A Highly Replicable Multi-Scale Whole-Brain Atlas of Canonical Functional Networks from Over 100,000 Resting-State fMRI Datasets

Motivation: There is a need for both standardized terminology as well as a universal reference space in the field of neuroscience. Goal(s): NeuroMark represents a data-driven approach towards establishing a common reference space in neuroimaging research. Here we present a functional brain atlas which translates the NeuroMark template into the field of neuroscience. Approach: We manually labeled each of the 105 networks based on visual inspection of their spatial maps, computed similarity with multiple other atlases, and ordered and grouped them into 7 domains and 14 subdomains based on functional and spatial similarity. Results: The resulting modules reveal many insights into the brain. Impact: There are many atlases in the field, however, this one is highly replicable across individuals, datasets, and studies, and the multi-scale approach offers great promise towards helping to standardize both terminology and methodology in the field of neuroscience.