An Eigenmode-Based GLM Method For Task-fMRI Data Analysis

Motivation: In traditional fMRI experiments, the BOLD signal is influenced by the spatial distribution of veins, which is closely linked to the morphological characteristics of the cortex. Goal(s): Cortical eigenmode decomposition represents a frequency-domain approach for analyzing brain structures, yielding a set of spatial bases for dissecting large-scale brain activities. Approach: In this study, we introduced an eigenmode-based General Linear Model method to investigate the influence of spatial patterns on the activation of specific fMRI tasks. Results: The results reveal a strong correlation in spatial distribution between the reconstructed z-map and the conventional activation map. Impact: Quantitative cortical eigenmode analysis offers a frequency-domain perspective for integrating structural and functional neuroimages. Eigenmodes encode connectivity patterns within the cortical structure, offering a promising avenue for unveiling implicit connections across cortical surface through their application to brain activity analysis.