Increased impairment of the hand is commonly reported as an outcome occurring after a stroke. This impairment is due mainly to the decreased ability to modulate appropriate muscle activation patterns. Issues with appropriately modifying activation patterns can profoundly affect tasks of daily living. This study is evaluating a novel software focusing on retraining hand muscle activation patterns through an Electromyographic (EMG) controlled game. Stroke survivors with chronic, severe hemiparesis of the hand will participate in a longitudinal study consisting of 3 weeks of training. We hypothesize that stroke survivors will experience a decrease in time to complete a test of EMG control, as well as improvement in hand motor control.