Increasing biological evidence suggests that the anatomical and functional organization of neuronal networks of the brain subtend their function. This talk will discuss a few questions related to this interplay and on which mathematical approach can shed new light.
First, I will discuss the characterization of large-scale dynamics of randomly connected neural networks, in particular taking into account their multi-scale connectivity patterns. I will investigate the role of noise and disorder in these macroscopic dynamics, with a particular focus on a mysterious and somewhat paradoxical role of noise or disorder in synchronizing neurons.
Closer from biology, I will present a new dataset and analyses on the functional organization of the visual cortex that tend to indicate that orientation and spatial frequency of simple stimuli organize within continuous maps with common punctual singularities that have a very particular pinwheel-dipole topology. Using arguments from planar topology, I will show that how these may reveal that exhaustivity and parsimony of representation may be fundamental organization principles of the primary visual area.