Design of BCI Based Multi-information System to Restore Hand Motor Function for Stroke Patients

The rehabilitation on limb paralysis after stroke is an international scientific and technological issue. The rehabilitation in the early stage could hardly realize active participation of patients´ intention, self-adaptive functional compensation and quantitative measurement on rehabilitation effect. This paper proposed a brain-computer interface (BCI) training system for rehabilitation of hand motor function after stroke. The mechanical hand mounted on patients´ hand was driven by their intention through online motor imagery electroencephalogram (EEG) signal. The mechanical hand provided real-time adaptive assistance or resistance according to motor condition measured by transducers. The system displayed real-time and dynamic information on EEG, hand force and angle synchronously for rehabilitation evaluation and program optimization. The system was tested on three healthy subjects. An average accuracy and information transfer rate of 89.7% and 0.5099bit/s were obtained respectively. The system provides new idea and approach for the rehabilitation training of stroke patients. In follow-up work, we will test our system in long-term rehabilitation experiment for stroke patients to evaluate the effectiveness on improving hand movement and neural system activation.