利用基于稳态视觉诱发电位的脑机接口控制功能性电刺激康复系统

In this paper, a functional electrical stimulation (FES) training system with steady-state visual evoked potential (SSVEP) based brain-computer interface (BCI) was designed to realize movements of upper limb. The subjects are required to initiatively focuse on one of the five flicking lights with different frequencies on the computer screen, and some corresponding reactions will occur at the visual cortex of the brain. At the same time, the electroencephalogram (EEG) signal where the subject's intention is encoded is acquired. After filtering the primary flicking frequencies, the harmonic components are extracted as classification features from the EEG channels at the visual cortex region, and then linear discriminant analysis (LDA) classifier is used to decode the subject's intention corresponding to the flicking light that the subject is focusing on. Thereafter a command with user's intention is sent to trigger the FES system to perform the frequency specific training action. The experimental results showed that the feature extraction and classification methods are efficient in on-line classification. Finally an energy bar model is applied to the machine user interface to enhance the output of the system and as a feedback to the user. The results imply that the subjects can accurately control the FES training system to realize the designed action sequencies with their own intention, and this work lays a solid fundation for the future clinical experiments on paralized patients.