Solving navigational circuits in the Drosophila larva. Marc Gershow¹, Mason Klein¹, Marta Zlatic², Matthew Berck¹, Elizabeth Kane¹, Bruno Afonso¹, Aravinthan Samuel¹. 1) Center for Brain Science, Harvard University, Cambridge, MA; 2) HHMI Janelia Farm, Ashburn, VA.

   The transparent Drosophila larva, with a simple nervous system, robust behaviors, and powerful genetic tools, is an ideal model in which to relate neural structure and function. Navigation, which requires nontrivial computations to turn sensory input into motor output, is especially elucidating as the inputs can be precisely controlled and the outputs directly observed and quantified.
   We developed large-scale high resolution assays to analyze tens of thousands of navigational decisions made by hundreds of larvae in response to spatially and temporally varying temperatures, light intensities, and concentrations of odors and carbon dioxide. We used these assays to determine a common navigational strategy across sensory modalities. In response to unfavorable changes, larvae interrupt runs, periods of forward movement, and initiate a series of head sweeps. During a head sweep, a favorable change increases the probability that a larva will begin a new run in the direction of its head, while an unfavorable change increases the probability that the larva will instead execute a new head sweep in a different direction. The behavior depends on temporal variation in the sensory input, even when, e.g. for light incident at an angle, a direct comparison between bilateral sensory organs might be expected.
   We used our behavioral assays in an inactivation screen to identify specific neural populations involved in each navigational behavior and those implicated as part of a common navigational circuit. We recently combined our olfactory assay with optogenetic stimulation to generate a fully automated training apparatus to probe learning and developed a microfluidic apparatus to enable calcium imaging of the entire larval nervous system in concert with controlled odor presentation.