Abstract
Contamination of blink artifacts in electroencephalographic (EEG) records often causes serious misconceptions in EEG analysis and/or the processing of evoked potentials because the amplitude of the blink artifact is much larger than that of the EEG. This paper proposes a method for real-time processing of blink artifact elimination on EEG records. The feature of the elimination method is acquisition of an appropriate feature of the blink artifact waveform in resards to the respective amplitudes and durations. For realizing real-time processing, an algorithm for blink artifact elimination, consisting of preprocessing of the EEG, blink artifact detection, normalized averaging and denormalization, was derived using on-line algorithms and implemented in real time. The proposed method exhibited satisfactory performance for eliminating blink artifacts on recorded EEG data acquired from five subjects.
