Spatiotemporal Analysis of Brain Activity Using EEG-fMRI Integration for Enhanced Neural Decoding

This study explores the integration of electroen- cephalography (EEG) and functional magnetic resonance imag- ing (fMRI) data for enhanced brain activity mapping. The analysis focuses on frequencies ranging from Delta to Gamma bands (0-120 Hz) with a specific emphasis on Gamma (30-45 Hz) during rest and visual stimulus tasks. EEG data were pre- processed and segmented into epochs from 0 to 400 seconds, excluding rest stimuli where correlation was minimal. To improve signal clarity, EEG data were whitened, and corresponding fMRI annotations were merged. Brain slices were temporally aligned, and regularized noise covariance was computed to filter out noise before performing inverse source localization. The visual cortex exhibited the highest activation levels during the stimuli. A decoding pipeline was constructed, with cross-validation and a spatio-temporal score of 99.2% achieved. Finally, sensor space patterns were projected to source space, utilizing linear classifiers to map brain activity with high precision.