Predictability and Hierarchy in Animal Behavior

Soft Condensed Matter and Physics of Living Systems: Prof. Gordon Berman, Emory University

How an animal choses to order its activities -- moving, resting, grooming, and so on -- is essential to its ability to survive, adapt, and reproduce, with even the simplest of animals exhibiting behavioral sequences with complex temporal dynamics.  Far from being a disordered collection of actions, however, there is thought to be an intrinsic structure to an animal's  behavioral repertoire and its temporal and functional organization.  We would like to understand the organizing principles behind this structure, providing quantitative and predictive formalisms that generalize across individuals, species, and phyla.  An oft-stated proposal for this type of principle has been the idea of hierarchical organization, wherein the movements an animal makes can be understood as a set of nested sub-clusters and programs that can be repurposed depending on the context.  Although this type of organization holds potential advantages in terms of motion control and neural circuitry, measurements demonstrating this for an animal's entire behavioral repertoire have been limited in scope and temporal complexity.  Here, I will introduce a recently developed unsupervised technique to discover and track the occurrence of all  stereotyped behaviors performed by fruit flies moving in a shallow arena.  Calculating the optimally predictive representation of the fly's future behaviors, we see that fly behavior exhibits multiple time scales and is organized into a hierarchical structure that is indicative of its underlying behavioral programs and its changing internal states like hunger, circadian rhythm, and age.  Lastly, I will delve into analysis of some recent experiments where we further explore these ideas through optogenetic manipulation of descending command neurons in freely-behaving fruit flies.