A hybrid ICA Kalman predictor algorithm for ocular artifacts removal

Abstract

Electroencephalogram (EEG) is a record of the electrical activity of the brain and is a tool which gives an insight into the brain functions thereby helping the physicians to diagnose various abnormalities. Blinking or moving the eyes produces large electrical potential around the eyes known as Electrooculogram (EOG), which is a non-cortical activity that spreads across the scalp and contaminates the EEG and such contaminating potentials are called Ocular Artifacts (OA). EEG recordings are significantly distorted by OA, causing problems for analysis and interpretation by clinicians and a nuisance for researchers who investigate the electrophysiology of the brain. Hence, a control procedure for removal of OA from EEG is essential for interpreting EEG properly. In this paper, a hybrid ICA-Kalman Predictor algorithm is proposed to classify the independent components obtained from JADE (Joint Approximation Diagonalization of Eigen matrices) algorithm and remove OA automatically. The proposed design is implemented in matlab and the Mean Squared Error (MSE) of proposed Kalman method was 0.0017, significantly lower compared to results using ICA and ANC, which were 0.0468 and 0.0052 respectively.