Complexity Reduction of Rate-Equations Models for Two-Choice Decision-Making

TL;DR

This paper presents a method to simplify complex stochastic differential equations modeling neuronal decision-making circuits by exploiting slow-fast dynamics, enabling accurate macroscopic predictions across the entire phase space.

Contribution

It introduces a global complexity reduction technique based on slow-fast behavior, extending previous local methods to the entire phase space for neuronal models.

Findings

Reduced models accurately predict performance and reaction times.

Method applies globally, not just near spontaneous states.

Results align with previous local reduction approaches.

Abstract

We are concerned with the complexity reduction of a stochastic system of differential equations governing the dynamics of a neuronal circuit describing a decision-making task. This reduction is based on the slow-fast behavior of the problem and holds on the whole phase space and not only locally around the spontaneous state. Macroscopic quantities, such as performance and reaction times, computed applying this reduction are in agreement with previous works in which the complexity reduction is locally performed at the spontaneous point by means of a Taylor expansion.