankle extensors to flexors.