A locally-restricted method for removal of eye-blink artifacts in electroencephalogram measurements

Eye blinks are among the dominant artifacts in electroencephalogram (EEG) recordings. Although several removal methods for eye-blink artifacts have been reported, most of them concurrently removed a part of the original EEGs as well. In this study, we focused on the number of pure EEGs that were retained, and proposed a locally-restricted removal method based on independent component analysis (ICA), empirical mode decomposition (EMD), and the Kalman filter. Simulated and measured results confirmed that the proposed method is more suitable for the local removal of eye-blink artifacts without the loss of EEG information when compared to previous methods.