An adaptive-threshold mechanism for odor sensation and animal navigation

Identifying the environmental information and computations that drive sensation is key for understanding animal behavior. Using experimental and theoretical analysis of AWC^ON, a well-described olfactory neuron in C. elegans, we here derive a general and broadly useful model that matches stimulus history to dynamic sensory and behavioral responses in odor sensation. We show that AWC^ON sensory activity is regulated by an absolute signal threshold that continuously adapts to odor history, allowing animals to compare present and past odor concentrations. The model predicts sensory activity and probabilistic behavior during animal navigation in different odor gradients and across a broad stimulus regime. Genetic studies demonstrate that the cGMP-dependent protein kinase EGL-4 determines the timescale of threshold adaptation, defining a molecular basis for a critical model feature. The adaptive threshold model filters stimulus noise efficiently, allowing reliable sensation in fluctuating environments, and represents a feed-forward sensory mechanism with implications in other sensory systems.