Alessandra Griffa1,
2, Richard Betzel3, Kim Q. Do4, Philippe Conus5,
Patric Hagmann, 16, Jean-Philippe Thiran1, 6
1Signal
Processing Laboratory (LTS5), Ecole Polytechnique Fdrale de Lausanne (EPFL),
Lausanne, Vaud, Switzerland; 2Department of Radiology, University
Hospital Center (CHUV) and University of Lausanne (UNIL), Lausanne, Vaud,
Switzerland; 3Department of Psychological and Brain Sciences,
Indiana University, Bloomington, IN, United States; 4Center for
Psychiatric Neurosciences, Lausanne University Hospital (CHUV) and University
of Lausanne, Lausanne, Vaud, Switzerland; 5Department of
Psychiatry, Lausanne University Hospital (CHUV), Lausanne, Vaud, Switzerland;
6Department of Radiology, Univerity Hospital Center (CHUV) and
University of Lausanne (UNIL), Lausanne, Vaud, Switzerland
Diffusion MRI, tractography and graph analysis allowed to characterize the brain structural architecture as a small-world, hierarchically modular network. The study of the brain modular topology is raising new interest, and could be a key approach for the understanding of neurodevelopmental disorder. In this framework, it is important to individuate representative partitions for whole groups of subjects. In this work we use information theory-derived measures to quantify the inter-subjects variability of structural network modular decomposition, and we propose different approaches (and particularly the consensus clustering algorithm) to individuate a group-representative partition.