Somatosensory graviception inhibits soleus H-reflex gain in humans during walking

To investigate the effects of gravity-related somatosensory information on spinal human reflexes, the soleus H-reflex was recorded in ten healthy subjects walking on a treadmill at 2.0 km/h on land and in water. The modulation pattern of the soleus H-reflex was determined in ten different phases of the step cycle. While the subjects were walking in water, the background electromyographic activity (BGA) of the soleus was lower than that on land; on the other hand, the soleus H-reflex amplitude while the subjects were walking in water showed no significant differences throughout the step cycle compared with that while the subjects were on land; the phase-dependent soleus H-reflex modulation pattern was well preserved while walking in water. There was a linear relationship between the BGA and the H-reflex amplitude in each condition; however, the soleus H-reflex gain while walking in water was significantly higher than that on land. These findings suggest that the somatosensory graviception can markedly reduce the spinal reflex excitability. Our findings are discussed in relation to human gait; therefore, further studies are needed to clarify the effect of somatosensory graviception on human neural mechanisms.     

On the soleus H-reflex modulation pattern during walking

In a recent paper it was claimed that in the majority (9/15) of subjects studied the soleus H-reflex increases progressively during the swing phase of walking. This pattern was at odds with our numerous observations made since 1986, as was the very large proportion of subjects reported to exhibit this pattern. We therefore reinvestigated the issue in an extensive series of experiments and detailed subsequent analysis on 21 subjects. In most subjects (13/21) the soleus H-reflex was completely inhibited during most or all of the swing phase (group A). In 8/21 subjects (group B) there was a small H-reflex mean 16% (SD=10.6%) of the value in quiet standing present during most or all of swing, but there was no systematic modulation pattern; the reflex amplitude fluctuated in a seemingly random manner. The difference between the two somewhat arbitrary groups could not be explained on the basis of greater electromyographic activity in the tibialis anterior (TA) during the swing phase or at the time of heel contact. However, by normalizing the mean level of TA activity to the peak level, the ratio was significantly greater for the group A subjects. This highlights the importance of reciprocal inhibition in accounting for the suppression of the soleus H-reflex in swing. In the discussion we emphasize that the presence of a small H-reflex during swing in the group B subjects is unlikely to have any functional role. What is of functional importance is the strong inhibition of the H-reflex during swing which reflects the ensemble of neural mechanisms at play to prevent the unwanted activation of the powerful ankle extensor muscles.Christian Ethier, Marie-Andrée Imbeault, and Visal Ung contributed equally to the study     

Excitability of the soleus H-reflex arc during walking and stepping in man

Summary In eight normal subjects, the excitability of the soleus (Sol) H-reflex was tested in parallel with Sol length changes, EMGs of leg and thigh muscles and ground contact phases, during three different pacing movements: bipedal treadmill walking, single limb treadmill walking, and single-limb stepping on one spot. A computerized procedure was used which compensated for changes in stimulus effectiveness that occurred during free motion. In the three paradigms examined, significant excitability modulations were observed with respect to a control level determined in standing weight-bearing position. During bipedal treadmill walking, excitability was decreased in the early stance, maximally enhanced in the second half of the stance, and again decreased during the end-stance and the whole swing phase, with a minimum value around the toe off period. The main modulation pattern was retained during single-limb treadmill walking. During single-limb stepping on one spot, the stance-phase increase in excitability and the swing phase depression were still present. However, in the second half of the swing phase, reflex responsiveness returned to reference level, which was maintained during the subsequent contact period. Moreover, a decrease in reflex excitability was detected around the mid-stance. The time course of the described modulations was only partly correlated with the EMG and length changes of the Sol muscle. Furthermore, in the three movements tested, during the early stance phase, the excitability of the H-reflex arc did not correspond to the one expected on the basis of the available H-reflex studies performed under static conditions. It is suggested that, at least in certain stride phases (e.g. around the early contact period), an active regulation affects the transmission in the Sol myotatic arc during the pacing movements investigated.     

