Brain-wide processing of gustatory information for gradient navigation in Drosophila larvae

Foraging in naturalistic environments is often challenging, as animals must evaluate varying sensory cues to locate optimal food sources. To successfully navigate a taste gradient, where perceived concentrations change over space and time, animals need to compare previously encountered taste qualities with current information. How such short-term taste memories are implemented in the brain and used to guide navigation remains poorly understood. Due to their powerful genetic toolkit and whole-brain connectome, Drosophila larvae are an excellent model organism for investigating the neural basis underlying taste gradient navigation. Using linear fructose and salt gradients, we show that larvae are attracted to high fructose concentrations, whereas they avoid high salt concentrations. To identify the neural basis underlying taste gradient navigation, we tested the role of cell types from the peripheral chemosensory system to higher brain regions. Contrary to conclusions from simple two-choice preference assays, we show that gradient navigation depends on different cell types across multiple layers of chemosensory processing and associative learning circuits, including mushroom body neurons. Attractive and aversive taste signals are conveyed through parallel, partially overlapping pathways that converge onto mushroom body output neurons, where they are integrated to shape motor behaviors during chemotaxis.