Synchronization domains in two coupled neural networks

Abstract

We investigate the dynamical properties of two coupled neural networks with 2,048 identical Hodgkin--Huxley type bursting neurons. The internal connection architecture of each network follows a small-world topology and the external connection scheme is based on the local mean field potential, where one network receives the signal from the other. To analyze the system, we use Kuramoto order parameter computed over each neuron data, and recurrence quantification analyses, particularly the ratio of recurrent points belonging to diagonal lines of the recurrence plot RP, namely the determinism, computed over the local mean field potential of the networks, an easier experimentally accessible data. We analyze the complex synchronization scenario depicted by the network as a function of internal and external coupling parameters. Particularly, we identify regions of non-monotonic dependence of the synchronization level as a function of the coupling strength; coupling induced phase desynchronization (PD), where the synchronization levels are similar to those expected for randomly distributed phases; almost complete spike synchronization (SS) for which even spikes composing a burst are synchronized. This regime occurs as a product of the fast modulated signal imposed by the coupling between networks and weak internal coupling; finally, bursting synchronization (BS) regions are associated with slow modulated internal coupling.