Neural mechanisms that contribute to cyclical modulation of the soleus H-reflex in walking in humans

The amplitude of the Hoffmann reflex (H-reflex) of the human soleus muscle is modulated in a cyclical way during walking. This paper addresses two questions associated with the neural mechanisms that might generate this modulation: (1) Does the amplitude of the H-reflex simply rise and fall as a function of the background excitability of the soleus motoneuron pool? (2) Is the modulation of the H-reflex dependent on events associated with activation of the antagonist muscle? The amplitude of the soleus H-reflex was compared under three conditions: natural walking, walking without activating the tibialis anterior muscle, and walking with activation of the soleus muscle in the swing phase. Human subjects were able to perform these three tasks with minimal training. The results indicated that the soleus H-reflex remained very depressed in the swing phase of walking, even when a voluntary contraction of the soleus muscle was superimposed during this time. Moreover, the presence of tibialis anterior activity had a very minor effect on the amplitude of the soleus H-reflex during walking. It is concluded that modulation of the soleus H-reflex is not simply a reflection of the background excitability of the motoneuron pool, and the modulation is not dependent on activation of the antagonist muscle. Other more powerful mechanisms are acting to modulate the reflex, most likely presynaptic inhibition of the primary afferents.     

Comparison of amplitude of human soleus H-reflex during sitting and standing.

The modulation of the H-reflex in the human soleus muscle under conditions of different length or of background EMG activity was compared in 7 healthy subjects under three conditions: sitting, standing with support, and standing without support. The amplitude of the H-reflex increased when the muscle was shortened in both the sitting and standing conditions. The degree of increase in H-reflex was smaller during standing than sitting for the same change in muscle length. The H-reflex was augmented according to the increase of the background EMG. The "reflex gain", the ratio of the increase in amplitude of the H-reflex to soleus muscle EMG activity, decreased on sitting, standing with support and standing without support, ranked in that order. From these observations, it is concluded that the H-reflex is modulated by both muscle length and the degree of postural stability. The modulation of the reflex could be interpreted in terms of gain compensation and would serve to stabilize posture. A decrease in reflex gain may be appropriate in stabilizing the spinal reflex feedback loop during standing, especially without support.     

Phase-dependent modulation of the soleus H-reflex during rhythmical arm swing in humans

The purpose of this study is to investigate modulation of the soleus H-reflex during rhythmic arm swing in humans. Significant depression of the soleus H-reflex was observed when subjects swung their ipsilateral arms or both arms reciprocally during testing. The degree of soleus H-reflex depression appeared directly proportional to the speed of the arm swing. This depression was observed in the conditioning-testing intervals of 400, 500, and 600 msec during the ipsilateral backward arm swing and at the onset of the ipsilateral arm forward swing. This phase of depression partially overlapped the phase of depression of the soleus H-reflex during walking. However, the pattern of modulation during arm swing was not exactly the same as that during walking. Therefore, we concluded that the ipsilateral arm swing may partially affect the depression of the soleus H-reflex during the arm swing phase of walking but is not responsible for depression of the soleus H-reflex throughout the entire walking cycle.     

Differential regulation of cutaneous and H-reflexes during leg cycling in humans

Reflexes undergo modulation according to task and timing during standing, walking, running, and leg cycling in humans. Both cutaneous and Hoffman (H-) reflexes are modulated by movement and task. However, recent evidence suggests that the modulation pattern for cutaneous and H-reflexes may be different. We sought to clarify this issue by reducing the effect of movement phase and altering the level of background muscle activation (low and high) in static and dynamic (leg cycling) conditions. Electromyography was recorded from the ankle extensors soleus and medial gastrocnemius (MG) and the knee extensor vastus lateralis (VL). Reflexes were evoked during the downstroke of stationary leg cycling. Cutaneous reflexes were evoked with trains of 5 x 1.0 ms pulses at 300 Hz delivered to the distal tibial nerve, whereas H-reflexes were evoked in soleus by stimulation with single 1.0-ms pulses. There were two main observations in this study: 1) middle latency cutaneous reflexes were facilitatory during static contraction but were dramatically attenuated or reversed to suppressive responses during cycling (task-dependent modulation); 2) soleus H-reflexes were larger in the high muscle activation condition but were unaffected by task (no task-dependent modulation). Thus opposite results were obtained in the two reflex pathways. It is concluded that cutaneous and H-reflexes are modulated by different mechanisms during active locomotor-like movements.     

