This seminar focuses on the statistical physics of complex network modelling for various physical problems, showing the effectiveness of such approach.
The meeting will also focus on the determination of their correct null case and in particular on analytic models reproducing the local features of the network.
The application of these model is valid both to detect statistically significant structural patterns in real networks (by testing them against their correctly-defined null hypothesis), and both in order to reconstruct the network structure in case of incomplete information. As case study, an analysis of brain networks from fMRI, showing how brain regions tend to coordinate by forming a highly hierarchical chain-like structure of homogeneously clustered anatomical areas. A maximum spanning tree approach revealed the centrality of the occipital cortex and the peculiar aggregation of cerebellar regions to form a closed core.
Finally, a report on the hierarchy of network segregation and the level of clusters integration as a function of the connectivity strength between brain regions, and an outline of applications to detect diseases.