Automatic Artifact Detection and Elimination for EEG Signal Based on AR-Copula and ICA

During the acquisition process, EEG (Electroencephalograph) is inevitably contaminated by various artifacts. In this paper, an AR-Copula model is employed to analyze the correlation between the contaminated EEG and the related artifacts, and a tail-dependence-based automation detection algorithm is proposed. According to the detection results, ICA is adopted to eliminate the artifacts for the contaminated EEG data segments. It is shown that the proposed algorithm can automatically detect the contaminated data segments and remarkably reduce the iteration numbers of ICA algorithm. Moreover, the efficiency of data processing can be also improved for real-time application.