A Review of Most Current Feature Extraction Methods for EEG Signal Processing

For investigation of the brain signal, using the appropriate method for feature extraction is essential. During these years different methods have been offered for processing the EEG signals such as frequency domain, time domain and time-frequency domain. The methods used in this study are Wavelet transforms, CSP, Fourier transform, Eigenvector and EMD. The main purpose of this research is to investigate the advantages and disadvantages of these methods and provide extensive comparison between them. This study shows that between these methods EMD is the best method for feature extraction of the EEG signal because it is more adaptive to non-stationary signal