Ubiquity of Uncertainty in Neuron Systems

TL;DR

This paper shows that unpredictability in neuron systems is widespread and inherent, arising from fundamental properties of multistable deterministic models, not just noise or complexity, impacting brain and cognition studies.

Contribution

It introduces the concept of 'chance synchronization' and demonstrates its role in unpredictability across various neuron models using basin analysis techniques.

Findings

Uncertainty is common in multistable neuron systems.

'Chance synchronization' explains emergence of unpredictability.

Deterministic models inherently exhibit unpredictability.

Abstract

We demonstrate that final-state uncertainty is ubiquitous in multistable systems of coupled neuronal maps, meaning that predicting whether one such system will eventually be chaotic or nonchaotic is often nearly impossible. We propose a "chance synchronization" mechanism that governs the emergence of unpredictability in neuron systems and support it by using basin classification, uncertainty exponent, and basin entropy techniques to analyze five simple discrete-time systems, each consisting of a different neuron model. Our results illustrate that uncertainty in neuron systems is not just a product of noise or high-dimensional complexity; it is also a fundamental property of low-dimensional, deterministic models, which has profound implications for understanding brain function, modeling cognition, and interpreting unpredictability in general multistable systems.