Dynamic spinal reflex adaptation during locomotor adaptation

Abstract

The dynamics and interaction of spinal and supraspinal centers during locomotor adaptation remain vaguely understood. In this work, we use Hoffmann reflex measurements to investigate changes in spinal reflex gains during split-belt locomotor adaptation. We show that spinal reflex gains are dynamically modulated during split-belt locomotor adaptation. During first exposure to split-belt transitions, modulation occurs mostly on the leg ipsilateral to the speed change and constitutes rapid suppression or facilitation of the reflex gains, followed by slow recovery to baseline. Over repeated exposure, the modulation pattern washes out. We further show that reflex gain modulation strongly correlates with correction of leg asymmetry, and cannot be explained by speed modulation solely. We argue that reflex modulation is likely of supraspinal origins and constitutes an integral part of the neural substrate underlying split-belt locomotor adaptation.NEW & NOTEWORTHY This work presents direct evidence for spinal reflex modulation during locomotor adaptation. In particular, we show that reflexes can be modulated on-demand unilaterally during split-belt locomotor adaptation and speculate about reflex modulation as an underlying mechanism for adaptation of gait asymmetry in healthy adults.

Keywords: H-reflex; human gait; locomotion; locomotor adaptation; spinal reflex.

Graphical abstract

Figure 1.

A: experimental setup. Participants walked on a split-belt treadmill instrumented with force plates. Ground reaction force was used to trigger the stimulator at midstance. The stimulus elicited an H-reflex that was recorded using a wireless electromyography (EMG) sensor. Body kinematics was tracked using a motion capture system. B: experimental protocol. Sessions consisted of 4 tied-belt and 3–6 split-belt trials. Tied-belt trials consisted of 5 min (∼250 strides) of walking each at 100% and 70% of the participants’ preferred speed before and after split-belt trials. Split-belt consisted of 5 min of walking: 1 min of tied-belt walking at 100% of preferred speed, then split-belt walking where one leg was kept at 100% (contra leg) while the other was reduced to 70% of the preferred speed (ipsi leg), then 2 min of tied-belt walking at 100% preferred speed. The same speed sequence was repeated 3–6 times.

Figure 2.

H-reflex measurements and step length asymmetry (SLA) for an example participant from the ipsi group (ID: ipsi05). H-reflex measurements are the raw GAS EMG data following nerve stimulation. SLA is normalized to individual step lengths (see Step Length Asymmetry Definition for details) and is smoothed by a 3-stride moving average filter.

Figure 3.

Adaptation. A: normalized H-reflex amplitudes during tied-to-split adaptation in the ipsi group during the first and third split-belt blocks. The pattern of adaptation washes out over blocks. B: step-length asymmetry (SLA) in the ipsi group during the first and third blocks. Similar to H-reflex adaptation, SLA pattern washes out over trials. C: correlation between SLA and H-reflex amplitudes during adaptation. The ipsi group showed a strong positive correlation, while the contra group showed no pattern of correlation. Error bars in A and B represent 1 means ± SE and 1 SD, respectively.

Figure 4.

De-adaptation. A: normalized H-reflex amplitudes during split-to-tied de-adaptation in the ipsi group during the first and third split-belt blocks. The pattern of adaptation washes out over blocks. B: step-length-asymmetry (SLA) in the ipsi group during the first and third blocks. Similar to H-reflex adaptation, the SLA pattern washes out over trials. C: correlation between SLA and H-reflex amplitudes during de-adaptation. The ipsi group showed a strong positive correlation, while the contra group showed no pattern of correlation. Error bars in A and B represent 1 means ± SE and 1 SD, respectively.

Figure 5.

Speed-dependent modulation. Unlike late adaptation, early adaptation shows a statistically significant difference from tied-belt walking at the same speed. This suggests that early adaptation suppression of the H-reflex cannot be explained by speed-dependent modulation.