State diversity in different neural network structures

Abstract

The objective of this work is to study the capability to replicate different states of some different topological configurations named as feed forward, sunrise and circular with different inter-module connections. To achieve this, the Izhikevich mathematical model has been used for excitatory regular spiking neurons. A sort of simulations has been done with different values of the noise intensity for a total of 7 different configurations, in order to analyze the evolution of the functional complexity ϕFC and to find the value of the maximum with an adjustment. Finally, the maximum values have been compared and it has been identified that the best configuration to replicate different states of the neural network is the feed forward with all the inter-modular connections in the same direction, showing the important effect of directionality.