Neural Circuit Architectures for Olfactory Navigation
A basic task faced by every nervous system is allowing an animal to navigate its environment in search of food. For many organisms, turbulent odor plumes emitted by food are a principle cue used for this form of navigation. Within a plume, odor cues are stochastic and rarely signal the direction of the source directly. Theoretical studies thus suggest that animals should integrate odor information over time and store this information in short-term memory in order to navigate. The fruit fly Drosophila is a premier model for dissecting the neural circuit architectures supporting odor-guided navigation. Work from our lab and others suggests that the Central Complex, a conserved navigation center, is critical for this behavior. Distinct input pathways to the Central Complex encode the nature and valence of the odor, and the orientation and movement of the fly through space, respectively. In ongoing work, we have been imaging from local neurons of the Central Complex as flies perform a virtual closed-loop plume navigation task. We have identified a small group of local interneurons that respond robustly to odor but show persistent activity that can outlast the odor stimulus by several seconds. While activity in this population remains high, flies maintain the goal trajectory adopted during odor, suggesting that this activity represents a form of directional working memory. In a simulated turbulent plume, activity gradually ramps up, suggesting that these neurons integrate olfactory evidence over time. Ongoing experimental and computational work suggests that these dynamics arise from both recurrent ring connectivity with other local interneurons, and global inhibition from tangential neurons that span the array of local neurons. Together our work identifies neural architectures that allow the fly to build abstract central representations of variables needed for navigation, and to accumulate and store information about the likely direction of a goal. Going forward, we hope to compare theses architectures to those of organisms that navigate towards odors in both simpler and more complex environments, in order to understand how the structure of neural circuits supports efficient navigation in diverse physical environments.