The use of functional magnetic resonance imaging (fMRI) to measure functional connectivity among brain areas has the potential to identify neural networks associated with particular cognitive processes. However, fMRI signals are not a direct measure of neural activity but rather represent blood oxygenation level-dependent (BOLD) signals. Correlated BOLD signals between two brain regions are therefore a combination of neural, neurovascular, and vascular coupling. Here, we describe a procedure for isolating brain functional connectivity associated with a specific cognitive process. Coherency magnitude (measuring the strength of coupling between two time series) and phase (measuring the temporal latency differences between two time series) are computed during performance of a particular cognitive task and also for a control condition. Subtraction of the coherency magnitude and phase differences for the two conditions removes sources of correlated BOLD signals that do not modulate as a fu...
Michael A. Silver, Ayelet N. Landau, Thomas Z. Lau