The Functional Role and Neural Basis of Rapid Turns in Olfactory Search Behavior - Summary When searching their environment for food, animals often switch from making random changes in direction to reflexively orient upstream when they smell an attractive odor, followed by rhythmic turns when they lose the plume. Random changes in direction are traditionally interpreted as manifestation of a stochastic search process. In this proposal, we test new hypothesis for the role of these random turns: animals turn deliberately to gather information before and after they change direction—allowing them to compute important internal and external state parameters such as wind direction, wind speed, ground speed, and altitude that they cannot directly measure. This idea is theoretically feasible based on the principle of observability, a powerful concept we borrow from control theory and informatics that is critical in the design of autonomous vehicles; here, we apply it to address fundamental questions in neuroscience. To do so, we exploit the natural behavior of flies, which execute rapid turns (called body saccades) as they search during flight. Our research program exploits state-of-the-art experimental techniques pioneered in the labs of team members, which we will leverage with genetic techniques and the rapidly emerging connectome databases available to the fly community—an integrated approach that has already provided our team with a solid foundation for our proposed effort. From wind tunnel experiments we have discovered that hungry flies execute a brief, stereotyped turn when they first experience an odor stimulus, which we postulate they use to estimate ambient wind direction and speed before either executing an upwind surge, or initiating a circling behavior in still air. From physiology experiments, we discovered a small network of identified cells that serve as command neurons for generating spontaneous saccades (DNa0X and DNb01) and another, specialized neuron that regulates their execution during flight (VS041). These preliminary results provide a strategic entry point for our proposed work, which is separated into four Specific Aims: (1) Test whether DNa0X and DNb01 are responsible for saccades that mediate olfactory search. (2) Identify upstream regulators of the saccade generating circuits responsible for the olfactory search behavior. (3) Determine the role of spontaneous saccades in gathering information. (4) Develop circuit-inspired, agent-based predictive models of olfactory search This effort to elucidate the circuitry that controls both stimulus-evoked and spontaneous turns during olfactory search and test their potential role in information-gathering links the ethologically important behavior of olfactory search across spatial and temporal scales—from the transient responses of individual neurons to the long flight trajectories of whole animals as they search for food. While not widely used in systems neuroscience, we believe that the concept of observability will provide a major conceptual advance in understanding how brains function; specifically, how the nervous system can collect information at different points in time and space and use it to compute parameters that are for critical subsequent behavioral actions.
