Using Drosophila Olfactory Navigation to Study Principles of Motor Encoding

Project Summary/Abstract Goal directed actions are often composed of shorter stochastic motor elements. How motor circuits are organized to translate a sensory-determined goal into a set of stochastic motor actions is unclear. Here I propose to use olfactory navigation behavior in the genetic model organism Drosophila melanogaster to

identify the circuitry and computational basis of motor control in a complex, goal-oriented task. Fly olfactory navigation is a highly robust behavior composed of shorter, stochastic motifs. Olfactory navigation involves three stages. At baseline, flies explore their environment in a stochastic fashion. When presented

with an appetitive odor, flies orient and run upwind. At odor offset, flies complete a search-like behavior, consisting of high angular velocity movements. Each phase has both reliable components (upwind orientation, increased angular velocity) and stochastic components (the precise timing of turns and runs). Our lab has

developed a high-throughput paradigm in which these three phases can be elicited repeatedly either though presentation of an attractive odor, or through presentation of a fictive optogenetic odor. The large datasets I can obtain with this paradigm are amenable to both computational and genetic analysis.

In my first Aim, I will perform a computational analysis of olfactory navigation behavior, identifying the timescales at which behavior is modulated following odor presentation or withdrawal, and decomposing fly trajectories into a series of behavioral motifs. Based on my motif analysis I will construct a Markovian model

that seeks to reproduce the complex statistics of navigation behavior, and to understand how the stochastic elements of navigation are concatenated to produce reliable goal-finding. In the second Aim, I will use genetic silencing and activation to identify descending neurons (DNs) that contribute to the behavior motifs and

temporal structure identified in the first Aim. DNs carry motor information from the brain to the ventral nerve cord, similar to neurons in the vertebrate that carry information from the brain to the spinal cord. This analysis will allow me to obtain a fairly complete circuit map of the motor circuitry the contributes to olfactory navigation.

Finally, in my third Aim, I will determine what features of sensory and motor information are encoded in the activity of particular DNs. Currently, two views of motor encoding exist in the fruit fly. Some studies support the notion that DNs relay motor information depending on behavioral context, while others suggest they encode for

specific movements, regardless of sensory driver. Olfactory navigation, composed of epochs of varying stimulus and behavioral goal, is poised to determine how movements of different sensory origin or behavioral context are encoded in motor circuitry. Using a closed loop behavioral apparatus, I will image from select DNs

during olfactory navigation and correlate activity with both behavioral motifs and navigational phase. Together, these experiments will help to uncover principles of motor encoding, which could aid in understanding how the brain is able to regain motor control after injury.