Numerical simulations of neuronal cultures with modular organization: uncovering the mechanisms that shape a rich dynamics

Abstract

The study of modular neuronal cultures has proven useful to provide insights into the mechanisms that shape a rich dynamics in the human brain, specifically the capacity to alternate between small-scale computation (segregation) and whole-network information exchange (integration). Motivated by experimental data, in this work we investigated the dynamics of simulated neuronal networks using the Izhikevich model, with the key ingredient that neurons were organized in modules coupled to one another with short- and long-range connections. We considered both cross-correlation and transfer entropy to analyze activity data and extract major metrics of network’s functional behavior, such as modularity and functional richness, and observed that these quantities were highly sensitive to the coupling among modules and noise, which is the main driver of spontaneous activity. We also assessed whether the resulting functional and effective networks retained information from the underlying structural connectivity. Our simulations, despite their simplicity and refinements left behind, may help experimentalists in the design of new in vitro systems to explore open questions of functionality and information processing in rich neuronal assemblies