The specificity of different-distance connections in human structural connectomes

Abstract

Brain structural connectomes underpin complex cognitive processes. To date, abundant organizational features have been distilled by network-based tools, including hubs, modules, and small-worldness. However, these features are often devoid of spatial characteristics which directly shape connection formation. By considering the spatial embedding of brain networks, we reveal the connection specificity, that is, the similarity of similar-distance connections and the dissimilarity of different-distance connections. It is induced by the whole-brain connection length distribution, allowing areas to send and receive diverse signals through different-distance connections. Based on it, areas' functional repertoires are associated with their connection length profiles, and meanwhile, length dispersion and clustering coefficients can be integrated into a hierarchy whose age-related degeneration may be related to cognitive decline. These results construct a putative bridge between brain spatial, topological, and functional features, expanding our understanding of how different architectures complement and reinforce each other to achieve complicated brain functions.