An ICA-based method for automatic eye blink artifact correction in multi-channel EEG

Eye blink artifact in EEG should be corrected before further analysis because it does the most remarkable harm to the analysis result. Independent component analysis (ICA) has already shown to be an effective and applicable method for EEG de-noising. While ICA is used to correct ocular artifact, the key lies in how to identify the artifact component. Now many processing methods have been proposed. However, most of these methods were manually or semi-automatically implemented, which was time-consuming. In this work, an ICA-based method was proposed to automatically correct eye blink artifact in the case of no reference EOG from multi-channel EEG recordings. Sample entropy, a measure of data regularity, was used to identify the artifact components after ICA decomposition. The method is easily realized and can automatically correct the eye blink artifact without artificial interference. The proposed method has been tested by EEG data collected during rapid a go/nogo visual categorization task. It is demonstrated that the proposed ICA-based method, combined with sample entropy, is appropriate for automatic eye blink artifact correction in multi-channel EEG recordings.