Acute effects of whole body vibration during passive standing on soleus H-reflex in subjects with and without spinal cord injury

Whole-body vibration (WBV) is being used to enhance neuromuscular performance including muscle strength, power, and endurance in many settings among diverse patient groups including elite athletes. However, the mechanisms underlying the observed neuromuscular effects of WBV have not been established. The extent to which WBV will produce similar neuromuscular effects among patients with neurological impairments unable to voluntarily contract their lower extremity muscles is unknown. We hypothesized that modulation of spinal motorneuronal excitability during WBV may be achieved without voluntary contraction. The purpose of our study was to describe and compare the acute effects of WBV during passive standing in a standing frame on the soleus H-reflex among men with and without spinal cord injury (SCI). In spinal cord intact participants, WBV caused significant inhibition of the H-reflex as early as 6 s after vibration onset (9.0 ± 3.9%) (p < 0.001). The magnitude of the H-reflex gradually recovered after WBV, but remained significantly below initial values until 36 s post-WBV (57.5 ± 22.0%) (p = 0.01). Among participants with SCI, H-reflex inhibition was less pronounced with onset 24 s following WBV (54.2 ± 18.7%) (p = 0.03). The magnitude of the H-reflex fully recovered after 60 s of WBV exposure. These results concur with prior reports of inhibitory effects of local vibration application on the H-reflex. Our results suggest that acute modulation of spinal motoneuronal excitability during WBV can be achieved in the absence of voluntary leg muscle contractions. Nonetheless, WBV has implications for rehabilitation service delivery through modulation of spinal motoneuronal excitability in individuals with SCI.     

Excitability of the soleus H reflex during graded walking in humans

The excitability of the soleus Hoffmann (H) reflex was measured in five healthy male subjects during graded treadmill walking. Uphill and downhill walking at an 8% grade as well as level walking were used to vary the demands for lengthening and shortening contractions of the soleus muscle. These changes were assumed to cause differences in control of the afferent input in the spinal cord and the voluntary output to the soleus muscle. The H reflex was strongly modulated in all three walking conditions, high during the stance phase and low or absent during the swing phase. The shape of the modulations was, however, different. At uphill walking the reflex increased gradually during the whole stance phase and seemed to follow the soleus electromyogram (EMG) pattern closely. In the downhill condition the reflex excitability increased rapidly at heel strike like the soleus EMG and co-contraction of the anterior tibial muscle was observed. At level walking a fast rise in reflex excitability was seen just after heel strike with low or absent soleus EMG. Mean soleus EMG was lower during downhill than during uphill or level walking, but the mean H reflex amplitude was similar in all three conditions. However, when the H reflex was related directly to the EMG activity by linear regression the reflex gain was lower during uphill walking than in the two other conditions. Furthermore, the ratio between H reflex and EMG amplitude was high during the first half of the stance phase at level walking indicating an elevated reflex excitability independent of the voluntary motor output. It is therefore concluded that the modulation of reflexes during walking cannot be interpreted in terms of the idea of automatic gain compensation. The reflexes must be controlled specifically and independently during the different phases of the motor output to meet the mechanical requirements of the movement task. Most explicitly this was seen during downhill walking, where an elevated reflex excitability together with co-contraction at the ankle joint seem to provide increased joint stiffness and security, when the kinetic energy of the body has to be brought under control at heel strike.     

Effect of sensory inputs on the soleus H-reflex amplitude during robotic passive stepping in humans

We investigated the modulation of the soleus (Sol) Hoffmann (H-) reflex excitability by peripheral sensory inputs during passive stepping using a robotic-driven gait orthosis in healthy subjects and spinal cord-injured patients. The Sol H-reflex was evoked at standing and at six phases during passive stepping in 40 and 100% body weight unloaded conditions. The Sol H-reflex excitability was significantly inhibited during passive stepping when compared with standing posture at each unloaded condition. During passive stepping, the H-reflex amplitude was significantly smaller in the early- and mid-swing phases than in the stance phase, which was similar to the modulation pattern previously reported for normal walking. No significant differences were observed in the H-reflex amplitude between the two unloaded conditions during passive stepping. The reflex depression observed at the early part of the swing phase during passive stepping might be attributed to the sensory inputs elicited by flexion of the hip and knee joints. The present study provides evidence that peripheral sensory inputs have a significant role in phase-dependent modulation of the Sol H-reflex during walking, and that the Sol H-reflex excitability might be less affected by load-related afferents during walking.     

