Temporal drift of sleep-wake representations in hypothalamic neuronal ensembles

Temporal drift of sleep-wake representations in hypothalamic neuronal ensembles | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Temporal drift of sleep-wake representations in hypothalamic neuronal ensembles Antoine Adamantidis, Yudong Yan, Niccolo Calcini, Thomas Rusterholz, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6960125/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted You are reading this latest preprint version Abstract Correlative and causal evidence implicate distinct genetically-defined and evolutionary-conserved hypothalamic neurons in regulating wakefulness, non-rapid eye movement (NREM), and rapid eye movement (REM) sleep. The prevailing view is that these circuits govern sleep-wake states by recruiting stable, invariant neuronal substrates, yet, this remains unknown. Here, we showed that inhibitory, excitatory, hypocretins/orexins-, and melanin concentrating hormone-expressing-neurons in hypothalamus did not exhibit stable state-specific activities using longitudinal single cell calcium imaging in freely-moving sleeping mice. Instead, their activity patterns shift across sleep-wake states over time, while the distribution of active neurons in each sleep state remained stable. While sleep deprivation minimally affected the selectivity of these activity patterns, we found that the sleep-promoting drug diazepam recruited NREM sleep-active cells that were previously inactive or wake-active. These findings indicate that while individual neurons exhibit dynamic, state-dependent shifts of their activity, the overall organization of sleep-wake neural populations remains stable. Biological sciences/Systems biology/Single-cell imaging Biological sciences/Neuroscience/Circadian rhythms and sleep/Sleep Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction The hypothalamus is a core brain structure involved in maintaining physiological functions including food intake ( 1 , 2 ), reproduction ( 3 , 4 ), fight-or-flight responses ( 5 , 6 ) and sleep-wake states ( 7 – 9 ). The lateral hypothalamic area (LH) encompasses diverse neuron populations characterized by distinct biological and chemical markers, including vesicular glutamate transporters 2 (Vglut2, a marker for glutamatergic neurons), γ-Aminobutyric acid (GABA, Vgat), as well as neuropeptides orexins/hypocretins (OX/Hcrt) and melanin-concentrating hormone (MCH) ( 10 ). Inhibitory LH-Vgat neurons and excitatory LH-Vglut2 neurons show high discharge rate during both wakefulness and REM sleep as compared to NREM sleep, and their activation promotes wakefulness ( 11 – 14 ). Similarly, the activity of LH-OX/Hcrt neurons, which co-release glutamate ( 13 , 15 , 16 ), correlate with sensory inputs, emotional responses, locomotion, and sleep-to-wake transitions ( 17 – 19 ), while their optogenetic activation causally induced awakening from both NREM and REM sleep ( 20 ). In contrast, the firing rate of LH-MCH neurons is highest during REM sleep and relatively low during both wakefulness and NREM sleep ( 21 ). Inhibitory neurons in the preoptic area (POA) are predominantly active during NREM, and to a lesser extent REM sleep ( 22 – 24 ), and have been implicated in the regulation of sleep homeostasis ( 25 , 26 ). The prevailing model is that hypothalamic neurons that govern sleep-wake states are clustered into wake-, NREMs- or REMs-active groups consistent with classical c-fos staining ( 23 , 27 ), electrophysiology ( 21 , 28 ), or population-based calcium activity recordings ( 29 , 30 ), suggesting that neurons within these functional clusters are highly stable across repeated sleep-wake transitions. To test this hypothesis, we conducted longitudinal in vivo single-cell calcium (Ca²⁺) imaging of genetically targeted wake- and sleep-promoting neuronal subpopulations in the LH and POA (LH-Vgat, LH-Vglut2, LH-Hcrt/OX, LH-MCH, and POA-Vgat) across successive sleep-wake state transitions in freely behaving mice. Results LH sub-populations show different sleep state-dependent activity. We first characterized the single-cell activity of GABAergic (LH-Vgat), glutamatergic (LH-Vglut2), orexins/hypocretins (LH-OX/Hcrt), and melanin-concentrating hormone (LH-MCH) neurons in the LH across sleep-wake states using longitudinal one-photon Ca²⁺ imaging combined with simultaneous electroencephalogram (EEG) and electromyogram (EMG) recordings in freely-behaving mice ( Fig. 1 , A and B; Extended data Fig. 1 and Fig. 2 ). To characterize the dynamics of single cell LH-Vgat neurons, we stereotactically injected a Cre-dependent AAV-Syn-flex-GCaMP6 virus into the LH of Vgat-Cre mice, resulting in specific expression in LH-Vgat neurons ( Fig. 1 A; and Extended data Fig. 1 A ) . In 3-day baseline sleep recordings, LH-Vgat neurons exhibited a significant increase of the frequency of Ca²⁺ event during REM sleep as compared to NREM sleep and wakefulness ( Fig. 1 , C and E) . We classified LH-Vgat neurons into wake-max, NREM-max, and REM-max populations based on their peak activity during each vigilance state. The activity of each population prior to state transitions confirmed the selectivity of the clustering method (Extended data Fig. 3 ) . LH-Vgat neurons were preferentially activated during either REM sleep (30%) or wakefulness (45%), while a minor subpopulation of cells remained active during NREM sleep (8%) or showed no activity modulation across states (17%; Fig. 1 D ) . These activity patterns were consistent across each of the baseline recording sessions ( Fig. 1 F ) . We next examined the activity of LH-Vglut2 neurons by expressing Cre-dependent AAV-Syn-flex-GCaMP6 virus in the LH of Vglut2-Cre mice ( Fig. 1 , G and H; and Extended data Fig. 1 A ) . We found that the majority of neurons were preferentially active during wakefulness (59%) or REM (21%) sleep, while a small subset of cells was active during NREM sleep (6%) or inactive (14%) during baseline sleep ( Fig. 1 I ) . Similar to LH-Vgat neurons, LH-Vglut2 neurons showed the highest frequency of Ca² + events during REM sleep as compared to NREM sleep and wakefulness. Their activity during wakefulness was greater than during NREM sleep, though lower than during REM sleep ( Fig. 1 , G and J) . This pattern of neuronal activity was consistent across all baseline recording sessions ( Fig. 1 K ) . The expression of GCaMP6s in LH-OX/Hcrt neurons was achieved by stereotactic injection of AAV1/2-hORX-GCaMP6s-hGHp(A) into the LH of C57BL/6 mice ( Fig. 1 , L and M; and Extended data Fig. 1 B ) . The majority of LH-OX/Hcrt neurons were preferentially active during wakefulness (60%) while few remained active during REM sleep (18%; Extended data Fig. 3 , C, F and H ) or NREM sleep (3%; Fig. 1 N ) . The frequency of Ca²⁺ event in LH-OX/Hcrt neurons was similar during wakefulness and REM sleep, and both were significantly higher than during NREM sleep ( Fig. 1 , L and O) . This activity pattern was consistent across all baseline sessions ( Fig. 1 P ) . To image the activity of LH-MCH neurons, we stereotactically injected AAV-Syn-flex-GCaMP6 in the LH of MCH-Cre mice ( Fig. 1 , Q and R; and Extended data Fig. 1 B ) . Our results showed that a large subset of MCH neurons were predominantly active during REM sleep (40.5%), while others were maximally active during wakefulness (36.5%) and to a lesser extent NREM sleep (7.5%; Fig. 1 S ) . LH-MCH