Sequentially activated groups in neural networks

¹  Department of Control Theory, Nizhny Novgorod University - Gagarin Avenue, 23, 603950 Nizhny Novgorod, Russia
²  K. U. Leuven, ESAT-SCD/SISTA - Kasteelpark Arenberg 10, B-3001 Leuven (Heverlee), Belgium, EU

Corresponding author: makomar@mail.ru

Received: 27 February 2009
Accepted: 2 June 2009

Abstract

The internal neuronal dynamics, network configurations and coupling are influencing the output of neuronal networks through different external actions. One possible output form is sequential neuronal activity. This activity typically comprises many subsequent components generated by neuronal networks with a different action and spatial location. A stimulus-dependent formation of certain spatio-temporal structures takes place with sequences of groups of neurons that have a switching activity in time. We obtain spatio-temporal structures of activity in networks of randomly synaptically coupled neurons modeled as a Hodgkin-Huxley system. Necessary conditions for appearance of functional structures are presented. It is found that the asymptotic information capacity of the neuronal network exhibits a power law dependency. Its exponent can be very large and depends on the number of groups and the number of neurons per group.