Removal of EOG Artifacts from EEG Signals Using Hybrid Wavelet Transform-Adaptive Filtering and Independent Component Analysis

Electroencephalography (EEG) is a vital tool for diagnosing and monitoring neurological disorders, but it is highly susceptible to artifacts, particularly those originating from eye movements and blinks (electrooculographic, EOG artifacts). These artifacts significantly degrade signal quality, leading to misinterpretation of brain activity. This paper proposes a novel hybrid approach for EOG artifact removal, combining Wavelet Transform (WT), Adaptive Filtering (AF), and Independent Component Analysis (ICA). The method leverages the multi-resolution analysis of WT to decompose EEG signals, uses AF for precise artifact cancellation, and employs ICA to separate neural and non-neural components. The performance of the proposed technique is evaluated using simulated and real EEG datasets, with metrics including Signal-to-Noise Ratio (SNR), Mean Square Error (MSE), and correlation coefficients. Results demonstrate that the hybrid method outperforms conventional techniques such as standalone ICA, regression, and filtering, achieving superior artifact removal while preserving underlying neural activity. This approach offers a robust solution for enhancing EEG signal quality in clinical and research applications.