Experience-Dependent Gain Modulation Drives Thermosensory Responses in Behavior

Sensory neurons must extract behaviorally relevant features from dynamic environments while maintaining sensitivity across wide stimulus ranges. To understand how sensory encoding adapts to experience during behavior, we combine long-duration calcium imaging in freely moving C. elegans with a temperature-trajectory playback paradigm to determine how the thermosensory neuron AFD extracts behaviorally relevant sensory features during navigation. We observe that AFD functions as a leaky integrator of recently experienced temperature changes, accumulating thermal inputs over a rolling window of tens of seconds, resulting in calcium levels that represent recent temperature dynamics during runs. Importantly, we determine that AFD selectively amplifies responses to temperature changes near its learned preferred temperature. This experience-dependent gain control aligns encoding with the navigational goal, providing a mechanism for representing temperature preference within a derivative-based sensory system. A minimal mathematical model incorporating derivative detection, leaky integration, and temperature-dependent gain captures the calcium dynamics over a range of stimuli, and a simulation based on the mathematical model predicts goal-oriented locomotor strategies across stimulus regimes. Together, these findings show how gain control allows a derivative-based sensory code to represent an absolute goal and guide locomotory strategies during navigation. HighlightsO_LIAFD encodes recent temperature changes over behaviorally relevant timescales. C_LIO_LIAFD amplifies temperature-change responses within a learned temperature-centered gain window. C_LIO_LIGain control provides a mechanism for representing distance to temperature preference within the derivative-based sensory system. C_LIO_LIA mathematical model and simulation of derivative sensing, leaky integration, and gain control predicts AFD calcium responses and goal-oriented locomotory strategies across stimulus regimes. C_LI