Independent Component Analysis of Localized Resting State fMRI Reveals Specific Motor Sub-Networks

Abstract

Recent studies have shown that blood oxygen level-dependent low-frequency (< 0.1 Hz) fluctuations (LFFs) during a resting-state exhibit a high degree of correlation with other regions that share cognitive function. Initial studies of resting-state network mapping have focused primarily on major networks such as the default mode network, primary motor, somatosensory, visual, and auditory networks. However, more specific or subnetworks, including those associated with specific motor functions, have yet to be properly addressed. We performed independent component analysis (ICA) in a specific target region of the brain, a process we name, ‘‘localized ICA.’’ We demonstrated that when ICA is applied to localized fMRI data, it can be used to distinguish resting-state LFFs associated with specific motor functions (e.g., finger tapping, foot movement, or bilateral lip pulsing) in the primary motor cortex. These ICA components generated from localized data can then be used as functional regions of interest to map whole-brain connectivity. In addition, this method can be used to visualize interregional connectivity by expanding the localized region and identifying components that show connectivity between the two regions.