Evolution of Residual Forearm Muscle Synergies During Intact-Limb sEMG Template-Guided Control Training

Residual-limb myoelectric control training involves motor learning and progressive adaptation of residual-limb muscle activation patterns. However, conventional assessments predominantly rely on behavioral metrics, providing limited insight into the evolution of underlying residual muscle control strategies. This preliminary study investigated phase-related changes during residual forearm motor training in five individuals with unilateral transradial limb absence. A closed-loop training paradigm with virtual-hand visual feedback was developed, in which intact-limb surface electromyography (sEMG) patterns during target movements served as individualized, task-specific templates for residual forearm training. Training progression was segmented into adaptation and stable phases to examine closed-loop performance, sEMG feature variability, and muscle synergy organization. Results showed that task completion time generally decreased during early training and reached a plateau in later trials. The stable phase was characterized by a higher proportion of correctly detected windows, shorter cumulative pause durations, and consistently lower residual forearm sEMG feature variability. Muscle synergy analysis showed reduced deviation from the corresponding terminal training patterns, with the most consistent change observed in the first component. Synergy weight-distribution measures also showed phase- and movement-related changes, while sensitivity analyses supported the principal synergy-deviation finding across synergy numbers and normalization strategies. These preliminary findings suggest that short-term improvement in closed-loop task performance is accompanied by phase-related changes in the consistency and low-dimensional organization of residual forearm muscle activity. The proposed measures may provide complementary indicators for characterizing myoelectric training.