Conference | Research

CORTEX conference by Demian Battaglia

On The October 14, 2014

Amphi ISC
67, bd Pinel
69500 Bron

Dynamics of Functional Connectivity

The dynamics of neuronal circuits are constrained by the anatomical connections of the brain (the so-called structural connectome), but are far from being fully determined by them. In particular multiple patterns of neuronal activity can be supported by a same underlying structural architecture, and these different dynamical states may lead to very different functional connectivity topologies, reflecting different modalities of information sharing. As illustrated by computational modeling of neuronal circuits at different spatial scales (from simple motifs involving a few interacting areas, to macroscale approximated simulations of an entire cortex), the sampling of a repertoire of internally-available dynamical states of neural activity is expected to give rise to stochastic switching transitions between alternative functional connectivity networks. Such spontaneous dynamics of functional connectivity might be easily biased by suitable control inputs, providing a substrate for the self-organized control of inter-areal communication. Our findings on dynamic functional connectivity in simulated systems provide a theoretical framework for the interpretation of correlation and information-theoretical analyses of actual electrophysiological and imaging data. Extensive local and global reconfigurations of information sharing networks are revealed in multichannel recordings from anaesthetised mice and awake monkey, and shown to be paced by global oscillatory configurations, despite their inherent "liquidity". At the macroscale, human resting state BOLD functional connectivity manifests a characteristic switching behavior, whose detailed statistical properties provide biomarkers which strongly correlate with the age of the considered subject. All together, these modelling and experimental results suggest a highly dynamic nature of functional interactions in the brain at different nested spatial and temporal scales, which might be completely hidden by the frequently performed averaging over long time-epochs, trials or subjects. We propose here that such functional connectivity dynamics is the emergent hallmark of self-organized complex activity of brain structural networks.