Independent Vector Analysis for Gradient Artifact Removal in Concurrent EEG-fMRI Data.

Area of Science:

• Neuroimaging
• Biomedical Signal Processing
• Machine Learning

Background:

• Simultaneous electroencephalogram (EEG) and functional magnetic resonance imaging (fMRI) acquisition is valuable for studying brain activity.
• Gradient artifacts are a major challenge in concurrent EEG-fMRI, compromising signal integrity.
• Existing methods struggle with uncontrollable factors like head movement and B0 field fluctuations.

Purpose of the Study:

• To develop and validate a novel method for removing gradient artifacts from EEG signals acquired during fMRI.
• To improve the robustness and accuracy of artifact removal by utilizing spatio-temporal information.
• To enhance the quality of EEG data for more reliable analysis in combined EEG-fMRI studies.

Main Methods:

• Independent Vector Analysis (IVA), an extension of Independent Component Analysis (ICA) for multiple datasets, was employed.
• The method exploits the quasi-periodicity of artifacts across epochs and channel similarity.
• Spatio-temporal information, including spatial dependencies across channels, is used for artifact estimation.

Main Results:

• The proposed IVA-based method effectively estimates and removes gradient artifacts from EEG signals.
• Simulated data demonstrated the method's ability to handle gradient artifacts.
• Real EEG data collected concurrently with fMRI confirmed the desirable performance and robustness of the technique.

Conclusions:

• Independent Vector Analysis offers a robust and effective solution for gradient artifact removal in concurrent EEG-fMRI.
• The method's utilization of spatio-temporal information enhances its performance against acquisition variabilities.
• This technique significantly improves the quality of EEG data, facilitating more accurate neuroscientific research.