Human-Computer Confluence for Rehabilitation Purposes after Stroke

In this publication, we present a Motor Imagery (MI) based BrainComputer Interface (BCI) for neurologic rehabilitation. The BCI is able to control two different feedback devices. The first one is a rehabilitation robot, moving the fingers of the affected hand according to the detected MI. The second one presents feedback via virtual reality (VR) to the subject. The latter one visualizes two hands that the user sees in a first perspective view, which open and close according to the detected MI. Four healthy users participated in tests with the rehabilitation robot, and eleven post stroke patients and eleven healthy users participated to tests with the VR system. We present all subjects’ control accuracy, including a comparison between healthy users and people who suffered stroke. Five of the stroke patients also agreed to participate in further sessions, and we explored possible improvements in accuracy due to training effects.