A stochastic explanation for observed local-to-global foraging states in Caenorhabditis elegans

TL;DR

This paper introduces a stochastic model for C. elegans foraging behavior that explains both abrupt and gradual reorientation changes without relying on discrete behavioral states, aligning with experimental observations.

Contribution

The model predicts local-to-global foraging transitions through stochastic reorientation rates, challenging the necessity of behavioral state changes.

Findings

Model reproduces observed foraging patterns

Abrupt and gradual changes explained without states

Behavioral variability across populations

Abstract

Abrupt changes in behavior can often be associated with changes in underlying behavioral states. When placed off food, the foraging behavior of C. elegans can be described as a change between an initial local-search behavior characterized by a high rate of reorientations, followed by a global-search behavior characterized by sparse reorientations. This is commonly observed in individual worms, but when numerous worms are characterized, only about half appear to exhibit this behavior. We propose an alternative model that predicts both abrupt and continuous changes to reorientation that does not rely on behavioral states. This model is inspired by molecular dynamics modeling that defines the foraging reorientation rate as a decaying parameter. By stochastically sampling from the probability distribution defined by this rate, both abrupt and gradual changes to reorientation rates can occur, matching experimentally observed results. Crucially, this model does not depend on behavioral states or information accumulation. Even though abrupt behavioral changes do occur, they may not necessarily be indicative of abrupt changes in behavioral states, especially when abrupt changes are not universally observed in the population.