neurons displayed significantly elevated Ca²⁺ event frequencies at the transition to, and during REM sleep (Extended data Fig. 3 D ) , while event frequencies remained very low during both wakefulness and NREM sleep across all baseline recording sessions ( Fig. 1 , Q, T and U) . Sleep state specificity of single neurons shifts over time. Although the overall event frequencies in each vigilant state within LH subpopulations remained stable over time, we investigated whether individual neurons maintained consistent state-specific activity or exhibited dynamic shifts. To address this, we assessed the stability and selectivity of single LH neuron activity within each cluster over 7-days longitudinal imaging ( Fig. 1 B and Extended data Fig. 2 ) . Interestingly, we observed that 58.0% of wake-active, 14.3% of NREM-active, and 57.9% of REM-active neurons remained in the same cluster between the first and second imaging sessions, however, a large majority of individual LH-Vgat neurons did not. Instead, their peak activity shifted to a different cluster between recording sessions, indicating a dynamic reorganization of state-dependent single cell dynamics over time ( Fig. 2 A; and movie S1) . This was also true for the LH-Vglut2, LH-OX/Hcrt and LH-MCH neurons ( Fig. 2 , B to D) . This shift was further confirmed by quantifying the state selectivity of each neuron, defined by their Ca²⁺ event rates across all vigilance states (Extended data Fig. 4 ; and movie S2) . Strikingly, a survival analysis revealed that only a small fraction of neurons (4/509 in LH-Vgat, 1/257 in LH-Vglut2, 6/425 in LH-OX/Hcrt and 0/28 in LH-MCH) retained their initial state-dependent activity after 7 days of recording, suggesting a rapid reorganization of state-specific dynamics across all LH cell types recorded ( Fig. 2 , E to H) . We found that each neuron shifted its predominant state-specific activity an average of 10 times across the 21 recording sessions (LH-Vgat: 11.06, LH-Vglut2: 9.13, LH-OX/Hcrt: 9.53 and LH-MCH: 10.30; Fig. 2 , I to L ). Interestingly, we observed that neurons initially identified as wake-active, particularly those from the LH-Vgat, LH-Vglut2, and LH-OX/Hcrt populations, exhibited greater stability and underwent less frequent shifts in state-dependent activity compared to neurons classified as REM- or NREM sleep-active ( Fig. 2 , M to O) . In contrast, neurons amongst the LH-MCH population originally clustered as REM sleep-active showed a significantly higher stability as compared to wake or NREM sleep-active cells ( Fig. 2 P ) . This is consistent with the observation that cells that belong to large clusters tend to remain within this cluster for a longer period of time ( Fig. 2 , E to H) . Sleep state-dependent neuronal populations remain stable over time. Next, we examined whether the distribution of wake-, NREM-, and REM sleep-active neurons within each recorded LH neuronal type remained stable over time at the population level, despite the rapid shifts observed at the single-cell level. Strikingly, although individual neurons frequently shifted their predominant activity between wake, NREM, and REM states ( Fig. 3 A ) , we found that the overall distribution of LH-Vgat neurons that are preferentially active during wakefulness (44.3 ± 4.4%; mean ± SD), NREM (8.6 ± 3.2%), and REM sleep (27.4 ± 5.8%) remained remarkably stable across all 21 recorded sessions ( Fig. 3 E ). This population-level stability was also true for other LH subpopulations, including LH-Vglut2, LH-OX/Hcrt, and LH-MCH neurons ( Fig. 3 , B to D and F to H) , indicating that stable ensemble-level organization persists despite dynamic shifts in individual neuronal activity. To further characterize the structure and variability of these neuronal activity patterns across sleep-wake states, we analyzed their functional interconnectivity between hypothalamic cell clusters by calculating the betweenness centrality across different sleep states. Betweenness centrality is a metric used to assess the importance of a neuron within a network by measuring how often it acts as a ‘bridge’ between other neurons ( 31 ). In our analysis, hypothalamic neurons with higher betweenness centrality showed a stronger functional correlation with other neurons ( Fig. 3 I ) . A network with a high betweenness indicates a more interconnected and structured organization, where specific neurons play a key role in facilitating communication. Additionally, betweenness variation reflects the stability of these connections over time; shifts in overall network betweenness ( Fig. 3 I, states 1 to 2) or specific connectivity patterns ( Fig. 3 I, states 3 to 4) can both lead to increased variation, suggesting fluctuations in the hypothalamus network structure. When computed for LH-Vgat, LH-Vglut2, and LH-OX/Hcrt neuron activities across sleep-wake states, we found that betweenness centrality remained relatively stable across wake, NREM, and REM sleep for these hypothalamic networks ( Fig. 3 , J to L) . Notably, LH-Vglut2 neurons during wake and REM sleep, as well as LH-Vgat and LH-OX/Hcrt neurons during wake, exhibited sharper-tailed centrality, indicating the presence of highly ‘central’ neurons within a structured functional network. However, most neurons showed betweenness values that were not significantly different from chance ( Fig. 3 , J to L) , suggesting a limited degree of interconnectedness within these populations and an overall lack of persistent network structure across sleep-wake states in line with our experimental data. Furthermore, betweenness variation in LH-Vgat, LH-Vglut2, and LH-OX/Hcrt neuron populations also displays a uniform pattern across wake, NREM, and REM sleep states ( Fig. 3 , M to O) . In all three LH subpopulations, the mean betweenness variation exceeded chance level, reflecting dynamic shifts of state-dependent activity over time, regardless of the behavioral state. These findings are consistent with our experimental results ( Fig. 2 ) and further indicate a high degree of flexibility of state-dependent activity patterns across recurrent sleep-wake states. To determine whether the neuronal activity of LH subpopulations is affected by sleep homeostasis, we subjected animals to a 4-hour total sleep deprivation (SD) protocol (see Methods) . In vivo imaging of single-cell activity was performed during the subsequent sleep rebound period ( Fig. 1 B and Extended data Fig. 5 , A and B) . In LH-Vgat mice, SD led to increased NREM sleep duration and reduced wakefulness during the recovery period (Extended data Fig. 5 C ) , along with enhanced slow wave activity (SWA; 0.5–4 Hz), evidenced by a greater number of slow waves (Extended data Fig. 5 D ) and increased peak-to-peak amplitude (Extended data Fig. 5 E ) . Importantly, we found that the rate of Ca²⁺ transients recorded from individual LH-Vgat neurons across sleep-wake states remained largely unchanged following SD (Extended data Fig. 5 F ) . Additionally, the distribution of neurons that active across sleep-wake states remained largely unchanged (Extended data Fig. 5 G ) . This pattern was similarly observed in LH-Vglut2, LH-OX/Hcrt and LH-MCH neurons (Extended data Fig. 5 ) . Furthermore, the amplitude of Ca²⁺ transients remained unchanged after SD, except for a reduction of MCH neuron activity during NREM sleep (Extended data Fig. 5 H ) . These findings suggested that the neuronal activity of LH subpopulations