Modulation of spinal reflex by assisted locomotion in humans with chronic complete spinal cord injury

Abstract

Objective

In healthy subjects, spinal reflexes (SR) evoked by non-noxious tibial nerve stimulation consist of an early (60–120 ms latency) and an occasional late-appearing (120–450 ms latency) component in the ipsilateral tibialis anterior. In chronic (>1 year) complete spinal cord injured (cSCI) subjects early components are small or lacking while late components are dominant. Here we report on the modulation of SR by assisted locomotion in healthy and chronic motor cSCI subjects.

Methods

SR was evoked by tibial nerve stimulation at the terminal stance phase during assisted locomotion and was compared to SR recorded during upright stance.

Results

In chronic cSCI subjects only a late SR component was consistently present during upright stance. However during assisted locomotion, an early SR component appeared, while amplitude of the late SR component became small. In contrast, in healthy subjects the early SR component dominated in all conditions, but a small late component appeared during assisted locomotion.

Conclusion

A more balanced activity of early and late SR components occurred in both subject groups if an appropriate proprioceptive input was provided.

Significance

Early and late SR components are assumed to reflect the activity of separate neuronal circuits, which are associated with the locomotor circuitry possibly by shaping the pattern.

Introduction

After a complete spinal cord injury (cSCI) spinal networks below the level of lesion are capable of producing rhythmical muscle activity in the mouse (Leblond et al., 2003), rat (Lavrov et al., 2006), cat (Rossignol et al., 2001) and humans (Dietz et al., 1994) if appropriate input, i.e. from hip joints and load receptors (Dietz et al., 2002), is provided. In cSCI subjects a pattern of leg muscle activation similar to that of healthy subjects can be recorded (Dietz et al., 1994), but electromyography (EMG) amplitude is much reduced. Nevertheless, after a cSCI several changes in the neuronal behaviour below the level of lesion can be observed over time after injury.

In chronic (>1 year after SCI) cSCI subjects an exhaustion of locomotor activity during assisted locomotion over 10 min of walking was recently described (Dietz and Muller, 2004). A degradation of spinal neuronal function was assumed to occur. In parallel also polysynaptic spinal reflexes (SR) show progressive changes in amplitude after a human cSCI over time (Hiersemenzel et al., 2000, Roby-Brami and Bussel, 1987).

Corresponding to rodent experiments, the term ‘spinal reflex’ was defined here as the EMG response to a probably below-nociceptive threshold stimulation to tibial nerve, as noxious stimulus threshold can hardly be determined in cSCI subjects (Dietz, 2010, Dietz et al., 2009). SR evoked by tibial nerve stimulation in cSCI subjects is assumed to be a polysynaptic reflex and most probably corresponds to the cutaneous reflex described elsewhere (Duysens et al., 2004). It also corresponds to the SR evoked in rats. In this condition the appearance of the polysynaptic SR was associated with the recovery of locomotor function in rats with a transected spinal cord (Lavrov et al., 2006) indicating a relationship between SR- and locomotor-generating neuronal networks. Also in humans SR- and locomotion-generating networks are believed to share spinal neuronal circuitries (Bussel et al., 1989, Rossignol et al., 2006). In healthy subjects SR evoked by tibial nerve stimulation consist of an early (latency 60–120 ms) and, in some cases, a small late (latency 120–450 ms) reflex component. In human cSCI SR are abolished during spinal shock (Hiersemenzel et al., 2000). The early component of the SR reappears during “the transition phase to spasticity” and around 5–6 months after injury an additional late component appears (Hiersemenzel et al., 2000). The late component becomes dominant while a successive loss of the early component occurs (Dietz et al., 2009). The time course of the loss of the early SR component correlates with the degree of the exhaustion of leg muscle EMG during assisted locomotion in chronic cSCI subjects (Dietz et al., 2009). These findings suggest a change in neuronal functions below the level of lesion after a cSCI. They are of major interest because chronic SCI subjects can only profit from future regeneration-inducing therapies if spinal neuronal networks below the level of lesion are preserved (Curt and Dietz, 2005).

The temporal relationship between the changes in SR behaviour and locomotor activity over time is an additional indicator that SR- and locomotor-generating networks might share spinal neuronal circuitries. An activation of locomotor generating networks should therefore also affect the behaviour of SR generating neuronal pathways. The aim of this study was to analyse the interaction between spinal reflexes and locomotor activity in chronic cSCI subjects. We hypothesise that the behaviour of SR components becomes affected even in chronic cSCI subjects if an appropriate afferent input is provided during assisted locomotion. The results might give further insight into the function of neuronal circuits mediating early and late SR components, as well as locomotor activity, during the course of a cSCI.