Measuring the hidden dynamics of animal behavior


January 18, 2018 - 4:00pm
Northwest Building Room 243
About the Speaker
Gordon Berman
Speaker Title: 
Assistant Professor
Speaker Affiliation: 

When we think of animal behavior, what typically comes to mind are actions – running, swimming, grooming, resting, flying, singing, resting. Behavior, however, is more than the catalogue of motions that an organism can perform. Animals typically organize their repertoire of actions into sequences and patterns whose underlying dynamics last much longer than any particular behavior. How an organism modulates these hidden dynamics affects its success at accessing food, reproducing, and myriad other tasks essential for survival. In this talk, I will show how we are using unsupervised techniques for measuring behavior to infer these types of hidden state dynamics in drosophilds moving within a shallow arena. Specifically, I will show that fly behavior exhibits multiple long time scales, and calculating the optimally predictive representation of the fly's future behaviors reveals a hierarchical structure that is indicative of its underlying behavioral programs and its changing internal states. Lastly, I will show how accounting for these internal states allows us to gain insight into the evolution of fly behavior through reconstructing ancestral behavioral repertoires.