remains robustly stable despite acute disruptions in sleep homeostasis. POA GABAergic neurons showed similar dynamics as LH neurons. We next tested whether similar single-cell and population dynamic apply to sleep-promoting neurons located in the anterior hypothalamus. To this end, we stereotactically injected AAV-Syn-flex-GCaMP6 into the preoptic area (POA) of Vgat-Cre (POA-Vgat) mice ( Fig. 4 A; and Extended data Fig. 1 C ) . Ca 2+ transients and EEG/EMG were monitored similarly as described in Fig. 1 B ( Fig. 4 B; see Methods) . Lateral POA-Vgat neurons were predominantly wake-active, with 53% of neurons displaying maximal event rate during wakefulness; 26.4% were selectively active during REM sleep, 8.6% during NREM sleep, and 12% remained inactive ( Fig. 4 C ) . These neurons displayed significantly higher Ca 2+ event frequency during REM sleep, while remaining low during wakefulness and NREM sleep—with wake-related activity being higher than that during NREM. ( Fig. 4 , D and E) . Similar to LH neurons, we found that lateral POA-Vgat neurons exhibited a rapid loss of initial state-selective activity, with only a small fraction of wake-selective neurons (12/442 cells) remaining in this cluster after seven days of recording ( Fig. 4 , F and G; and Extended data Fig. 4 ) . Similar to LH neurons, individual lateral POA-Vgat neurons shifted on average 9.7 times their state-selectivity over the 21 recording sessions ( Fig. 4 H ) . Despite this dynamic shift, the overall distribution of neuronal activity profile across sleep-wake states remained stable at the population level over time ( Fig. 4 , I and J) . These results suggested that the observed drift of neuronal state-selective activity is independent of the subregions and cell types in the hypothalamus. DZP increased POA-Vgat neuron activity and their selectivity for NREM sleep. To determine whether pharmacologically-induced changes in sleep architecture alter the state-dependent activity of lateral POA-Vgat neurons, we next recorded their response to Diazepam (DZP), a GABAergic sedative known to enhance NREM sleep ( 32 ). As expected, a single dose of intraperitoneal (i.p) DZP (5 mg/kg) at ZT12, when sleep pressure is minimal, extended the duration of NREM sleep at the expense of wakefulness ( Fig. 5 , A-C ) and significantly reduced the peak-to-peak amplitude of slow waves without altering their number, compared to saline-injected control conditions ( Fig. 5 , D and E) . Importantly, we found that DZP significantly increased the number of lateral POA-Vgat neurons activated during NREM sleep as compared to the control group, by either recruiting inactive neurons or switching a large part of the wake-active into NREM-active neurons ( Fig. 5 , G and H) . In addition, we observed that the average rate of Ca 2+ transients in lateral POA-Vgat neurons was significantly increased during NREM sleep upon DZP treatment ( Fig. 5 , F and I) and was suppressed during the transition from NREM sleep to wakefulness ( Fig. 5 J ) . Of note, DZP minimally altered NREM sleep, Ca 2+ transients or the distribution of wake, NREM and REM sleep active neurons when administered at ZT0 (Extended data Fig. 6) . Discussion These results indicate that while individual hypothalamic neurons exhibit dynamic, state-dependent shifts of their encoding of sleep-wake states, the overall organization of neural populations remains stable. This shift is similar to the ‘representational drift ( 33 ), described in highly plastic brain structures including the hippocampus, visual and auditory cortex, where individual neuron responses to specific stimuli or behavioural tasks fluctuate over time, while the overall population-level pattern remains consistent ( 34 – 38 ). Such mechanism is thought to support functional stability through redundancy and adaptability during learning or sensory perception ( 33 , 39 , 40 ). Our findings extend this mechanism to innate behavior, possibly to ensure the stability of sleep-wake states upon external or internal challenges. This is consistent with the relatively minor impact of experimental ablation of the hypothalamic system on sleep-wake architecture ( 41 – 43 ). This shift in the cellular substrates of sleep and wakefulness likely reflects the heterogeneous inputs to Vgat, Vglut2, OX/Hcrt and MCH neurons ( 44 – 46 )), as well as the molecular heterogeneity of hypothalamic neuronal populations recorded ( 13 , 30 , 47 ). Consistent with this, amassing studies suggested a higher functional heterogeneity amongst hypothalamic circuits, in particular LH-OX/Hcrt cells ( 30 ), than previously thought ( 17 , 48 ). Moreover, their activity prior to sleep onset may shape subsequent selective patterns of activity ( 49 ), while the sparse intrinsic connectivity among LH neurons ( 50 ) could further support this functional flexibility in sleep regulation. Importantly, the relatively rapid timescale of shifts in state-selective activity patterns suggested that these changes are not solely input-driven and may include synapse and dendritic spine plasticity ( 51 – 53 ), specific gene expression ( 54 , 55 ), protein translation and phosphorylation ( 56 , 57 ) or axonal bouton dynamics ( 58 , 59 ). While both sleep deprivation (SD) and diazepam (DZP) increased NREM sleep, only DZP significantly altered the cellular encoding of sleep through the recruitment of NREM-active cells and the shifting of the excitation-inhibition balance in hypothalamic circuits ( 32 , 60 ). The DZP effect was most prominent under low sleep pressure, a condition typical of insomnia, when endogenous sleep drive is low and sleep-promoting systems are presumably less active ( 61 , 62 ). Altogether, these findings highlight the delicate balance between genetically encoded stability (or rigidity) and functional flexibility in the regulation of sleep–wake states. Such flexibility likely supports the brain’s ability to maintain predictable and restorative sleep while adapting to changing internal and external demands ( 61 – 63 ). In this context, a flexible encoding strategy may allow sleep-regulating circuits to respond to changes, such as hormonal fluctuations, stress, temperature, or metabolic demands ( 64 , 65 ), and to support multifunctional roles such as thermoregulation and energy balance ( 66 – 68 ). On a broader scale, our observation of shifts in the cellular encoding of sleep-wake states suggested a network-level resilience in case of neuronal damage or dysfunction, that preserve essential homeostatic functions such as sleep ( 69 , 70 ). Methods Animal subjects C57Bl6JRj, Vgat-ires-Cre (Strain #016962), Vglut2-ires-Cre (Strain #:028863) and MCH-ires-Cre (Strain #014099) adult male mice, aged 8–24 weeks, were used in this study. Before surgeries, the mice were group-housed in ICV cages (Green Line, Techniplast), while post-surgery was single-housed in custom, open polyacryl cages (18 x 29 cm). In both conditions, animals stayed at constant temperature (22 ± 1°C) and humidity (30–50%), under a circadian cycle of a 12-hour light-dark cycle with lights on at 08:00. Food and water were available ad libitum . All procedures were conducted following protocols and guidelines approved by the Veterinary Office of the Canton of Bern, Switzerland (license number BE 129/2020 and BE122/2023). Viral targeting In this study, we utilized male mice, aged between 8 to 12 weeks, for in vivo calcium imaging to examine