Robotically facilitated training of the hemiparetic upper extremity as an integrated functional unit in virtual environments

This study describes a novel robotic system using haptic effects and objects, in rich, three- dimensional virtual environments (VEs) for the sensorimotor training of the hemiparetic hand. This system is used to train the hand and arm together as an integrated functional unit. Eight subjects with subacute or chronic strokes practiced two to three hours/day for 8 days using four different robotic simulations that were controlled with a combination of arm and finger movement. Subjects improved 27% in the Wolf Motor Function Test, 22% in the Jebsen Test of Hand Function and 9% in the Box and Blocks Test. Subjects also improved in several kinematic measures during their interactions with the VR simulations. Subjects demonstrated a more efficient trajectory in a VE-based activity that involved reaching and interacting with an object using their fingers. Smoothness of these trajectories also improved suggesting that subjects required fewer sub-movements to approach the target, which is consistent with improved motor control. During virtual piano training, subjects showed improvement in the ability to isolate finger movement and increase speed of the movement with minimal effect on accuracy We feel that our initial findings suggest that training the arm and hand as a unit following stroke may be effective for improving upper extremity function.