Modified constraint induced movement therapy is a focus of rehabilitation research for stroke survivors with hemiplegia that relies on physically constraining a patient's unaffected limb to modify a patient's behavior and force the use of the affected limb in a rehabilitation setting. In this study, we attempt to cognitively induce a constraint in patients engaging in virtual therapy via a Kinect and Unity based platform. The platform included four games that each utilized a different active range of motion (ROM) exercise for control of game objects.

Five stroke survivors with hemiparesis were recruited for a two week study consisting of five, 25-minute therapy sessions each. Users were allowed to use either their unaffected or affected limb to play each game at any time and allowed their choice of games during each session. For each game, the difficulty level remained at baseline while the user used their affected side to play but the difficulty was increased at varying rates when the user used their unaffected side to play. Results were evaluated for compliance to the therapy (usage rate of affected side), choice of games, performance, efficacy of therapy (ROM), and qualitative behavior of users.

Compliance during virtual mCIT seems to be higher than that of traditional mCIT but a more extensive user test is necessary to validate this result. ROM measurements did not provide a clear trend due to the small size and duration of the data set. Trends were found in several areas of user behavior: (1) users preferred games utilizing only one axis of control and not reliant on orientation of the limb, (2) users also tended to play games in a certain order, (3) user performance increased over time, and (4) users often incorrectly performed the desired active ROM exercise for each game which may indicate the necessity to observe and coach patients during virtual rehabilitation.