specific neuron sub-populations in the brain. The mice were initially anesthetized using 5% isoflurane and subsequently maintained at a level of 1.5-2% throughout the procedure. During the experiments, each mouse was fixed in a digital stereotactic frame, and body temperature was maintained at 37℃ using a feedback-coupled heating device (Panlab/Harvard Apparatus). Eye hydration was preserved with Bepanthen ointment (Bayer), and pain management was handled using 100 µl of lidocaine (1%) locally and a subcutaneous injection of Carprofen (10 mg/kg) administered 30 minutes before surgery. The preparation for surgery included shaving the hair and disinfecting the skin over the skull with 70% modified ethanol and betadine. Following a precise skin incision, alignment of anatomical landmarks (bregma and lambda) was performed. A small hole was then drilled using a pneumatic dental drill (Foredom, K1030 Portable) for targeted injections into the brain. Specific targeting involved unilateral injections into the lateral hypothalamus (LH) and the preoptic area (POA) to reach various neuronal types. To target GABAergic, glutamatergic, and MCH neurons in the LH, we administered unilateral injections of 0.6 µl of recombinant AAV5-Syn-flex-GCaMP6s (Addgene, 100845) into Vgat-Cre, Vglut2-Cre, and MCH-Cre mice, respectively. For targeting Orexins/Hypocretins neurons in the LH, 0.6 µl of AAV1/2-hORX-GCamp6s-hGHP(A) (VVF ETHZ) was injected into C57Bl6JRj mice. The coordinates used for LH were from Bregma anterior-posterior (AP): -1.35 mm, medial-lateral (ML): ±0.9 mm, dorso-ventral (DV): -5.30 mm. Additionally, to target GABAergic neurons in the preoptic area (POA), Vgat-Cre mice received 0.6 µl injections of AAV5-Syn-flex-GCaMP6s (Addgene, 100845) at AP: 0 mm, ML: +0.7 mm, DV: -5.2 mm. The viral infusions were performed at a maximum rate of 100 nl/min using a micro-infusion pump (PHD Ultra, Harvard Apparatus) through a 28 G stainless steel cannula (Plastic One). GRIN lens and EEG/EMG electrodes implantation Two weeks after virus injection, mice were anesthetized with an intraperitoneal (i.p.) injection of a mix containing medetomidine (0.5 mg kg-1), midazolam (5 mg kg-1) and fentanyl (0.05 mg kg-1) in sterile NaCl 0.9% (MMF-mix). Analgesia was achieved by local application of 100 µl of lidocaine (lurocaine, 1%) and subcutaneous (s.c.) injection of Carprofen (10 mg kg-1). Dexamethasone (0.5 mg.kg-1) was injected intramuscularly into the quadriceps to prevent potential inflammation that would be caused by the drilling and implantation. The animals were then shaved and placed in the stereotactic frame as described above. The skull bone was cleaned with 4% hydrogen peroxide and saline to remove remaining tissue. To ensure precise and controlled placement, the GRIN lens (diameter: 0.6 mm; length: 7.3 mm, Inscopix) was descended into the LH (-1.35mm AP; +0.9 mm ML; -5.35mm DV) or POA (0 mm AP; +0.7 mm ML; -5.25 mm DV) at 8 µm.s-1 for the first two-thirds of DV and then 0.8 µm.s-1 for the the final third with the help of stereotaxic motorized manipulator (Scientifica IVM Single). The lens was then fixed to the skull with a three-component dental cement (C&B-Metabond, Parkell Inc.). For polysomnographic recordings, four EEG electrodes made of stainless-steel screws were placed into the skull to record EEG signals (bilaterally into the frontal regions and parietal cortices, avoiding lens implanted region) and two EMG bare-ended wire loops were sutured to the trapezius muscle of the neck to record muscle activity. The EEG/EMG implants were then anchored to the skull with dental cement (Paladur, Patterson dental) and a cement plateau was built around the lens. The top of the GRIN lens was cleaned and covered with a plastic lid. To end the anesthesia, mice were injected subcutaneously with an antagonist mix (ABF-mix) containing atipamezole (2.5 mg.kg-1), buprenorphine (0.1 mg.kg-1) and flumazenil (0.5 mg.kg-1). After the surgery, mice underwent a recovery period with three days of subcutaneous administration of analgesic buprenorphine (0.1 mg.kg-1). After a minimum of two weeks of recovery, the animals were anesthetized again with isoflurane and placed in the stereotactic frame. Black nail polish was applied around the lens to prevent light propagation during the imaging. The optimal imaging focus plane for the micro-endoscope (nVista 3, Inscopix) was determined. The baseplate that attached to the micro-endoscope was fixed with UV glue and cement onto the existing cement layer, then black nail polish was applied around the cement again. The EEG/EMG cables were plugged into the EEG/EMG implants, taped with parafilm, and covered by aluminum foil. A tethered dummy micro-endoscope was connected to the baseplate for habituation to the recording device. After the procedures, the mice returned to their home cages. The cables from the dummy miniscope and EEG/EMG electrodes were securely taped to the top of the cage. This setup ensured that the animals could move freely within their environment. Calcium Imaging and Polysomnography Data Acquisition After an additional week of recovery and habituation to the cables, the mice were restrained to remove the microscope dummy and attach the real miniature microscope to the baseplate 1–2 days before the start of the experiment. Mice were recorded using a longitudinal protocol consisting of three days of baseline (BL1-3), 4 hours of sleep deprivation followed by acute sleep recovery (R0), and three days of chronic sleep recovery (R1-3) ( Fig. 1 B ) . Each day included three recording sessions, with each session lasting approximately 30 minutes. Before each recording session, mice were thoroughly acclimated to the experimenter and the recording environment to minimize stress. At the beginning of the longitudinal calcium imaging protocol, the focus, gain, and power settings of the microscope were carefully optimized to ensure clear visualization of neuronal activities. For each animal, the focus of the miniature microscope remained unchanged throughout the study, and the same field of view (FOV) was consistently imaged across all sessions to track the same neurons over time. Minor adjustments to the light power were occasionally made during later sessions to compensate for photo bleaching. The miniature microscope remained attached to the mouse throughout the entire recording period to ensure the stability of the FOV across all sessions. Calcium imaging data were collected at a rate of 10 frames per second using Inscopix Data Acquisition Software (IDAS) with typical excitation light intensities ranging from 0.6–1.2 mW/mm² (30–60% of maximum power, nVista 3, Inscopix). Simultaneously, EEG and EMG signals were recorded using an AM Systems 3500 amplifier, sampled at 512 Hz, and digitized with a National Instruments USB X DAQ device. To synchronize the calcium imaging data with the polysomnography recordings, binary timestamps for each frame from the miniscope were integrated with the AM systems. Animal behavior was recorded as a reference using a FireFly MV camera (Point Grey Research) at 15 frames per second. All EEG, EMG, behavioral video data, and synchronization signals were compiled and analyzed using SleepScore software (ViewPoint). During the baseline sleep phase (days 1–3) and the sleep deprivation recovery phase (days 5–7), the morning recording session began at around 10:00 AM (ZT2), followed by two