Multi-channel EEG Classification Based on Fast Convolutional Feature Extraction

In this paper, we develop a novel feature extraction approach for multi-channel electroencephalography (EEG) classification. Inspired by convolutional neural networks (CNNs), we devise a fast convolutional feature extraction approach for EEG classification. In our approach, convolutional filters are first applied to extract features of multi-channel EEG signals. Then weak classifier selection is adopted to adaptively choose important features, which will be used for final classification. After that, we evaluate the performance of selected features through classification accuracy. Experiments on BCI III IVa competition dataset demonstrate the superior performance of our method, compared with the same classifier without feature extraction and deep learning methods, such as CNNs and long short term memory (LSTM). This work can be used to form the framework of deep neural networks for EEG signal processing.