An Exoskeleton BCI System for Stroke Rehabilitation Using Multi-modality Training Mode

With the advent of the global aging era, the impact of stroke on humans is continuously increasing. Therefore, new therapeutic approaches to enhance patient recovery are receiving growing attention. Stroke is a severe neurological disorder that causes brain damage, requiring extended periods of rehabilitation for patients to return to normal life. Currently, common clinical methods largely rely on traditional passive limb rehabilitation, primarily targeting peripheral nerve training. Nowadays, an increasing number of people are recognizing the importance of post-stroke neuroplasticity. This study aims to activate patients' central nervous system and proposes a brain-machine interface (BCI) exoskeleton system based on human cognition. By designing three cognitive brainwave training models, including an attention-based standing-up model detecting θ-band power decrease, a relaxation-based sitting-down model detecting α-band power increase, and a walking model detecting MRCP, patients can actively control lower limb exoskeleton movement using brainwaves.This study provides cross-validation accuracy through both normal subjects and stroke patients, with accuracies of 87.13±9.38 and 80.83±9.68, respectively. Envisioning a rehabilitation system guided by human cognition, our team has developed an intuitive BCI exoskeleton system.