A reaction diffusion-like formalism for plastic neural networks reveals dissipative solitons at criticality

TL;DR

This paper introduces a reaction-diffusion-like formalism for plastic neural networks with STDP-inspired learning, revealing the emergence of dissipative solitons at criticality, which are analytically characterized and studied for their dynamics.

Contribution

It develops a novel reaction-diffusion formalism for neural plasticity and analytically describes the formation and interaction of dissipative solitons near critical points.

Findings

Dissipative solitons emerge at the transition to instability.

Stable dynamics depend on a non-linearity criterion.

Analytical solutions for solitons are derived.

Abstract

Self-organized structures in networks with spike-timing dependent plasticity (STDP) are likely to play a central role for information processing in the brain. In the present study we derive a reaction-diffusion-like formalism for plastic feed-forward networks of nonlinear rate neurons with a correlation sensitive learning rule inspired by and being qualitatively similar to STDP. After obtaining equations that describe the change of the spatial shape of the signal from layer to layer, we derive a criterion for the non-linearity necessary to obtain stable dynamics for arbitrary input. We classify the possible scenarios of signal evolution and find that close to the transition to the unstable regime meta-stable solutions appear. The form of these dissipative solitons is determined analytically and the evolution and interaction of several such coexistent objects is investigated.