Meta-analytic maps of brain activity evoked by cognitive function diverge from resting state networks

Abstract

Spontaneous human brain activity is organized into resting state networks (RSN), complex patterns of synchronized activity that account for the majority of brain metabolism. The correspondence between these patterns and those elicited by task performance suggests that spontaneous brain activity originates from the stream of ongoing cognitive processing. Here we investigate a large number of meta-analytic activation maps obtained from Neurosynth (www.neurosynth.org) to establish that task-rest similarity can be inflated by two potential sources of bias. Applying a hierarchical module detection algorithm to a network representation of activation map similarity, we showed that the correspondence between RSN and task-evoked activity tends to hold only for the largest spatial scales. Second, we established that this correspondence is biased by the inclusion of maps related to neuroanatomical terms (e.g. "parietal", "occipital", "cingulate", etc.). Our results challenge the cognitive origin of spontaneous brain activity, suggesting that anatomically-constrained homeostatic processes could also play an important role in the inception and shaping of human resting state activity fluctuations.