Computational study of neuronal networks with spatial constraints by means of the Izhikevich model

Abstract

A huge effort has been made along the last 20 years to map the detailed structural organisation of neural networks. The main reason behind this effort is to understand the relationship between the structure of the network and its dynamics or functional traits. To advance in this quest, numerical simulations have emerged to help exploring the relation between structure and function. Here we used the Izhikevich model to simulate neuronal networks with spatial constraints. We launched numerical simulations of 1000 neurons in two different modular networks, and mimicking designs reported in experiments. We observed that some information about structure can be glimpsed when the spatial constraints are very strong. In general, however, the properties of the underlying structural network differ greatly from those obtained from the simulations, indicating that the assessment of structural connectivity from just dynamics is not possible. We also applied damage to the networks, and observed that targeted attacks strongly affects the activity and functional traits of the networks under study