Neurophysiological substrates of stroke patients with motor imagery-based brain-computer interface training

We investigated the efficacy of motor imagery-based Brain Computer Interface (MI-based BCI) training for eight stroke patients with severe upper extremity paralysis using longitudinal clinical assessments. The results were compared with those of a control group (n = 7) that only received FES (Functional Electrical Stimulation) treatment besides conventional therapies. During rehabilitation training, changes in the motor function of the upper extremity and in the neurophysiologic electroencephalographic (EEG) were observed for two groups. After 8 weeks of training, a significant improvement in the motor function of the upper extremity for the BCI group was confirmed (p < 0.05 for ARAT), simultaneously with the activation of bilateral cerebral hemispheres. Additionally, event-related desynchronization (ERD) of the affected sensorimotor cortexes (SMCs) was significantly enhanced when compared to the pretraining course, which was only observed in the BCI group (p < 0.05). Furthermore, the activation of affected SMC and parietal lobe were determined to contribute to motor function recovery (p < 0.05). In brief, our findings demonstrate that MI-based BCI training can enhance the motor function of the upper extremity for stroke patients by inducing the optimal cerebral motor functional reorganization.