Coherent multi-level network oscillations create neural filters to favor quiescence over navigation in Drosophila

For all animals, undisturbed periods of rest are essential for undergoing recuperative processes. How neural interactions create brain states capable of dissociating an animal from its external world to promote quiescence remains a fundamental question. Here, we show how coherent network oscillations can create neural filters that favor a quiescent brain state over a state that promotes navigation. Circadian regulation and excitability of the Drosophila sleep homeostat (dFSB) generate nighttime specific slow-wave coherence between neural networks mediating sleep need (R5) and networks gating locomotion (helicon cells). Optogenetically mimicking coherent activity reveals that temporally fine-tuned R5 oscillations promote a quiescent state and reduce responsiveness to visual stimuli by hierarchically overruling locomotion-promoting helicon cells. We uncover that R5 and helicon bidirectionally regulate behavioral responsiveness by providing antagonistic inputs to head direction targets (EPG). Thus, coherent oscillations can form the mechanistic basis of neural filters by temporally associating antagonistic inputs and therefore reducing the functional connectivity between locomotion gating and navigational networks.