Critical neuronal avalanches in levels model under noisy drive

TL;DR

This paper investigates how different noisy stimuli affect the critical avalanche scaling in a neuronal levels model, revealing that noise can alter the scaling exponent and introduce finite-size effects.

Contribution

It demonstrates the robustness of critical avalanche scaling under various noisy drives and shows how noise can modify the scaling exponent and finite-size behavior.

Findings

Scaling exponent varies with noise type

Finite-size effects depend on system size

Noise can shift the scaling exponent from 5/4

Abstract

We consider a neuronal levels model that exhibits critical avalanches satisfying power-law distribution. The model has recently explained a change in the scaling exponent from 3/2 to 5/4, accounting for a change in the drive condition from no input to moderate strength, along with a relaxed separation of time-scale between drive and dissipation. To understand the robustness of the scaling features, we examine the effect of different noisy stimuli in the moderate input regime. Our tool of analysis is the scaling method. We compute scaling functions associated with the avalanche size distribution, revealing striking finite-size scaling. For a class of noisy drives, we find that the scaling exponent can take a value different from 5/4, with an explicit system size dependence of the distribution.