A method for automatic removal of EOG artifacts from EEG based on ICA-EMD

Abstract

According to the interference of the ocular artifacts in the measurement process of EEG, a method combined independent component analysis (ICA) and empirical mode decomposition (EMD) is proposed. Firstly, ICA is applied to the mixed signal including EEG and EOG so as to obtain the independent components. Secondly, EMD threshold denoising is used to remove the ocular artifacts which have larger amplitude in the independent components, then the EEG signals are rebuilt by using the inverse ICA based on the new independent components. In order to evaluate the effect of the method quantitatively, the simulation data containing EOG interference is constructed. The correlation coefficient and the mean square error are used as indexes to evaluate the denoising performance. The results show that the proposed method can automatically and effectively remove the EOG interference, the reserved EEG information provide good conditions for further feature extraction and pattern recognition.