Characterizing and utilizing fMRI fluctuations, patterns, and dynamics

Abstract

Functional MRI is fundamentally grounded in the hemodynamic response. With an increase in neuronal activity, blood flow increases, causing an increase in blood oxygenation, leading to an increase in transverse relaxation rate T2*. This increase in blood flow is slow and highly variable and shows a considerable spatial heterogeneity. In spite of these limitations, the hemodynamic response has been proven to be exquisitely sensitive to subtle differences in neuronal activity in time, over space, and between subjects. This paper is a brief review of my Keynote address describing some of the effort coming from my group that further demonstrates methods to robustly extract ever more information from both resting state fMRI and activation-induced fMRI. Specifically, I discuss 1) our new method to use multi-echo fMRI time series data collection to separate blood oxygen level dependent (BOLD) signal from non-BOLD signal, 2) activation of the whole brain obtained using a simple task and, importantly massive averaging and a model-free analysis approach, and 3) fMRI decoding of left vs right eye ocular dominance column activation with a timing offset as low as 100 ms. © 2013 SPIE.