Common Spatial Pattern Using Multivariate EMD for EEG Classification

Brain-computer interface (BCI) is a system to translate humans thoughts into commands. For electroencephalography (EEG) based BCI, motor imagery is considered as one of the most effective ways. This paper presents a method for classifying EEG during motor-imagery by the combination of well-known common spatial pattern (CSP) with so-called multivariate empirical mode decomposition (MEMD), which is effectively suitable for processing of multichannel signals of EEG. In the proposed method, the EEG signal is decomposed into intrinsic mode functions (IMF) using the MEMD. Different from EMD, the number of IMF is the same in each channel. Then by removing some of the IMFs, the reconstructed signal can carry more useful information than the original signal. Based on the MEMD, weights of CSP are found. By off-line simulation, the use of MEMD in CSP has shown to perform well in the application to the classification of EEG signals.