Intrasession and intersession reliability of the soleus H-reflex in supine and standing positions

The Hoffmann reflex (H-reflex) is a measure of motoneuron pool excitability, which is valuable in determining muscle inhibition caused by joint damage (arthrogenic muscle inhibition). In order to detect changes in H-reflex due to injury, the reliability of such a measurement must be established. The purpose of this study was to establish the intrasession and intersession reliability of soleus H-reflex in a supine and standing position. Thirteen healthy volunteers (age 10 +/- 2.63 yr, height 171.35 +/- 10.19 cm, mass 69.62 +/- 13.03 Kg) with no lower extremity orthopedic or neurological disorders within the past year participated in this study. To determine the intrasession and intersession reliability of this measure in a supine resting position and a one-leg standing position, EMG data were collected from the soleus while the tibial nerve was stimulated in the popliteal space. A high voltage (120-200 V), short duration (1.0 msec) stimulus was automatically triggered, eliciting a reflex twitch detected by surface EMG. Several of these measurements were performed with 20 second rest intervals to find the maximum H-reflex. The maximum H-reflex was located by adjusting the intensity of the stimulus. Once a maximum H-reflex was found, 12 measurements were taken in that position with 20 second rest intervals. These steps were repeated for each position (supine and standing) at the same time for 5 consecutive days. Intrasession reliability was computed using 12 measurement trials (12), 12 measurement trials dropping the high and low score (12x), the first 7 measurement trials dropping the high and low score (7x), and the first 5 measurement trials (5). Intrasession and intersession reliability over five consecutive days was estimated using intraclass correlation coefficients (ICC (3, 1)). The supine intrasession reliability measurements were as follows: 0.932 (12), 0.932 (12x), 0.935 (7x), and 0.932 (5). The standing intrasession reliability was 0.853 (12), 0.852 (12x), 0.865 (7x), and 0.862 (5). The intersession reliability was 0.938 in the supine position and 0.803 in the standing position. These results indicate that the H-reflex measured using our protocol in a supine and standing position is a reliable assessment within sessions and between sessions. Five measurements are sufficient to observe reliable measurements within a single session. Most importantly, this data shows that the H-reflex is a reliable assessment that may be used to measure small changes in motoneuron pool excitability over time.     

Phase-specific modulation of the soleus H-reflex as a function of threat to stability during walking

The purpose of the present study was to determine whether the soleus H-reflex is modulated with changes in the level of postural threat during walking. H-reflexes were tested at four points in the step cycle when subjects walked in 5 conditions representing different levels of postural threat. H-reflexes were significantly increased in amplitude at heelstrike in conditions of increased postural threat compared to normal treadmill walking with only minimal changes in H-reflex amplitude at other step cycle points. Conversely when subjects walked while holding stable handles, to decrease postural threat, the amplitude of the H-reflex was significantly smaller at heelstrike and midstance compared to normal walking. The changes in the amplitude of the H-reflex between walking conditions were not accompanied by changes in ongoing electromyographic activity or movements. Our findings suggest that the amplitude of the reflex is adjusted in a phase-specific manner, related to the postural uncertainty of the task. These adaptations in reflex amplitude may be related to changes in the amplitude of corrective responses following perturbations during walking. The adaptations in the amplitude of the H-reflex specific to heelstrike may be important in the control of foot placement at ground contact.     

On the origin of the soleus H-reflex modulation pattern during human walking and its task-dependent differences

