Collective Correlations of Brodmann Areas fMRI Study with RMT-Denoising

TL;DR

This study uses fMRI data and Random Matrix Theory to analyze collective activity patterns in Brodmann brain regions during rest and tapping tasks, revealing multi-region collective behavior.

Contribution

It introduces a novel application of RMT to interpret collective brain activity in Brodmann areas using PCA on fMRI correlation data.

Findings

Multiple Brodmann regions participate collectively in brain activity.

Eigenvector analysis reveals widespread regional involvement.

RMT provides insights into multivariate brain data structure.

Abstract

We study collective behavior of Brodmann regions of human cerebral cortex using functional Magnetic Resonance Imaging (fMRI) and Random Matrix Theory (RMT). The raw fMRI data is mapped onto the cortex regions corresponding to the Brodmann areas with the aid of the Talairach coordinates. Principal Component Analysis (PCA) of the Pearson correlation matrix for 41 different Brodmann regions is carried out to determine their collective activity in the idle state and in the active state stimulated by tapping. The collective brain activity is identified through the statistical analysis of the eigenvectors to the largest eigenvalues of the Pearson correlation matrix. The leading eigenvectors have a large participation ratio. This indicates that several Broadmann regions collectively give rise to the brain activity associated with these eigenvectors. We apply random matrix theory to interpret the underlying multivariate data.