A least across-segment variance (LASV) method for the correction of EEG-fMRI desynchronization

Abstract

Simultaneous collection of electroencephalogram (EEG) and functional magnetic resonance imaging (fMRI) is a promising neuroimaging technique, which can provide high resolution in both spatial and temporal domain. Because EEG recorded in MRI scanners is heavily contaminated with gradient artefact (GA), removal of GA from EEG is a crucial step in EEG-fMRI data analysis. To date, the most efficient methods to remove GA are the average artefact subtraction (AAS) method and its extensions. However, these methods assume perfect synchronization between EEG and fMRI recording, which could be violated in practice. In this paper, a least across-segment variance (LASV) method is proposed for correcting EEG-fMRI desynchronization. Simulation and real data tests were conducted to check the performance of LASV method. The results suggested that the LASV method is able to efficiently correct EEG-fMRI desynchronization in both synthetic and real data, providing a powerful tool for improving the performance of GA removal for desynchronized EEG-fMRI data.

Gan Huang

My research interests include Neural Modulation, Brain Computer Interface and Neural Prosthetics.