Resting-state functional connectivity magnetic resonance imaging (fcMRI) has recently emerged as a leading method for non-invasively characterizing the functional connections of the human brain. In fcMRI studies, a standard measure of connectivity between two brain regions is the temporal correlation between their respective resting-state functional MRI (fMRI) time series. For the computation of correlations, many studies use a pre-processing step known as global signal regression (GSR), in which a global mean signal is subtracted from all voxel time courses. While GSR can greatly improve the consistency and reliability of functional connectivity maps, its use is controversial because it may also introduce spurious negative correlations. Currently, there is not a clear consensus regarding how to best handle global signal confounds, with many fcMRI studies still continuing to use GSR, while others have adopted alternate methods due to concerns about GSR. This lack of agreement makes it difficult to compare fcMRI studies, as differing approaches can yield significantly different connectivity measures. The goals of this project are to develop a better understanding of the global signal and to use that knowledge to develop new methods for global signal correction.
The aims of the study are to (1) Develop and evaluate a new approach for global signal correction and (2) Determine whether global signal variations reflect changes in underlying neuroelectric coherence.
The proposed project will improve our ability to accurately measure functional connections in the brain using resting-state fMRI. Because the strength of functional connections may be an important indicator of the health of the brain, the proposed methods will benefit the study of a wide range of diseases.