additional sessions at 3:00 PM (ZT7) and 5:00 PM (ZT9). Sleep deprivation was carried out from 8:00 AM to 12:00 PM (ZT0-ZT4) for a total of 4 hours. The first recording session after sleep deprivation started as soon as the animal entered the first long NREM sleep bout, with subsequent sessions at 3:00 PM and 5:00 PM. Data acquisition was performed while the mice remained in their home cages. The experimenter entered the room at least 30 minutes before the start of each session to allow the mice to naturally settle. Recordings began once the animals reached a stable NREM sleep state. Due to the large amount of data streaming, individual imaging sessions were limited to approximately 30 minutes. Sleep deprivation Total sleep deprivation was performed manually using a gentle handling technique. Mice remained in their home cages with real-time EEG and EMG monitoring. When the experimenter observed signs of the mouse entering NREM sleep, gentle stimulus such as touching the tail or body with a cotton swab, offering enrichment materials for exploration, or adding fresh bedding to prompt nesting behavior—were used to wake the mouse. The aim was to keep the mouse awake without causing stress. This approach has been widely employed in sleep studies and has been shown to have minimal impact on stress hormone levels in a short period, particularly corticosterone, thereby minimizing confounding stress effects during the deprivation process ( 71 , 72 ). Polysomnography data scoring Three vigilance states—wakefulness, NREM sleep, and REM sleep—were identified using established criteria based on EEG and EMG characteristics, as outlined before ( 73 ). Briefly, wakefulness was characterized by low-amplitude, desynchronized EEG patterns and high-activity EMG signals featuring phasic bursts. NREM sleep was defined by synchronized EEG oscillations within the slow- and delta-frequency bands (0.5-4 Hz) that exhibited high amplitude and low frequency, coupled with significantly reduced EMG activity. REM sleep was identified by pronounced theta oscillations (6–9 Hz) in the EEG, or an increased presence of desynchronized, high-frequency oscillations, alongside an almost complete absence of EMG tone, except for brief muscle twitches. Micro-arousals during NREM sleep were determined by the presence of at least one second of cortical fast rhythms and EMG bursts. All recordings were analyzed at an epoch length of one second using the Matlab-based SlipAnalysis software (v.2.9.98, Lyon Neuroscience Research Center). The scoring data were compiled into a Hypnogram for further analysis, allowing alignment of neuronal activities with vigilance states. Automatic single slow wave detection Individual slow waves were detected during NREM sleep epochs based on the single EEG channel of each mouse using the SWA-MATLAB toolbox ( 74 ), with detection parameters adjusted for mice according to the settings described by ( 75 ). Slow waves were identified in various conditions, including baseline sleep, sleep recovery after sleep deprivation, and sleep following NaCl or DZP treatment. Briefly, the EEG data underwent initial processing to compute the negative envelope, which was then filtered within the 0.5 to 4 Hz frequency band using a Chebyshev Type II filter to isolate slow wave frequencies. Subsequent analysis involved consecutive zero-crossing detection, where transitions from negative to positive and vice versa were identified to mark potential starts and ends of slow waves. For a waveform segment to be classified as a slow wave, it had to meet two criteria: the duration between its negative-going zero-crossing and subsequent upward zero-crossing needed to be between 100 milliseconds and 1 second, and the peak negative amplitude had to exceed a threshold of three standard deviations above the median amplitude of all negative peaks recorded. This approach provides a structured process for identifying slow wave events based on specific amplitude and duration parameters, helping to mitigate potential individual differences due to electrode reference type, distance to those references, and electrode depth, which could affect the recorded amplitude. Calcium traces extraction Stacks of raw calcium imaging recordings were preprocessed to crop the redundant imaging area, correct time-invariant pixels, or reconstruct frames lost during the acquisition, and were then adjusted for motion artifacts in the x-y plane using Inscopix Data Processing software 1.9.1 (Inscopix). Subsequently, the image stacks were exported to Fiji (ImageJ), where the imaging background was subtracted. Regions of interest (ROIs) corresponding to individual neurons were manually defined by scrolling through all the recording sessions for each mouse, and the average calcium intensity in each ROI from each recording session was extracted and exported using Fiji software, providing a single time-series of raw fluorescence. Raw traces are further processed and normalized using custom-made scripts (Matlab, MathWorks) as described before ( 49 ). In brief, due to the potential bleaching of the GCaMP6 fluorescence within the session, an exponential curve was fitted and subtracted from the raw signal of each cell during each session to ensure consistency in the fluorescence measurements. Subsequently, F 0 was determined using Gaussian mixture modeling, where the observed intensity values from each cell in each session were approximated as a mixture of Gaussian distributions via an expectation-maximization procedure. The mean of the distribution containing the lowest values was identified as the F 0 for each neuron and session. ΔF/F 0 is then calculated based on bleaching-corrected calcium traces and estimated F 0 , and is normalized using envelope normalization. Normalized traces are referred to as ‘ΔF/F 0 ’ throughout the paper. Artifacts detection and removal The ΔF/F 0 traces were further pre-analyzed for the removal of any leftover movement artifacts. The measured raw EMG signal was used to determine sharp movements and ignore the corresponding frames. For this purpose, the EMG was thresholded on 5 times its mean + standard deviation, and the parts of the recordings that corresponded to peaks in absolute value higher than the forementioned threshold were marked as motion artifacts and ignored by the analysis. Furthermore, since sudden movements of the field of view with respect to the drawn ROI usually correspond to brief, sharp decreases in the ROI average intensities, causing sudden decreases in the signal, as a second step of motion artifact detection, the ROI traces were averaged to obtain an overall population signal. Then, a negative peaks detection using Matlab’s ‘findpeaks’ function was run on the average trace, and the frames corresponding to negative peaks, removed from the analysis. Ca 2+ event detections Calcium transient events (‘events’ in short) were detected using custom software made in our group, with Matlab 2021b. Each ROI ΔF/F 0 trace was analyzed separately after motion correction and artifacts detection. Traces were first upsampled via interpolation to be smoother with cubic spline interpolation (spline function in Matlab). The baseline median ‘M’ and noise level ‘E’ for each trace was estimated by computing the median and standard deviation of the signal in three iterations, the points exceeding 2 times the standard deviation were excluded from the estimation of M and E in the