Enhancing EOG Artifact Removal from EEG Through Independent Component, Band Power, and Pearson’s Correlation

Eye movements introduce artifacts in EEG recordings, thus hindering accurate analysis. To detect and remove electrooculogram (EOG) components from EEG data, we commonly use independent component analysis (ICA). However, existing methods may be inconsistent, thus necessitating better EOG artifact removal. Our study presents an improved EOG artifact removal method that employs band power analysis and Pearson’s correlation coefficient (PCC). This method enhances accuracy and efficiency in identification and removal of EOG artifacts from EEG data. We applied the fast Fourier transform to compute band power (1–2 Hz) for each ICA-derived component. A threshold of 10
⁶
μV
²
identified EOG components with excessive band power. Subsequently, PCC values quantified the impact of each component on EEG channels. This approach accurately pinpointed EOG components for removal. Our proposed method significantly improved EEG signal quality. Combining band power and PCC analysis refined EOG component identification, thus facilitating targeted artifact removal. This method can potentially enhance EEG studies, particularly in eye movement-rich scenarios.