Removal of ocular artifacts from EEG signals using ICA-RLS in BCI

Abstract

Ocular artifacts are the most important form of interference in electroencephalogram (EEG) signals. An adaptive filter based on reference signals from an electrooculogram (EOG) can reduce ocular interference, but collecting EOG signals during a long-term EEG recording is inconvenient and uncomfortable for the subject. For removing ocular artifacts from EEG in EEG based brain computer interfaces (BCIs), a method named independent component analysis recursive least squares (ICA-RLS) is proposed. Firstly, ICA is used to decomposing multiple EEG channels into an equal number of independent components (ICs). The ocular artifacts significantly contribute to some ICs but not others. ICs that include ocular artifacts can be identified. Then adaptive filtering based on RLS uses reference signals from identified ocular ICs to reduce interference, which avoids the need for parallel EOG recordings. Based on the EEG data collected from seven subjects, the new method achieved a higher 6.7% classification accuracy than that of standard ICA method, which demonstrates a better ocular-artifact reduction by the proposed method.