next iteration. Candidate calcium events were identified via the use of the findpeaks function in Matlab, using twice the estimated M + E as threshold for minimum peak height, and M + E as minimum peak prominence. A window of 6 seconds was considered before every identified peak and the maximum of the second derivative of the signal taken as the starting point of the candidate Calcium Event. The end point of the candidate event was taken as the point after the peak where the signal would return to the ‘local baseline’ value, computed as the average of the signal over the 6 seconds preceding the event start. The characteristics of the candidate Events, such as duration, peak amplitude, half-width, integral, duration of rise and decay phase were computed. Each measured characteristic was then compared with minimum and maximum physiological characteristics, and if not respecting these, excluded. The candidate events were then filtered and confirmed as Ca 2+ events if their measured characteristics were satisfying some minimum physiological parameters. Physiological parameters used for filtering candidate events were: Minimum Duration of an Event = 1 [seconds]; Maximum Duration of an Event = 20 [seconds]; Maximum Ratio between rise and decay duration = ⅓; Minimum Integral = 1 [a.u. (ΔF/F) * seconds]; Minimum Peak Amplitude = 2 [a.u. (ΔF/F)] for OX/Hcrt and VGlut2 neurons, or = 1 for VGAT neurons. State Selectivity and Stability Analysis The calcium traces were analyzed for each mouse and each session separately, to determine their state selectivity and their changes in different recording sessions and across multiple days. Calcium transients were identified as described above and separated into each different wake-sleep cycle state (awake, non-REM, and REM), according to where the calcium peak would occur. States with duration shorter than 20 seconds were removed from the analysis, as considered unstable. The frequency of calcium transient events (frequency of events) was then computed per cell, for each state, as the number of events happened during all stable states of a certain type, divided by the total duration of the stable states of that type. Cells with events’ frequency in the lower 25 percentile of all cells, in all three states, were considered separately as ‘inactive’ or ‘low active’ cells, and given the lack of activity, their state preference could not be estimated. All remaining cells were assigned as having state preference for the state where their frequency of events is maximal. Calcium traces integrals and integrals frequency Single-cell calcium traces were baseline corrected (baseline subtraction), and the integral of the entire trace was computed as the simple sum of all data points belonging to a specific state. The value obtained was divided by the duration of the state to obtain the ‘integral frequency’. Only states with a duration of 20 seconds or longer were considered to exclude unstable states. Network and Betweenness analysis The betweenness is a measure of the centrality of a node in a network, based on an estimation of how ‘disruptive’ to the network the removal of that specific node would be. To compute it, we first make a correlation-based network analysis between each calcium trace. As a pre-processing step for the correlation analysis, white noise, of the same amplitude as the estimated noise of the calcium traces, was summed to the raw calcium traces, to minimize the influence of noise correlations. State-specific traces were built by isolating all the stable states (longer than 20 seconds) of the same type, and gluing them together into three single traces, one for each state, ending up with 3 x number of cells traces. These state-specific traces were then correlated with each other over a sliding time window of length = 60 frames (6 seconds). Shuffled traces were used to compute a threshold for minimal significant correlation, for each state: this threshold was then used to weed out any non-significant correlation, setting it to zero. The remaining significant correlations were used to construct the upper triangular adjacency matrix for the undirected weighted graph, built via the “graph” Matlab function. Autoconnections, inactive cells, and cells that were isolated were removed from the graph and not considered in the rest of the analysis. At this point, we computed the undirected betweenness centrality C of a node u as ( 1 ) ( 1 ). C(u) = ∑ s,t≠u (n st (u) / N st ), where n st is the number of shortest paths from s to t that pass through node u, and N st is the number of shortest paths from s to t. Pharmacology treatment and recording Diazepam solution (DZP, Valium, 5 mg/ml) was prepared by diluting it in NaCl to a concentration of 2.5 mg/ml. To mitigate the stress associated with intraperitoneal injections, mice underwent a three-day acclimatization prior to the commencement of the experiment, during which they received 0.1 ml of NaCl intraperitoneally each day. This pre-treatment aimed to familiarize the mice with the handling and injection procedures. To explore the effects of DZP on sleep phases in mice, either 0.1 ml NaCl or DZP, at the dose of 5 mg/kg, was administered 10 minutes before the onset of the light phase (ZT0, 8 am) for three consecutive days. EEG and EMG recordings were then continuously taken for 12 hours following the initial administrations to monitor changes in sleep patterns. Additionally, calcium imaging targeting POA-Vgat neurons was performed for 30 minutes at 1-, 4-, and 8-hours post-administration each day to evaluate the neuronal activities response to the treatments. Following this phase, the animals were left undisturbed and handled minimally for at least 3 days to ensure complete metabolism of DZP before beginning treatments in the active phase. Subsequently, a single injection of either 0.1 ml NaCl or DZP (5 mg/kg) was administered 10 minutes before the start of the night phase (ZT12, 8 pm). EEG and EMG were again recorded for 12 hours to assess the sleep patterns post-treatment. Concurrently, calcium imaging was conducted 1-hour post-treatment for approximately 30 minutes to monitor the neuronal activities. Immunohistochemistry To assess viral vector expression and confirm accurate anatomical targeting of brain regions, brain slices were imaged after the conclusion of all experiments. Mice were deeply anesthetized using an intraperitoneal injection of pentobarbital (250 mg/kg, Streuli Pharma). Once fully anesthetized, transcardial perfusion was performed with 5–10 ml of cold saline (0.9% NaCl), followed by 20–25 ml of 4% formaldehyde (Grogg Chemie) to fix the tissues. After perfusion, the brains were removed and stored overnight in formaldehyde at 4°C for complete fixation. The following day, the brains were transferred to a phosphate-buffered saline (PBS) solution containing 30% sucrose and kept there until they sank, ensuring full cryoprotection before sectioning. The brains were sectioned into 40 µm slices using a cryostat (Hyrax C 25, Zeiss) and arranged in triplicate sets of 1:3 series, which were collected in PBS. To verify the co-expression of GCaMP6 with MCH or OX/Hcrt, sections were washed in PBST (PBS with 0.1% Triton X-100, Sigma-Aldrich) for 5 x 5 minutes and incubated for 45 minutes at room temperature in a blocking solution containing PBST and 4% bovine serum albumin (Sigma-Aldrich). The sections were then incubated with anti-pMCH (goat, 1:500, sc-14509, Santa Cruz Biotechnology) or anti-Orexin A (rabbit, 1:1000, H-003-30, Phoenix Pharmaceuticals) for 24–48 hours. After incubation, the sections were washed again for 5 x 5 minutes and then incubated for 2 hours at room temperature with AlexaFluor555-conjugated secondary antibody (1:1000 dilution, Invitrogen, A21432 for MCH and OX/Hcrt staining). The sections were then washed for 3 x 10 minutes and mounted on glass slides. Cover slips were applied using Fluoromount (F4680, Sigma-Aldrich). The GCaMP6s fluorescence was easily detected, and no additional staining was necessary. Slides were imaged using a Nikon-PLAN Fluor 10x/0.3NA objective on a Nikon Eclipse Ti-E fluorescence microscope controlled by Nikon NIS software. Excitation was achieved with a Solar LED lamp using Cy3 (mCherry) and FITC (eGFP) filters. For display purposes, image brightness and contrast were moderately adjusted in Fiji (ImageJ), and figures were compiled using Adobe Illustrator 2020. Statistical Analysis All data are presented as means ± standard error of the mean (SEM), with a significance level (α) of 0.05 for all two-sided tests. Statistical analysis was performed using MATLAB (MathWorks) and GraphPad Prism 10, focusing on graphed data. Significant results are detailed in the text, and sample sizes are provided in the corresponding figure legends. No formal power calculations were performed, but sample sizes were consistent with those used in similar studies. Data were compared using t-tests for parametric data, one-way repeated measures ANOVA with Tukey’s post-hoc correction for multiple comparisons, or two-way RM ANOVA with Sidak’s multiple comparisons test. While data distribution was assumed to be normal, this was not formally tested. Experimenters were not blinded to the conditions during data acquisition or analysis. Declarations Acknowledgments: We thank all the Tidis Laboratory members, M. Schmidt, M. Baud, and C. Bellone, B. Engelhardt, T. Korotkova for their insightful discussions. Funding : This work was supported by the Inselspital University Hospital Bern, the Swiss National Science Foundation (A.A. 310030_188761), the China Scholarship Council (Y.Y), and the University of Bern (A.A.). Author contributions : A.A. conceived the study. Y.Y performed the experiments. N.C developed the Ca 2+ imaging analysis tools. Y.Y, N.C. and T.R. analyzed the data. Y.Y. wrote the original draft and A.A, N.C, C.G.H reviewed & edited the manuscript. A.A. and C.G.H supervised the research. Competing interests : Authors declare that they have no competing interests. Data and materials availability : All presented data and analysis scripts, including mat-files and Matlab scripts and functions, are available on https://github.com/ZENLabCode . In addition, all data that supports the findings of this study are available from the corresponding authors upon request. References J. K. Elmquist, C. F. Elias, C. B. Saper, From lesions to leptin: hypothalamic control of food intake and body weight. Neuron 22 , 221–232 (1999). C. Roger, A. Lasbleiz, M. Guye, A. Dutour, B. Gaborit, J.-P. Ranjeva, The Role of the Human Hypothalamus in Food Intake Networks: An MRI Perspective. Front Nutr 8 , 760914 (2021). E. K. McIlwraith, D. D. Belsham, Hypothalamic reproductive neurons communicate through signal transduction to control reproduction. Mol Cell Endocrinol 518 , 110971 (2020). L. Mei, R. Yan, L. Yin, R. M. Sullivan, D. Lin, Antagonistic circuits mediating infanticide and maternal care in female mice. Nature 618 , 1006–1016 (2023). J. P. Herman, M. M. Ostrander, N. K. Mueller, H. Figueiredo, Limbic system mechanisms of stress regulation: hypothalamo-pituitary-adrenocortical axis. Prog Neuropsychopharmacol Biol Psychiatry 29 , 1201–1213 (2005). T. Kuwaki, Orexin (hypocretin) participates in central autonomic regulation during fight-or-flight response. Peptides 139 , 170530 (2021). C. Von Economo, Sleep as a problem of localization. Journal of Nervous and Mental Disease 71 , 249–259 (1930). C. B. Saper, T. E. Scammell, J. Lu, Hypothalamic regulation of sleep and circadian rhythms. Nature 437 , 1257–1263 (2005). A. R. Adamantidis, L. de Lecea, Sleep and the hypothalamus, Science . 282 (2023). https://doi.org/10.1126/science.adh8285. P. Bonnavion, L. E. Mickelsen, A. Fujita, L. de Lecea, A. C. Jackson, Hubs and spokes of the lateral hypothalamus: cell types, circuits and behaviour. J Physiol 594 , 6443–6462 (2016). O. K. Hassani, P. Henny, M. G. Lee, B. E. Jones, GABAergic neurons intermingled with orexin and MCH neurons in the lateral hypothalamus discharge maximally during sleep. Eur J Neurosci 32 , 448–457 (2010). A. Venner, C. Anaclet, R. Y. Broadhurst, C. B. Saper, P. M. Fuller, A Novel Population of Wake-Promoting GABAergic Neurons in the Ventral Lateral Hypothalamus. Curr Biol 26 , 2137–2143 (2016). L. E. Mickelsen, M. Bolisetty, B. R. Chimileski, A. Fujita, E. J. Beltrami, J. T. Costanzo, J. R. Naparstek, P. Robson, A. C. Jackson, Single-cell transcriptomic analysis of the lateral hypothalamic area reveals molecularly distinct populations of inhibitory and excitatory neurons. Nat Neurosci 22 , 642–656 (2019). R.-F. Wang, H. Guo, S.-Y. Jiang, Z.-L. Liu, W.-M. Qu, Z.-L. Huang, L. Wang, Control of wakefulness by lateral hypothalamic glutamatergic neurons in male mice. J Neurosci Res 99 , 1689–1703 (2021). P. Henny, F. Brischoux, L. Mainville, T. Stroh, B. E. Jones, Immunohistochemical evidence for synaptic release of glutamate from orexin terminals in the locus coeruleus. Neuroscience 169 , 1150–1157 (2010). C. Schöne, J. Apergis-Schoute, T. Sakurai, A. Adamantidis, D. Burdakov, Coreleased Orexin and Glutamate Evoke Nonredundant Spike Outputs and Computations in Histamine Neurons. Cell Reports 7 , 697–704 (2014). B. Y. Mileykovskiy, L. I. Kiyashchenko, J. M. Siegel, Behavioral correlates of activity in identified hypocretin/orexin neurons. Neuron 46 , 787–798 (2005). O. K. Hassani, M. R. Krause, L. Mainville, C. A. Cordova, B. E. Jones, Orexin Neurons Respond Differentially to Auditory Cues Associated with Appetitive versus Aversive Outcomes. The Journal of Neuroscience 36 , 1747 (2016). C. Concetti, D. Burdakov, Orexin/Hypocretin and MCH Neurons: Cognitive and Motor Roles Beyond Arousal. Front. Neurosci. 15 (2021). A. R. Adamantidis, F. Zhang, A. M. Aravanis, K. Deisseroth, L. de Lecea, Neural substrates of awakening probed with optogenetic control of hypocretin neurons. Nature 450 , 420–424 (2007). O. K. Hassani, M. G. Lee, B. E. Jones, Melanin-concentrating hormone neurons discharge in a reciprocal manner to orexin neurons across the sleep-wake cycle. Proc Natl Acad Sci U S A 106 , 2418–2422 (2009). N. Suntsova, R. Szymusiak, N. Md. Alam, R. Guzman-Marin, D. McGinty, Sleep-waking discharge patterns of median preoptic nucleus neurons in rats. The Journal of Physiology 543 , 665–677 (2002). H. Gong, D. McGinty, R. Guzman-Marin, K.-T. Chew, D. Stewart, R. Szymusiak, Activation of c-fos in GABAergic neurones in the preoptic area during sleep and in response to sleep deprivation. The Journal of Physiology 556 , 935–946 (2004). G. Miracca, B. Anuncibay-Soto, K. Tossell, R. Yustos, A. L. Vyssotski, N. P. Franks, W. Wisden, NMDA Receptors in the Lateral Preoptic Hypothalamus Are Essential for Sustaining NREM and REM Sleep. J Neurosci 42 , 5389–5409 (2022). M. A. Alam, S. Kumar, D. McGinty, M. N. Alam, R. Szymusiak, Neuronal activity in the preoptic hypothalamus during sleep deprivation and recovery sleep. J Neurophysiol 111 , 287–299 (2014). N. L. S. Machado, W. D. Todd, S. Kaur, C. B. Saper, Median preoptic GABA and glutamate neurons exert differential control over sleep behavior. Curr Biol 32 , 2011-2021.e3 (2022). L. Verret, R. Goutagny, P. Fort, L. Cagnon, D. Salvert, L. Léger, R. Boissard, P. Salin, C. Peyron, P.