Changes in the number of muscles synergies (MSs) and in the weighting of muscles composing each MS are typically altered following an incomplete spinal cord injury (iSCI). To overcome this problem, locomotor training with a wearable robotic exoskeleton (WRE) represents a promising rehabilitation option though the effects of the various WRE control modes on MSs remain unknown.

This case series aims to characterize how WRE control modes affect the number of MSs and the weighting of muscles composing each MSs in individuals with iSCI.

Three participants with a chronic iSCI walked at a self-selected comfortable speed without and with a WRE set in different trajectory controlled (maximal assistance [MAX]; assistance-as-needed [ADAPT]) and non-trajectory controlled modes (high assistance [HASSIST], high resistance [HRESIST], and NEUTRAL). Recorded surface EMG of eight L/E muscles was used to extract the MSs using a non-negative matrix factorization algorithm. Cosine similarity and weighting relative differences characterized similarities between MSs of individuals with iSCI and against references obtained in a healthy control.

The mode providing movement assistance within a self-selected L/E trajectory (HASSIST) best replicates MSs observed in healthy control during overground walking. The MSs extracted with the passive (MAX) and assistance-as-needed (ADAPT) modes differed to the greatest extent from MSs characteristics extracted from a healthy reference.

Most of the control modes selected during overground walking with a WRE did not closely replicate the motor control required for the production of coordinated L/E movements during stereotypical overground walking. These first results may allow rehabilitation professionals to refine WRE locomotor training protocols.