Gaussian Elimination-Based Novel Canonical Correlation Analysis Method for EEG Motion Artifact Removal.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Neuroscience

Background:

• Motion artifacts significantly degrade electro-encephalography (EEG) signal quality, impacting diagnostic accuracy.
• Current artifact removal methods, including Canonical Correlation Analysis (CCA), Ensemble Empirical Mode Decomposition (EEMD), and Wavelet Transform (WT), face challenges in highly noisy conditions and computational efficiency.

Purpose of the Study:

• To introduce a novel CCA-based approach utilizing Gaussian elimination for enhanced motion artifact removal in EEG signals.
• To improve filtering performance and reduce computational time compared to existing methods, particularly in noisy environments.

Main Methods:

• A new CCA approach employing Gaussian elimination for calculating correlation coefficients via backslash operation.
• Solving linear equations to compute Eigen values, thereby reducing CCA's computational cost.
• Testing the proposed method against EEMD-CCA and WT using synthetic and real EEG signal data.

Main Results:

• The proposed Gaussian elimination-based CCA method demonstrated improved artifact removal efficiency.
• Significant reduction in filter computation time was observed, especially under highly noisy conditions.
• Evaluation metrics including DSNR, lambda (λ), RMSE, elapsed time, and ROC parameters confirmed the method's effectiveness.

Conclusions:

• The novel Gaussian elimination-based CCA algorithm is a suitable supplement to existing EEG artifact removal techniques.
• This approach offers a computationally efficient and effective solution for motion artifact removal in noisy EEG data.