Quantifying Task-Specific Active Patient Participation during Robot-Assisted Gait Training for Spinal Cord Injury

Active patient participation is a key element for effective robot–assisted gait training, but available studies have assessed it only without focusing on task-specific demands, i.e., conditions requiring joint– and phase–specific engagement. The aim of this study was to evaluate active patient participation, quantified as human–robot interaction work at joint level, at the beginning and end of a 12–session Lokomat training in 10 individuals with incomplete spinal cord injury, under controlled task–specific conditions. In each session, participants performed eight tasks targeting specific combinations of joints (hip or knee) and gait phases (stance or swing), enabling the systematic evaluation of side–, joint– and phase–specific selectivity in their voluntary activity. Results showed side–specific selectivity, as participants were able to actively engage the correct limb. Phase-specific selectivity emerged only in swing–targeted tasks. During swing, joint–specific selectivity was observed, although participants were not able to fully decouple joints of the same limb. Finally, selectivity and total positive interaction work did not deteriorate throughout the training period.