Removing eye movement and power line artifacts from the EEG based on ICA

The presence of different artifacts has long been a problem for the analysis and interpretation of electroencephalographic (EEG) recordings. Independent Component Analysis (ICA) is a technique for Blind Source Separation (BSS) and has been used to remove biomedical artifacts. In order to subtract the power line noise effectively and robustly, we use two additional channels of artificial power line signals and the EEG recordings to form a new data set as the input of ICA, and the data were separated into independent components based on the extended ICA method, and the EEG signal was reconstructed by excluding those components related to eye movement and line artifacts. Experiment results indicated the method is effective and satisfying.