Learning how to average brain networks (i.e., build a brain network atlas) constitutes a key step in creating a reliable ‘mean’ representation of a set of normal brains, which can be used to spot deviations from the normal network atlas (i.e., abnormal cases). However, this topic remains largely unexplored in neuroimaging field. In this work, we propose a network atlas estimation framework through a non-linear diffusion along the local neighbors of each node (network) in a graph. Our evaluation on both developing and aging datasets showed a better ‘centeredness’ of our atlas in comparison with the state-of-the-art network fusion method.
This abstract and the presentation materials are available to members only; a login is required.