Correction of eye-blink artifacts in EEG recordings using Wiener filtering

摘要

A novel method for removing eye-blink artifacts from brain EEG recordings is presented. It is proposed to use multichannel Wiener filter (MWF) indirectly. Namely, the eye-blink signal estimate is obtained first. For this purpose only a small subset of the frontal electrodes is used (so that extra EOG sensors are unnecessary). Then, the eye-blink estimate is subtracted from the noisy EEG signal in accordance with the regression technique. We have performed numerical experiments and compared this method to the commonly used independent component analysis (ICA) approach. Our experiments show that the proposed method can perform similarly to or even better than the ICA-based method. However, the MWF-based approach is much simpler and less computationally demanding.