ESTUDO DA FAIXA DINAMICA EM REDES NEURONAIS MODELADAS POR AUTOMATO CELULAR

Abstract

Cellular automata are mathematical models that can be used, among other things, to describe neuronal firing and interactions between neurons. They are simpler than coupled map lattices and oscillator chains, since cellular automata are characterized by discrete space, time, and state variables. The state variables are updated simultaneously according to the values of the variables in their neighborhood, by using deterministic or stochastic rules. The main goal of this work is the investigation of the dynamic range in a neuronal network modelled by celullar automata. Dynamic range is characterized by the capacity of a biological system to discriminate the intensity of an external stimulus. It is a range of intensities for which receptors can encode stimuli. In order to study the dynamic range in a neuronal network modeled by cellular automata, we consider a deterministic evolution rule to describe the neuronal dinamics. To model the interactions between the neurons through chemical and electrical synapses we have used deterministic and non-deterministic rules. By means of this models, it was possible to analyze the influence of chemical synapses and chemical time delay on the dynamic range. For a network with only electrical synapses, which describes intraglomerular region of the olfactory bulb, we found that the addition of weighted chemical synapses may contribute to improve the system sensitivity in response to external perturbation, i.e., to increase the dynamic range of neuronal network. We also verified the existence of bistability regions of the average firing rate and this system behavior is directly related to the region for which there is an increase in the dynamic range value.