Berman Lab @ Emory

Measuring Animal Behavior

How should we measure behavior? What are the numbers one should use to describe how an animal moves and interacts with the world? Our lab develops novel metrics and measurement techniques for studying behavior, leveraging ideas from nonlinear dynamics, statistical physics, information theory, computer vision, and machine learning. In particular, we have developed a set of techniques that take videography data as raw input and create a two-dimensional representation of behavior that can be used as the basis for many different analyses.

Neural Control of Behavior

In almost all animals, there is a bottleneck of information transmission bewteen the brain and the muscles that drive movement. By systematically activating neurons within this bottleneck optigenetically in Drosophila, our collaborators and we can attempt to decode how signals from the brain translate into movement, potentially gaining insight into the underlying aspects of behavioral organization and control. What is the encoding scheme that the nervous system is using within these ~400 neurons, and can it be viewed as being optimal in some sense?

Postural Dynamics and Control

Underlying all animal movement is the production and control of mechanical forces. In particular, much work has been performed studying how fast, periodic motions like running or flight are controlled or built-into the organism’s morphology, from which a rich mathematical framework has arisen that analyzes the strategies for controlling perturbations. Using postural measurement techniques developed in the lab that can find many thousands of instances of a species performing a particular set of repeated movements, we aim to test some of these theories, asking questions about the nature of stability and control in these systems.

Temporal Organization of Behavior

Behaving animals perform a complex sequence of actions across multiple temporal scales, with patterns ranging from locomotor activities occurring in less than a second to grooming patterns and courtship rituals over several minutes to circadian rhythms over the course of a day to even longer time-scale activities such as migration or child rearing. Our lab studies how animals organize these patterns of activity, using individual stereotyped behaviors as primative components. How does animal use these motions and create long time scales? What types of higher-order representations make this possible? Can these representations tell us something about how the brain organizes behavior?

Evolution of Behavior

Our understanding of how behaviors evolve has lagged far behind our knowledge of the mechanisms behind morphological evolution. Although a great deal of this is the result of the fact that there is little to no available fossil record of behavior, modern experimental approaches have also been limited by an inability to reliably and meaningfully measure behavior in evolutionarily-similar extant animals. Together with our collaborators, we apply the behavioral measurement techniques that we have developed to nearby species of Drosophila , attempting to reconstruct ancestral behavioral repertoires and providing hints as to the genetic and neurobioligcal mechanisms of evolution between these closely-related species.

Quantifying Vocalizations and Social Communication