Small-world structure induced by spike-timing-dependent plasticity in networks with critical dynamics

TL;DR

This study demonstrates that activity-dependent spike-timing-dependent plasticity can induce small-world network structures in integrate-and-fire neuron models at critical dynamics, affecting connectivity motifs and network efficiency.

Contribution

It identifies a specific learning rule that promotes small-world properties in neural networks during critical states, linking plasticity, network topology, and critical dynamics.

Findings

Small-world structures emerge at critical dynamics due to plasticity.

Bidirectional connectivity is reduced by the plasticity rule.

Power-law activity distributions characterize the critical state.

Abstract

The small-world property in the context of complex networks implies structural benefits to the processes taking place within a network, such as optimal information transmission and robustness. In this paper, we study a model network of integrate-and-fire neurons that are subject to activity-dependent synaptic plasticity. We find the learning rule that gives rise to a small-world structure when the collective dynamics of the system reaches a critical state which is characterised by power-law distributions of activity clusters. Moreover, by analysing the motif profile of the networks, we observe that bidirectional connectivity is impaired by the effects of this type of plasticity.