The brain possesses a highly fluctuating dynamic, with time scales much shorter than the changes in connectivity. Many factors are responsible for the emergence of this behavior, that favor a metastable or critical dynamic and not allowing the system to settle down in a single attractor.
Using numerical simulations of different models of large-scale brain activity, we are disentangling the precise factors that contribute to the particular behavior of human brain-inspired networks. We have shown that the human connectome contains a critical core of highly interconnected nodes (s-core) that have the possibility of high and low activity states at a low value of global connectivity weight G. As G is further increased, the recruitment of new nodes able to sustain high activity is fairly gradual, a sign of dynamical richness because the network can have multiple ‘ignited’ states. This gradual recruitment is lost when the s-core is disrupted by randomizing the network, even if some other topological features are conserved such as degree distribution or small-worldness. We are also studying the high-order interactions in the activity of the brain, that can be measured in rs-fMRI recordings and become more redundant (less synergistic) with aging. This result is also reproduced in mean-field models connected by the topology of the human connectome, and our simulations help us to understand the emergence of synergistic relationships from the structural connectivity and how this is related to the multistable dynamics.
Acknowledgments: Fondecyt 1181076, ANID-Basal FB0008, Instituto Milenio ICM-ANID P09-022-F
Send your questions to our Q & A Space: https://neurostars.org/t/workshop-spatiotemporal-dynamics-in-neuroimaging-models-and-analysis-q-a/7608?u=psanzleon
