Excitation-inhibition balance in cortical networks with heterogeneous cluster sizes and its applications

TL;DR

This paper investigates how excitation-inhibition balance is maintained in cortical networks with heterogeneous cluster sizes, proposing a reweighing method to restore balance and enable controlled information propagation.

Contribution

It introduces a novel reweighing approach to restore balance in heterogeneous clustered networks and demonstrates controlled synchronization and stimulus propagation.

Findings

Reweighing connection strengths based on community sizes restores balance.

Partial balancing allows tuning of synchronization within communities.

Stimuli can propagate hierarchically without affecting weakly connected clusters.

Abstract

Insight into how information can propagate within cortical networks is essential for a more complete understanding of neural dynamics and computation in complex networks. Networks with clustered connections have previously been shown to give rise to correlated dynamics in individual clusters. However, this same model applied to a network with highly heterogeneous cluster sizes leads to a clear breakdown of the balanced state. In this article, using a formal definition of the balance matrix, we show why the balance condition breaks and propose a solution to restore balance in heterogeneous networks by reweighing the connection strengths based on community sizes. We introduce a method of partially balancing a heterogeneous network and show that the degree of spontaneous synchronization within communities can be varied using a single parameter describing the reweighing. We further show that stimuli can propagate through a hierarchically clustered network, where stimulating one cluster of neurons in a densely connected pair induces correlated firing in the other without propagating to other weakly connected clusters.