Eye blink region detection in EEG using amplitude thresholding

Eye blink is one of the physiological artifacts that is commonly presents in electroencephalogram (EEG) data. Eye blink artifact is very impactful to process output because of its high energy content. Hence, eye blink-contaminated EEG region detection method is necessary most of the time during EEG pre-processing whether for eye blink removal or utilization. This paper presents a straightforward method to detect eye blink regions within EEG data without the need for EEG low-pass filtering or decomposition step that causes removal of some EEG spectrum. Eye blink region detection presented here utilized amplitude thresholding since it is composed of high amplitudes component. An experiment is conducted on the eye blink-contaminated EEG that were recorded from 10 participants using OpenBCI EEG equipment to test the ability of proposed method in detecting eye blink region presented in the EEG. The proposed method obtained accuracy percentage of 94.5% which may be comparable with method that utilizing EEG spectrum compromising method like data smoothing or filtering.