Context, circuit and modulation of courtship signal selection in Drosophila

Abstract

Communication is multi-modal -- when we interact, we speak, gesticulate, and touch. However, the neural computations and circuits that select and coordinate these communication signals are unclear. We address this issue in Drosophila melanogaster which, thanks to its complex social behavior and genetic toolbox for manipulating neural activity, is ideal for dissecting the neural basis underlying communication. The courtship ritual, crucial for the survival of the species, is a social behavior which relies on successful communication between male and female flies. The male fly produces a range of signals based on feedback cues from the female to increase his chance of mating. When close to a female, he extends and vibrates one wing to generate air-borne signals, known as the courtship song, but also generates substrate-borne vibrations, providing an additional communication cue to her. Here, we study the specific behavioral context in which vibration signals are chosen over courtship song. On top of that, we propose a central circuit controlling courtship signal selection and test how the circuit's dynamics are modulated by internal state and sensory cues to support context-dependent signaling in social interactions. We developed an experimental setup to record the interactions of courting flies alongside the male song and vibrations. Statistical modeling of the flies' pose properties suggests that males vibrate in response to female immobility. We confirm this hypothesis by optogenetically controlling female walking during courtship. Asking how the selection of courtship signals is centrally controlled, we found two neuron clusters which control vibration and song production with complex dynamics: P1a and pC2l. While optogenetic activation of P1a neurons primarily drives persistent vibrations, pC2l neurons induce song production followed by vibrations. We propose an underlying circuit in which the two clusters drive the output of either signal type while cross inhibiting each other when active. A recurrent network acts presumably on P1a to maintain the persistent vibrations outlasting the neurons activity. This suggests that the two communication modes -- song and vibration -- are not only controlled by a cross-talk between different neuron clusters but also with distinct temporal dynamics, to provide persuasive communication output of the male. Lastly, we show that internal state and sensory cues modify the intrinsic circuit dynamics. Sexually satiated males and naive, highly motivated males differ in their internal state. A lower courtship drive reduces signal output of both clusters P1a and pC2l on intermediate time scales. Sensory cues provided by a receptive female quickly overwrite downstream effects of our circuit motif but also reveal that the mutual inhibition between both outputs is agnostic to social cues. With this, we provide insight into 1) how a social being such as the vinegar fly chooses for a communication signal type depending on its behavioral context and 2) how a neural circuit driving the output of different signal modalities is modulated by internal state and social cues to achieve context-dependent signal selection.