loading page

SpikeAnts, a spiking neuron network modelling the emergence of organization in a complex system
  • Sylvain Chevallier
Sylvain Chevallier
Université de Versailles Saint-Quentin

Corresponding Author:[email protected]

Author Profile

Abstract

The emergence of organization is at the core of many complex systems, from neural cell assemblies to living insect societies. Since information processing by a Spiking Neuron Network (SNN) is based on temporal dynamics, SNNs appear to be powerful neural information processing systems for modelling the emergence of collective phenomena in complex systems. A spatio-temporal model, called SpikeAnts, is proposed for reproducing the emergence of synchronized activities in social insect colonies: Each ant agent is modelled by only two spiking neurons and the ant colony is a sparsely connected SNN. Individual decision making results from the competition of an active and a passive neurons which process information received from neighbouring agents. The spatio-temporal coupling of the ant agents results in three types of temporal patterns at the population scale: asynchronous, synchronous aperiodic, and synchronous periodic, similar to real-world behaviour observed in ant societies. A phase diagram of the emergent synchronization patterns with respect to two control parameters, respectively modelling the ant sociability and receptivity, is presented and discussed.