Study on pattern recognition of EEG based on imagination and hand movement

Electroencephalography (EEG) is the reaction of the overall activities of the brain neurons. In the researches of Brain Computer Interface (BCI), the pattern recognition of EEG which is associated with mental tasks is the most important part of the BCI system. In this paper, data of ¿ wave and ß wave of C3, C4, P3 and P4 channels are certificated to be the proper sources for feature extraction, and the power spectral densities and the modules means of the data are selected to be the main components of the eigenvector. Then, an improved neural network model is established to complete the classification using eigenvectors above. In addition, some experiments have also been done using a four-channel EEG measurement system. The result shows that the recognition model proposed in this paper has a good characteristic of classification for mental tasks based on imagination and hand movement. And the method selected in this paper is valuable and effective.