An estimated 15,000 Veterans suffer a stroke per year, more than 75% of whom have persistent hand impairment even after a full course of rehabilitation that significantly diminishes their independence. Fundamentally, post-stroke upper limb impairments arise from altered muscle activation patterns. Thus, direct training of muscle activation patterns holds promise. Unfortunately, altered muscle activation patterns are not fully addressed in current rehabilitation. Previous efforts have been limited to training a single pair of muscles, failing to address aberrant intersegmental coupling. To address this gap, a novel EMG-controlled game system has been developed: EMG signals are mapped to the 2-dimensional motion of a cursor on a computer screen to play games. In a preliminary study, the target EMG patterns were derived from 4 nonparetic distal upper limb muscle EMGs, while games were played using 4 paretic distal upper limb muscle EMGs mapped to the target patterns, thereby promoting use of the nonparetic muscle activation patterns in the paretic limb. Twenty stroke survivors improved upper limb function assessed by the Wolf Motor Function Test Functional Abilities (WMFT FA) significantly after 9 game sessions. Objective: To expand the capacity of the EMG-controlled game system and integrate it into a full therapy paradigm. Aim 1: Improve the system. Muscle activation patterns will be characterized with nonnegative EMG amplitudes to make the game control more intuitive. The number of EMG input channels will be increased to include both proximal and distal upper limb muscles for intersegmental training. More games will be created for engagement. Feasibility of the improved system will be evaluated with neurologically intact Veterans. Aims 2-3 will evaluate the system with Veteran stroke survivors. Veterans stroke survivors will be randomly assigned to one of two groups. The treatment group will receive 9 EMG game sessions and 9 task practice therapy sessions in an alternating order. The control group will receive 18 task practice therapy sessions. Evaluation will occur at baseline, post, and 1-month follow-up. Aim 2: Determine the effect of the EMG game system plus task practice therapy on upper limb function. Hypothesis: The group training with the EMG games in conjunction with task practice therapy will improve their upper limb function assessed by WMFT FA (primary) more than the group training with dose-matched task practice therapy only. Aim 3: Determine the neuromechanical changes associated with training with the EMG-controlled games. Hypothesis: The paretic muscle activation patterns will assimilate to the nonparetic pattern only for the treatment group. The similarity will be quantified by computing the angle between the hyperplane defined by the paretic EMG patterns and the plane defined by the nonparetic EMG patterns. Impact: This research will address the current treatment gap by developing/evaluating a novel EMG-controlled game system to...