Complete artifact removal for EEG recorded during continuous fMRI using independent component analysis.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Simultaneous electroencephalography (EEG) and functional magnetic resonance imaging (fMRI) offers combined electrophysiological and hemodynamic insights into brain dynamics.
• EEG recordings in MRI scanners are susceptible to significant artifacts from imaging, ballistocardiography (BCG), and eye movements.
• Existing artifact removal methods, like averaged artifact subtraction (AAS), have limitations, particularly with BCG variability and lack of ocular artifact handling.

Purpose of the Study:

• To develop and validate a comprehensive method using Independent Component Analysis (ICA) for simultaneous removal of BCG and ocular artifacts from EEG data acquired during fMRI.
• To assess the efficacy of the proposed ICA method in cleaning EEG signals contaminated by MRI environment disturbances, including residual artifacts after AAS.
• To compare the performance of the ICA method against AAS for BCG artifact removal and evaluate its impact on the availability of analyzable data.

Main Methods:

• Application of Independent Component Analysis (ICA) to simultaneously identify and remove ballistocardiographic (BCG) and ocular artifacts from EEG data.
• Integration of the ICA method with existing averaged artifact subtraction (AAS) techniques to address residual MRI contamination.
• Testing the ICA method on event-related potentials (ERPs) recorded during a visual oddball paradigm in a simultaneous EEG/fMRI setup.

Main Results:

• The ICA method demonstrated high effectiveness in attenuating BCG and ocular artifacts, leading to clearer reconstructed brain signals from EEG acquired within the MRI scanner.
• ICA significantly outperformed AAS in removing the ballistocardiographic artifact, addressing its limitations related to cardiac wave variability.
• Complete suppression of ocular artifacts by ICA resulted in a greater number of available trials for subsequent event-related potential (ERP) analysis.

Conclusions:

• The developed ICA-based method provides a comprehensive solution for artifact removal in simultaneous EEG/fMRI recordings, significantly improving data quality.
• ICA offers a superior alternative to AAS for mitigating BCG artifacts and introduces effective ocular artifact suppression, which is currently lacking in standard EEG/fMRI processing.
• The ICA method does not introduce systematic bias into the ERP waveform, as confirmed by comparisons with EEG data acquired outside the MRI scanner, validating its utility for accurate neurophysiological analysis.