Automatic Ocular Artifact Suppression From Eeg Data By Using Statistics And Time-Frequency Properties Of Independent Components

Contamination of Electroencephalographic (EEG) recordings with different kinds of artifacts is the main obstacle to the analysis of EEG data. Independent Component Analysis (ICA) is now a widely accepted tool for detection of artifact in EEG data. This component-based method segregates artifactual activities in separate sources hence, the reconstruction of EEG recordings without these sources leads to artifact reduction. Identification of the artifactual components is a major challenge to artifact removal using ICA is the. Although, during past several years, it has been proposed for automatic detecting the artifactual component, there is still little consensus on criteria for automatic rejection of undesired components. In this paper we present a new identification procedure based on statistics and time-frequency properties of independent components for fully automatic ocular artifact suppression. By comparing the statistics and time-frequency properties of independent components, the artifactual components were identified and removed. The results on 2000 4-s EEG epochs indicate that the artifact components can be identified with an accuracy of 92.8%. Moreover, statistical test indicates that the statistics and time-frequency properties of artifactual components are significantly different from that of non-artifactual components.