Network-specific synchronization of electrical slow-wave oscillations regulates sleep in Drosophila

Abstract

Slow-wave rhythms characteristic of deep sleep oscillate in the delta band (0.5 -4 Hz) andcan be found across various brain regions in vertebrates. Across systems it is however unclear how oscillations arise and whether they are the causal functional unit steering behavior. Here, for the first time in any invertebrate, we discover sleep-relevant delta oscillations in Drosophila.We find that slow-wave oscillations in the sleep-regulating R2 network increase with sleep need. Optical multi-unit voltage recordings reveal that single R2 neurons get synchronized by sensory and circadian input pathways. We show that this synchronization depends on NMDA receptor (NMDARs) coincidence detector function and on an interplay of cholinergic and glutamatergic inputs setting a resonance frequency. Genetically targeting the coincidence detector function of NMDARs in R2, and thus the uncovered mechanism underlying synchronization, abolished network-specific slow-wave oscillations. It also disrupted sleep and facilitated light-induced wakening, directly establishing a causal role for slow-wave oscillations in regulating sleep and sensory gating. We therefore propose that the synchronization-based increase in oscillatory power likely represents an evolutionarily conserved, potentially 'optimal', strategy for constructing sleep-regulating sensory gates. that these delta oscillations depend on multi-unit synchronization mediated through NMDA receptor (NMDAR) coincidence detection. Disrupting this synchronization and thus the emergence of compound delta oscillations disrupts sleep and alters sensory gating during sleep. We thus identify slow-wave oscillations as an electrophysiological correlate for sleep regulation in invertebrates and place these oscillatory patterns at the basis of behavior. The sleep-regulating oscillations are comparable to sleep-regulating thalamic oscillations 20-22 as well as network-specific oscillations observed during sleep deprivation in vertebrates (local sleep) 2,23,24 . Our work demonstrates that slow-wave oscillations and sleep are fundamentally interconnected across systems, potentially representing an evolutionarily conserved strategy for network mechanisms regulating internal states and sleep. Results Sleep deprivation increases network-driven delta oscillations in the sleep-regulating R2 networkExamples of rhythmic activity patterns have previously been reported in insects 9,10 .However, their source, function and interdependence with internal states (such as sleep drive), remain largely unclear. We targeted expression of the GEVI ArcLight specifically to R2 neurons in the Drosophila brain. This defined network of 10 cells per hemisphere (Fig. 1a) resides within the ellipsoid body and is involved in sleep regulation [14][15][16] and multi-sensory relay [17][18][19] .In vivo recordings of the dendritic processes (bulb) of R2 neurons (Fig. 1a) identified electrical compound activity (Fig. 1b) that oscillated at delta-band frequencies between 0.5-1.5 Hz (Fig. 1b, d) in rested flies. We sleep-deprived f...

Integration of sleep homeostasis and navigation in Drosophila

“…For these ring neurons, ring attractor networks offer a compelling structure-function relationship which can provide a rationale for their ring-shaped structure. On the other hand, sleep related ring neurons serve as homeostatic sleep integrator, encoding sleep drive through structural 18,24 and activity changes 18,23 . To elucidate the relationship between these suggested navigation and sleep functionalities of ring neurons, we therefore asked what role the homeostatic integrator could play in the ring attractor framework.…”

“…The introduction of plasticity was motivated by the observation of structural, synaptic and functional changes in homeostatic ring neurons 18,23,24 as well as their likely connectivity to the head direction system 14,28 . In past work, plasticity has been introduced in ring attractor models to circumvent the problem of fine tuning [34][35][36] , in contrast to the role of plasticity proposed here, which is to maintain bump activity at a setpoint.…”

“…a concentric ring in the ellipsoid body, a substructure of the central complex. There are different classes of ring neurons: the sleep homeostatic integrator was identified in 18 as R2 neurons or as R5 neurons in 23 . Whether these are the same as R2 neurons with orientation-tuned receptive fields described in 20 is unclear, although sleep-related R2 neurons also respond to visual input 26 .…”

“…The morphological similarity of the involved ring neurons and the spatial proximity of the circuits as well as the connectome 14 , suggest however that they interact. Given the observed structural changes in the sleep-related ring neurons 18,23,24 , this suggests that the head direction system needs to operate in the face of substantial synaptic and functional changes in connected circuits.…”

“…Motivated by these experimental observations of plasticity 18,23,24 , we here use theoretical modeling to discuss how sleep homeostasis and navigation can be understood as a combined system (see Figure 1A for a schematic summary). We show that ring neurons can be assigned the role of integrating homeostatic sleep drive (referred to in the following as 'homeostatic integrator', see Figure 1B), while at the same time maintaining an operational head direction system (wedge neurons) in the face of plasticity.…”