Leila Cammoun1, Xavier Gigandet1, Olaf Sporns2, Jean-Philippe Thiran1, Kim Q. Do3, Philippe Maeder4, Reto Meuli4, Patric Hagmann1,4
1Signal Processing Laboratory 5, Ecole Polytechnique Fdrale de Lausanne, Lausanne, VD, Switzerland; 2Department of Psychological and Brain Sciences, Indiana University, Bloomington, IN, USA; 3Center of Psychiatric Neuroscience, University Hospital Center and University of Lausanne (CHUV), Lausanne, VD, Switzerland; 4Department of Radiology, University Hospital Center and University of Lausanne (CHUV), Lausanne, VD, Switzerland
We propose to map the human connectome by constructing normalized whole-brain structural connection matrices derived from diffusion spectrum MRI tractography at 5 different scales. These connection matrices are then examined across scales for several important network characteristics. We show that key network measures can be robustly estimated across multiple scales with results that are consistent with previous single-scale investigations. In particular, we confirm that cortical networks exhibit exponential distributions of node degree and fiber densities, robust small world characteristics, as well as consistent estimates for node centrality at all scales examined.