These studies used both seed correlation maps and component maps to usefully demonstrate spatial features of correlated rs-fcMRI signal, but they also reveal limitations inherent to such approaches. For example, if one is interested in multiple regions highlighted in a seed correlation map, a new map needs to be generated for each additional region, and comparing more than a few maps at once quickly becomes overwhelming. Alternatively, if one is interested in interactions within or between components, one must adopt some other technique to examine those interactions, since component analyses only indicate that variance is shared to some extent by voxels within components. Such problems are amplified when multiple cohorts are examined (e.g. clinical or developmental cohorts), imposing substantial limits on the complexity and scope of brain relationships that can be grasped and communicated by researchers using these methods. Such limitations suggest the need for a more comprehensive framework for analyzing interactions between brain regions.
