BCG Artifact Removal for Reconstructing Full-scalp EEG inside the MR Scanner

Area of Science:

• Neuroimaging
• Biomedical Engineering
• Signal Processing

Background:

• Simultaneous EEG/fMRI acquisition is crucial for brain research.
• Ballistocardiogram (BCG) artifacts severely contaminate EEG signals, hindering analysis, especially in continuous recordings.
• Existing BCG removal methods struggle with continuous data, necessitating new approaches.

Purpose of the Study:

• To develop and evaluate a novel method for estimating and removing BCG artifacts from EEG data acquired during simultaneous EEG/fMRI.
• To introduce the Direct Recording Temporal Spatial Encoding (DRTSE) approach for BCG artifact reduction.
• To demonstrate the efficacy of DRTSE compared to established methods like Optimal Basis Set (OBS).

Main Methods:

• A high-density EEG cap was used in a specially designed experiment.
• A subset selection scheme identified 20 optimal channels out of 256 for BCG estimation.
• The Direct Recording Temporal Spatial Encoding (DRTSE) method integrated learning and inference for artifact estimation.
• DRTSE was combined with direct subtraction and optimization for EEG signal reconstruction.

Main Results:

• The DRTSE method, utilizing an optimization-based approach, significantly improved EEG signal reconstruction.
• Normalized Root Mean Square Error (RMSE) was reduced by approximately 13-fold compared to the OBS method.
• The study demonstrated that BCG artifacts from all channels can be accurately estimated from a small subset of channels.

Conclusions:

• DRTSE offers a robust and efficient solution for BCG artifact removal in simultaneous EEG/fMRI.
• The proposed method enhances the quality of EEG data, enabling more reliable interpretation of brain activity.
• DRTSE shows particular promise for continuous, nonevent-related EEG studies where averaging is not feasible.