EEG-EMG Analysis Method in Hybrid Brain Computer Interface for Hand Rehabilitation Training

Brain-computer interfaces (BCIs) have demonstrated immense potential in aiding stroke patients during their physical rehabilitation journey. By reshaping the neural circuits connecting patient’s brain and limbs, these contribute to restoration of motor functions, ultimately leading a significant improvement overall quality life. However, current BCI primarily relies on Electroencephalogram (EEG) imagery (MI), which has relatively coarse recognition granularity struggles accurately recognize specific hand movements. To address this limitation, paper proposes hybrid framework based Electromyography (EEG-EMG). The utilizes combination techniques: decoding EEG by using Graph Convolutional LSTM Networks (GCN-LSTM) subject’s motion intention, EMG convolutional network (CNN) identify In decoding, correlation between channels is calculated Standardized Permutation Mutual Information (SPMI), process further explained analyzing matrix. experiments are conducted two task paradigms, both achieving promising results. proposed validated publicly available WAL-EEG-GAL (Wearable for function recovery Electroencephalography Grasp-And-Lift) dataset, where average classification accuracies 0.892 0.954, respectively. This research aims establish an efficient user-friendly EEG-EMG BCI, thereby facilitating training patients.