Recently, Brooke and colleagues have suggested "that the strong inhibition arising from passive movement about the knee and hip joints, lays down the base for the soleus H-reflex gain modulation seen during human gait." In particular stretch-evoked afferent activity from the quadriceps muscle was emphasized as the most important source of movement-induced inhibition of the H-reflex. To test this hypothesis we examined the kinematics and electromyographic (EMG) activity of the leg during human walking and correlated these with the modulation pattern of the soleus H-reflex. To further test the possible contribution of stretch-evoked quadriceps afferent activity to the soleus H-reflex modulation pattern during walking different walking gaits were studied. In one condition subjects were asked to walk with their knee locked in full extension by a rigid knee brace. In a second condition subjects were asked to walk backwards. During normal walking, the soleus H-reflex modulation pattern is strongly correlated with the EMG events of the soleus and tibialis anterior (TA), but not with hip, knee, or ankle angular displacement or velocity. When subjects walked with the knee locked in full extension, the amplitude of the H-reflex, its modulation pattern, and the task-dependent changes of its amplitude were the same as during normal walking. During backward walking, the H-reflex increases in late swing before activity of the soleus has begun and while the knee is flexing, an observation that highlights central control of the H-reflex amplitude. The effects of imposed flexion of the knee in passive subjects were also reexamined. The knee flexion imposed by the experimenter followed the same trajectory as that which occurred during the swing phase of the subject's step cycle. It was found that imposed knee flexions elicited a burst of TA EMG activity with an average latency of 81.6 ms (SD = 21 ms) in six out of eight subjects. Inhibition of the H-reflex, when it occurred, was associated with the occurrence of this burst. When subjects voluntarily flexed their right knee from an initial quiet standing posture, the inhibition of the soleus H-reflex began before flexion of the knee or that of any other leg segment. Once again the onset of inhibition was closely associated with the onset of activity in the TA. In the discussion section the present observations are examined in light of the predictions made by the movement-induced inhibition hypothesis of Brooke et al. It will be concluded that none of the predictions of this hypothesis were corroborated by present tests done during human walking. In consequence, we suggest that the modulation pattern of the H-reflex observed during normal human walking is centrally determined, as are the task-dependent differences of its amplitude (e.g., standing versus the stance phase of human walking).     

Progressive adaptation of the soleus H-reflex with daily training at walking backward

When untrained subjects walk backward on a treadmill the amplitude of the soleus H-reflex in midswing is equal to or exceeds the value in stance. This is a surprising result because during the swing phase of backward walking the soleus is inactive and its antagonist, the tibialis anterior, is active. We suggested that the high amplitude of the soleus H-reflex in late swing reflects task uncertainties, such as estimating the moment of foot contact with the ground and losing balance. In support of this idea we show that when untrained subjects held on to handrails the unexpected high-amplitude H-reflex during midswing was no longer present. We therefore asked whether daily training at this task without grasping the handrails would adaptively modify the H-reflex modulation pattern. In this event, within 10 days of training for 15 min daily, the anticipatory reflex activity at the beginning of training was gradually abated as the subjects reported gaining confidence at the task. However, when adapted subjects were made to walk backward with their eyes shut, the anticipatory reflex activity in midswing returned immediately. The reflex changes as a result of training were not due to changes in the motor activity or kinematics; they are likely part of the motor program controlling backward walking. This adaptive phenomenon may prove to be a useful model for studying the neural mechanisms of motor learning and adaptive plasticity in humans and may be relevant to rehabilitation programs for neurological patients.     

Contribution of peripheral afferents to the activation of the soleus muscle during walking in humans

Summary Small, rapid stretches were applied to the soleus muscle during the stance phase of walking by lifting the forefoot with a pneumatic device. Stretch responses were induced in the soleus muscle by the disturbance. The amplitude and time course of the responses from the soleus muscle were a function of both the kinematics of the disturbance and the time in the step cycle when the disturbance was applied. The step cycle was divided into 16 equal time parts, and data obtained within each of these parts were averaged together. The electromyographic (EMG) response of the soleus muscle showed a time course that was similar to the time course of the angular velocity induced by the disturbance at the ankle. Three linear equations were used to predict the EMG response from the soleus muscle as a function of the angular kinematics of the disturbance: 1) velocity, 2) velocity and displacement, 3) velocity, displacement and acceleration. Introduction of a pure delay between the EMG and the kinematics substantially improved the predictions. Most of the variance (70%) in the EMG response could be accounted for by the velocity of the disturbance alone with an optimal delay (average 38 ms). Inclusion of a displacement term significantly increased the variance accounted for (85%), but further addition of an acceleration term did not. Since the velocity of the disturbance accounted for most of the variance, the reflex gain was estimated from the velocity coefficient. This coefficient increased in a ramp-like fashion through the early part of the stance phase, qualitatively similar to the increase in the H-reflex. Based on these identified gains, this reflex pathway was estimated to contribute substantially (30% to 60%) to the activation of the soleus muscle particularly during the early part of the stance phase.     

