Many methods have been proposed to remove artifacts from EEG recordings es- pecially those arising from eye movements and blinks. Often regression in time and frequency domain on parallel EEG and electrooculographic recordings is used, but this approach can become problematic in some cases. Use of Principal Com- ponent Analysis (PCA) has been proposed to re- move eye artifacts from multichannel EEG. This method is not effective when the activations from cerebral activity and artifacts have comparable amplitudes. In this paper it is presented a gener- ally applicable method for removing a wide vari- ety of artifacts from EEG recordings based on In- dependent Component Analysis (ICA) with high- order statistics. The method is applied with good results in the analysis of a sample lowpass event -related potentials (ERP) data.
Semantic parafoveal processing in natural reading: Insight from fixation‐related potentials & eye movements
