The Frequent Subgraphs of the Connectome of the Human Brain

Abstract

In mapping the human structural connectome, we are in a very fortunate situation: one can compute and compare graphs, describing the cerebral connections between the very same, anatomically identified small regions of the gray matter among hundreds of human subjects. The comparison of these graphs has led to numerous recent results, as the (i) discovery that women's connectomes have deeper and richer connectivity-related graph parameters like those of men, or (ii) the description of more and less conservatively connected lobes and cerebral regions, and (iii) the discovery of the phenomenon of the Consensus Connectome Dynamics. Today one of the greatest challenges of brain science is the description and modeling of the circuitry of the human brain. For this goal, we need to identify sub-circuits that are present in almost all human subjects and those, which are much less frequent: the former sub-circuits most probably have functions with general importance, the latter sub-circuits are probably related to the individual variability of the brain structure and functions. The present contribution describes the frequent connected subgraphs (instead of sub-circuits) of at most 6 edges in the human brain. We analyze these frequent graphs and also examine sex differences in these graphs: we demonstrate numerous connected sub-graphs that are more frequent in female or the male connectome. While our results describe subgraphs, instead of sub-circuits, we need to note that all macroscopic sub-circuits correspond to an underlying connected subgraph. Our data source is the public release of the Human Connectome Project, and we are applying the data of 426 human subjects in this study.

Abstract: arXiv descriptive metadata (CC0). Thank you to arXiv for use of its open access interoperability. This service was not reviewed or approved by, nor does it necessarily express or reflect the policies or opinions of, arXiv.