Temporal depression of the soleus H-reflex during passive stretch

Synaptic efficacy associated with muscle spindle feedback is regulated via depression at the Ia-motoneurone synapse. The inhibitory effects of repetitive Ia afferent discharge on target motoneurones of different sizes were investigated during a passive stretch of ankle extensors in humans. H-reflex recruitment curves were collected from the soleus muscle for two conditions in ten subjects. H-reflexes were elicited during passive stretch at latencies of 50, 100, 300, and 500 ms after a slow (20°/s) dorsiflexion about the right ankle (from 100 to 90°). Control H-reflexes were recorded at corresponding static (without movement) ankle angles of 99, 98, 94, and 90° of flexion. The slope of the H-reflex recruitment curves (Hslp) was then calculated for both conditions. H-reflex values were similar for the static and passive stretch conditions prior to 50–100 ms, not showing the early facilitation typical of increased muscle spindle discharge rates. However, the H-reflex was significantly depressed by 300 ms and persisted through 500 ms. Furthermore, less than 300 ms into the stretch, there was significantly greater H-reflex depression with a lower stimulus intensity (20 % Mmax) versus a higher stimulus intensity (Hmax), though the effects begin to converge at later latencies (>300 ms). This suggests there is a distinct two-stage temporal process in the depression observed in the Ia afferent pathway for all motoneurones during a passive stretch. Additionally, there is not a single mechanism responsible for the depression, but rather both heterosynaptic presynaptic inhibition and homosynaptic post-activation depression are independently influencing the Ia-motoneurone pathway temporally during movement.     

Reciprocal inhibition of soleus motor output in humans during walking and voluntary tonic activity

1. The extent to which an active, human motoneuron pool can be inhibited via short-latency inhibitory pathways was studied by stimulating the common peroneal nerve and recording the inhibition of on-going soleus electromyographic (EMG) activity. The responses were compared at the same EMG level during walking and tonic voluntary activity to determine whether the inhibition was task dependent. 2. In both tasks the amount of inhibition (measured as the depression in rectified, filtered, and averaged EMG activity) increased approximately linearly with the amount of motor activity, as determined from the mean EMG level before stimulation (correlation coefficient greater than or equal to 0.9). No difference in the amount of inhibition was found between the two tasks at the same stimulus and EMG levels. 3. Previously published studies based on the H-reflex method have reported that the amount of inhibition decreases with the amount of motor activity. On the contrary, single-unit studies and the present results suggest that segmental inhibitory reflexes retain their capacity to mediate a rapid reduction of motoneuronal discharge during voluntary activity. This inhibition may be important in regulating the amount of activity early in the stance phase of walking and during the transition from stance to the swing phase. 4. Analytic results are derived in an APPENDIX that should be of general interest in interpreting the inhibition of motor units from a peristimulus time histogram (PSTH). The linear correlation between inhibition and level of voluntary activity can be explained if newly recruited units are strongly inhibited by the stimulus, whereas previously active motor units are inhibited relatively less, as their firing rate increases with increasing background activity.     

Somatosensory graviception inhibits soleus H-reflex during erect posture in humans as revealed by parabolic flight experiment

The purpose of this study was to investigate how gravity level affects the excitability of the soleus muscle (SOL) motoneuron pool to Ia afferent input while erect posture is maintained in humans. Three healthy male subjects participated in an experiment whereby three different gravity conditions [microgravity (MG), normal gravity (NG), and hypergravity (HG)] were imposed using a parabolic flight procedure. The SOL H-reflex was evoked every 2 s while the subjects kept an erect posture. The stimulus intensity was controlled automatically on a real-time basis by personal computer to induce the constant amplitude of M-wave (10±5% of maximal M-wave amplitude). The background electromyographic activity (BGA) of the SOL was largest during HG, while it was almost absent during MG. The SOL H-reflex amplitude was significantly larger during HG and MG than during NG (P<0.05). During NG and HG, there was a linear relationship between the BGA and the H-reflex amplitude; the difference in the SOL H-reflex amplitude between both gravity conditions could be explained in terms of the BGA level. However, during MG, despite the absence of BGA, the SOL H-reflex amplitude was larger than that during NG. Furthermore, when the subjects voluntarily activated the SOL by applying a load to the lower limb joints and spine by pulling a handle upward, this H-reflex enhancement almost disappeared. These results suggest that the somatosensory systems detecting a load at the lower limbs and/or vertebral column might play a role in reducing the excitability of the SOL motoneuron pool to Ia afferent inputs by presynaptic inhibition. Electronic Publication