Control of neural chaos by synaptic noise

Abstract

We studied neural automata -or neurobiologically inspired cellular automata- which exhibits chaotic itinerancy among the different stored patterns or memories. This is a consequence of activity-dependent synaptic fluctuations, which continuously destabilize the attractor and induce irregular hopping to other possible attractors. The nature of the resulting irregularity depends on the dynamic details, namely, on the intensity of the synaptic noise and on the number of sites of the network that are synchronously updated at each time step. Varying these details, different regimes occur from regular to chaotic. In the absence of external agents, the chaotic behavior may turn regular after tuning the noise intensity. It is argued that a similar mechanism might be at the origin of the self-control of chaos in natural systems.