Independent Component Analysis of Eyeball Movements

Lee, Il Keun; Ha, Choong Kun

Abstract

Independent Component Analysis (ICA) is a signal processing algorithm to separate independent sources from unknown mixed signals and can be applied to separate artifacts and independent neural sources from EEG recordings. This study was designed to extract individual components of eyeball movements from scalp EEG. Digital EEG signals were recorded using the international 10-20 system during eye closure, eye opening, and blinking. 18 EEG tracings using bipolar montage were analyzed by ICA algorithm into 18 independent components. Each of the components was reviewed, selected, and reconstructed into an original montage. Among 18 components, two components which were thought to represent eyeball movements were obtained. Each of the components was inversely projected into the original bipolar montage. This inverse projection showed separated vertical and horizontal eyeball movements components. These results suggest that the ICA analysis of EEG can separate vertical and horizontal eyeball movements and may be applied to separate other EEG artifacts and source signals from unknown mixed sources recordings of EEG.