Temporal processing and context dependency in C. elegans mechanosensation

TL;DR

This study reveals that C. elegans' mechanosensory responses depend on temporal signal properties and behavioral context, with a predictive linear-nonlinear model capturing these dynamics.

Contribution

It demonstrates that C. elegans processes mechanosensory signals based on temporal features and context, introducing a new model for predicting behavior.

Findings

Behavior depends on stimulus temporal properties like integral and derivative.

Responses are modulated by behavioral context, such as during turns.

A linear-nonlinear model accurately predicts behavioral responses.

Abstract

A quantitative understanding of how sensory signals are transformed into motor outputs places useful constraints on brain function and helps reveal the brain's underlying computations. We investigate how the nematode C. elegans responds to time-varying mechanosensory signals using a high-throughput optogenetic assay and automated behavior quantification. In the prevailing picture of the touch circuit, the animal's behavior is determined by which neurons are stimulated and by the stimulus amplitude. In contrast, we find that the behavioral response is tuned to temporal properties of mechanosensory signals, like its integral and derivative, that extend over many seconds. Mechanosensory signals, even in the same neurons, can be tailored to elicit different behavioral responses. Moreover, we find that the animal's response also depends on its behavioral context. Most dramatically, the animal ignores all tested mechanosensory stimuli during turns. Finally, we present a linear-nonlinear model that predicts the animal's behavioral response to stimulus.