Analysis of the instantaneous phase signal of a fMRI time series via the Hilbert transform

Abstract

Typically, functional magnetic resonance imaging (fMRI) data is processed in the time domain with linear methods such as regression and correlation analysis. We propose that the theory of phase synchronization may be used to understand the dynamics of interacting systems more completely and can be applied to fMRI data as a novel method of detecting activation. We performed phase synchronization analysis on the data from five volunteers for an event-related finger-tapping task. Functional maps were created that provide information on the interrelations between the instantaneous phases of the reference function and the voxel time series in a whole-brain fMRI activation data set. We conclude that this method of analysis is useful for revealing additional information on the complex nature of the fMRI time series.