A feature extraction technique of EEG based on EMD-BP for motor imagery classification in BCI

The aim of this paper is to investigate a nonlinear approach for feature extraction of Electroencephalogram (EEG) signals in order to classify motor imagery for Brain Computer Interface (BCI). This approach consists of combining the Empirical Mode Decomposition (EMD) and band power (BP). Considering the non-stationary and nonlinear characteristics of the motor imagery EEG, the EMD method is proposed to decompose the EEG signal into set of stationary time series called Intrinsic Mode Functions (IMF). These IMFs are analyzed with the bandpower (BP) to detect the caracteristics of sensorimotor rhythms (mu and beta). Finally, the data were reconstructed with only with the specific IMFs and then the band power is employed on the new database. Once the new feature vector is reconstructed, the classification of motor imagery is applied using Hidden Markov Models (HMMs). The results obtained show that the EMD method allows the most reliable features to be extracted from EEG and that the classification rate obtained is higher and better than using only the direct BP approach. Such a system appears as a particularly promising communication channel for people suffering from severe paralysis, for instance for persons with myopatic diseases or muscular dystrophy (MD) to move a joystick to a desired direction corresponding to the specific motor imagery.