A hybrid method for artifact removal of visual evoked EEG

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Visual evoked Electroencephalogram (EEG) signals offer insights into brain circuitry.
• EEG signals are often corrupted by artifacts from eye movements, muscle activity, and electrical interference.
• Existing denoising methods can distort the underlying neural signals.

Purpose of the Study:

• To develop and evaluate a novel hybrid method for denoising visually evoked EEG signals.
• To effectively remove artifacts without compromising neural information.

Main Methods:

• A hybrid approach combining digital filters, Independent Component Analysis (ICA), and Transient Artifact Reduction Algorithm (TARA).
• ICA is used for automatic ocular artifact removal.
• TARA addresses remaining artifact interference.

Main Results:

• The proposed ICA-TARA method significantly improved Signal-to-Noise Ratio (SNR).
• SNR increased by 13.47% (simulated) and 26.66% (real data) after ICA.
• TARA further boosted SNR by 6.98% (simulated) and 71.51% (real data).
• Statistically significant improvements were observed (p<0.05).

Conclusions:

• The ICA-TARA method effectively eliminates major artifacts from visual evoked EEG.
• This approach preserves the integrity of underlying neural signals in both simulated and real data.
• It outperforms existing denoising techniques like wavelets and empirical mode decomposition.