Self-organized criticality hierarchical modular networks of Galves-Löcherbach neurons
De Flavio Rusch
One of the observed features of the neocortex is the existence of activity avalanches characterized by power-law size and duration distributions, putatively suggesting that the neocortex operates at a critical state. Several models have been proposed to explain how neocortical networks can reach a critical state through self-organizing mechanisms. These self-organized criticality (SOC) models have explored networks with different topologies, e.g. fully connected and Erdős-Rényi. Here we study SOC in a hierarchical modular network using stochastic neurons of the Galves-Löcherbach type. The system has two mechanisms that make the critical region an attractor of the SOC dynamics: (i) dynamical gains, which produces adaptation in the neuronal firing rates, and (ii) dynamical synapses, which represent homeostatic mechanisms. We characterize the size and duration avalanches displayed by the model and study the emergence of synchronized activity across the network modules.