Extracting components containing maximal information for EEG based-brain computer interface

Always, one of the issues in the brain-computer interface (BCI) is to extract components from raw EEG data that have more information in order to separate task-related potentials from other neural and artifactual EEG sources. In this paper, a new method is proposed for extracting components from raw EEG data such that these components have maximal information for separating task-related potentials from other potentials. The effectiveness of the proposed method is evaluated by using the classification of EEG signals. The tasks to be discriminated are the imaginative hand movement and the resting state. The results demonstrate that the proposed mutual information-based component extraction (MIEC) algorithm performed well in several experiments on different subjects and can improve the classification accuracy in the BCI systems. The results show that the classification accuracy obtained by MIEC is higher than that achieved by independent component analysis (ICA) and original EEG signals.