Brainwide dynamics of a social interaction

Communication is a fundamental aspect of life among animals, and yet we lack a comprehensive understanding of the underlying neural mechanisms. During communication, brains process sensory cues of different modalities, integrate those cues over time relative to changing internal states, and continually adjust ongoing actions. Here, we propose an ambitious, collaborative project to examine how the complete wiring diagram of a brain gives rise to brainwide patterns of activity that shape a social interaction, from sensation to action. In particular, we will leverage a novel connectomic resource and functional imaging methods uniquely available for Drosophila, combined with recent breakthroughs in deep learning and statistical methods for analyzing neural and behavioral data, to investigate how a female fly dynamically communicates with her social partner during courtship.

The Drosophila brain is compact, comprising only 140,000 neurons, yet it is capable of sophisticated behaviors. During courtship, males sing to, chase and circle females, while females integrate auditory and visual cues from the male to dynamically adjust locomotion. Female locomotion is a social behavior, because the movements shape the moment-to-moment patterning of male song. Our modeling of this behavior has revealed that it is also sculpted by fluctuating internal states of the brain, presenting an opportunity to leverage a quantitative understanding of the social interaction to investigate mechanisms underlying state-dependent sensorimotor transformations. Importantly, we are generating a whole-brain connectome for a female brain, which we will use here, in combination with brainwide cellular-resolution recordings from behaving females, to unify anatomy and function. Large-scale wiring diagrams of brains are emerging rapidly, but relating them to functional data and behavior, critical for interpreting these diagrams, remains a significant challenge. The relative stereotypy of the Drosophila brain makes it possible to compare functional activity recorded in one brain to connectivity information from another brain, but this has not been accomplished systematically at the scale of the brain yet. Doing so will allow us to build connectivity-constrained recurrent neural network models that capture dynamic activity and computations critical to behavior. Preliminary data suggest that the fly brain is densely recurrent, motivating the use of novel methods developed within our collaboration for understanding the dynamics of computation in this model system. Ultimately, these studies will reveal general principles that govern how brainwide connectivity and activity shape complex behavior, which will serve as a critical steppingstone for similar analyses in larger animals.