Classification of four-class motion imagination tasks based on EEG by combining EEG source imaging with convolution neural networks

Goal: With the rapid development of electroencephalogram (EEG) technology, the brain-computer interface (BCI) based on motion imagination (MI) has been widely used. Aiming at the problem of low classification accuracy of multi-task MI, this paper adopted an innovative method. Method: This paper combines EEG source imaging with convolution neural networks to optimize the classification problem. Result: The results showed that the proposed method improved the classification accuracy compared with other studies. Significance: Scouts and convolution neural networks are applied to classify EEG signals, which provides a new idea for classifying EEG signals.