-H. Luppi, A role of melanin-concentrating hormone producing neurons in the central regulation of paradoxical sleep. BMC Neurosci 4 , 19 (2003). M. N. Alam, H. Gong, T. Alam, R. Jaganath, D. McGinty, R. Szymusiak, Sleep-waking discharge patterns of neurons recorded in the rat perifornical lateral hypothalamic area. J Physiol 538 , 619–631 (2002). J. C. Borniger, L. de Lecea, Peripheral Lipopolyssacharide Rapidly Silences REM-Active LHGABA Neurons. Front Behav Neurosci 15 , 649428 (2021). H. Ito, N. Fukatsu, S. M. Rahaman, Y. Mukai, S. Izawa, D. Ono, T. S. Kilduff, A. Yamanaka, Deficiency of orexin signaling during sleep is involved in abnormal REM sleep architecture in narcolepsy. Proceedings of the National Academy of Sciences of the United States of America 120 (2023). L. C. Freeman, A set of measures of centrality based on betweenness. Sociometry 40 , 35–41 (1977). L. E. McKillop, S. P. Fisher, L. Milinski, L. B. Krone, V. V. Vyazovskiy, Diazepam effects on local cortical neural activity during sleep in mice. Biochem Pharmacol 191 , 114515 (2021). L. N. Driscoll, L. Duncker, C. D. Harvey, Representational drift: Emerging theories for continual learning and experimental future directions. Curr Opin Neurobiol 76 , 102609 (2022). T. D. Marks, M. J. Goard, Stimulus-dependent representational drift in primary visual cortex. Nat Commun 12 , 5169 (2021). C. E. Schoonover, S. N. Ohashi, R. Axel, A. J. P. Fink, Representational drift in primary olfactory cortex. Nature 594 , 541–546 (2021). Y. Ziv, L. D. Burns, E. D. Cocker, E. O. Hamel, K. K. Ghosh, L. J. Kitch, A. E. Gamal, M. J. Schnitzer, Long-term dynamics of CA1 hippocampal place codes. Nat Neurosci 16 , 264–266 (2013). D. Deitch, A. Rubin, Y. Ziv, Representational drift in the mouse visual cortex. Curr Biol 31 , 4327-4339.e6 (2021). Z. N. Roth, E. P. Merriam, Representations in human primary visual cortex drift over time. Nat Commun 14 , 4422 (2023). D. J. Cai, D. Aharoni, T. Shuman, J. Shobe, J. Biane, W. Song, B. Wei, M. Veshkini, M. La-Vu, J. Lou, S. E. Flores, I. Kim, Y. Sano, M. Zhou, K. Baumgaertel, A. Lavi, M. Kamata, M. Tuszynski, M. Mayford, P. Golshani, A. J. Silva, A shared neural ensemble links distinct contextual memories encoded close in time. Nature 534 , 115–118 (2016). M. E. Rule, A. R. Loback, D. V. Raman, L. N. Driscoll, C. D. Harvey, T. O’Leary, Stable task information from an unstable neural population. Elife 9 , e51121 (2020). D. Gerashchenko, C. Blanco-Centurion, M. A. Greco, P. J. Shiromani, Effects of lateral hypothalamic lesion with the neurotoxin hypocretin-2–saporin on sleep in Long–Evans rats. Neuroscience 116 , 223–235 (2003). C. J. Hung, D. Ono, T. S. Kilduff, A. Yamanaka, Dual orexin and MCH neuron-ablated mice display severe sleep attacks and cataplexy. eLife 9 , e54275 (2020). D. Gerashchenko, T. C. Chou, C. A. Blanco-Centurion, C. B. Saper, P. J. Shiromani, Effects of lesions of the histaminergic tuberomammillary nucleus on spontaneous sleep in rats. Sleep 27 , 1275–1281 (2004). T. Sakurai, R. Nagata, A. Yamanaka, H. Kawamura, N. Tsujino, Y. Muraki, H. Kageyama, S. Kunita, S. Takahashi, K. Goto, Y. Koyama, S. Shioda, M. Yanagisawa, Input of orexin/hypocretin neurons revealed by a genetically encoded tracer in mice. Neuron 46 , 297–308 (2005). C. Schöne, D. Burdakov, Orexin/Hypocretin and Organizing Principles for a Diversity of Wake-Promoting Neurons in the Brain. Curr Top Behav Neurosci 33 , 51–74 (2017). J. A. González, P. Iordanidou, M. Strom, A. Adamantidis, D. Burdakov, Awake dynamics and brain-wide direct inputs of hypothalamic MCH and orexin networks. Nat Commun 7 , 11395 (2016). C. Blanco-Centurion, E. Bendell, B. Zou, Y. Sun, P. J. Shiromani, M. Liu, VGAT and VGLUT2 expression in MCH and orexin neurons in double transgenic reporter mice. IBRO Rep 4 , 44–49 (2018). M. G. Lee, O. K. Hassani, B. E. Jones, Discharge of identified orexin/hypocretin neurons across the sleep-waking cycle. J Neurosci 25 , 6716–6720 (2005). L. T. Oesch, M. Gazea, T. C. Gent, M. Bandarabadi, C. G. Herrera, A. R. Adamantidis, REM sleep stabilizes hypothalamic representation of feeding behavior. Proceedings of the National Academy of Sciences of the United States of America 117 (2020). D. Burdakov, M. M. Karnani, Ultra-sparse Connectivity within the Lateral Hypothalamus. Current Biology 30 , 4063-4070.e2 (2020). N. G. Hedrick, Z. Lu, E. Bushong, S. Singhi, P. Nguyen, Y. Magaña, S. Jilani, B. K. Lim, M. Ellisman, T. Komiyama, Learning binds new inputs into functional synaptic clusters via spinogenesis. Nat Neurosci 25 , 726–737 (2022). G. Yang, C. S. W. Lai, J. Cichon, L. Ma, W. Li, W.-B. Gan, Sleep promotes branch-specific formation of dendritic spines after learning. Science 344 , 1173–1178 (2014). M. N. Insanally, B. F. Albanna, J. Toth, B. DePasquale, S. S. Fadaei, T. Gupta, O. Lombardi, K. Kuchibhotla, K. Rajan, R. C. Froemke, Contributions of cortical neuron firing patterns, synaptic connectivity, and plasticity to task performance. Nat Commun 15 , 6023 (2024). G. Tononi, C. Cirelli, Sleep and the Price of Plasticity: From Synaptic and Cellular Homeostasis to Memory Consolidation and Integration. Neuron 81 , 12–34 (2014). Y. Vanrobaeys, Z. J. Peterson, E. N. Walsh, S. Chatterjee, L.-C. Lin, L. C. Lyons, T. Nickl-Jockschat, T. Abel, Spatial transcriptomics reveals unique gene expression changes in different brain regions after sleep deprivation. Nat Commun 14 , 7095 (2023). Y. Wang, S. Cao, D. Tone, H. Fujishima, R. G. Yamada, R.-I. Ohno, S. Shi, K. Matsuzawa, S. Yada, M. Kaneko, H. Sakamoto, T. Onishi, M. Ukai-Tadenuma, H. Ukai, C. Hanashima, K. Hirose, H. Kiyonari, K. Sumiyama, K. L. Ode, H. R. Ueda, Postsynaptic competition between calcineurin and PKA regulates mammalian sleep-wake cycles. Nature 636 , 412–421 (2024). F. Brüning, S. B. Noya, T. Bange, S. Koutsouli, J. D. Rudolph, S. K. Tyagarajan, J. Cox, M. Mann, S. A. Brown, M. S. Robles, Sleep-wake cycles drive daily dynamics of synaptic phosphorylation. Science 366 , eaav3617 (2019). L. De Vivo, M. Bellesi, W. Marshall, E. A. Bushong, M. H. Ellisman, G. Tononi, C. Cirelli, Ultrastructural evidence for synaptic scaling across the wake/sleep cycle. Science 355 , 507–510 (2017). X.-F. Du, F.-N. Li, X.-L. Peng, B. Xu, Y. Zhang, G. Li, T. Liu, Y. Li, H. Wang, J. Yan, J.-L. Du, Circadian regulation of developmental synaptogenesis via the hypocretinergic system. Nat Commun 14 , 3195 (2023). F. M. R. de Mendonça, G. P. R. R. de Mendonça, L. C. Souza, L. P. Galvão, H. S. Paiva, C. de Azevedo Marques Périco, J. Torales, A. Ventriglio, J. M. Castaldelli-Maia, A. Sousa Martins Silva, Benzodiazepines and Sleep Architecture: A Systematic Review. CNS Neurol Disord Drug Targets 22 , 172–179 (2023). J. Lim, D. F. Dinges, A meta-analysis of the impact of short-term sleep deprivation on cognitive variables. Psychol Bull 136 , 375–389 (2010). C. M. McDermott, G. J. LaHoste, C. Chen, A. Musto, N. G. Bazan, J. C. Magee, Sleep deprivation causes behavioral, synaptic, and membrane excitability alterations in hippocampal neurons. J Neurosci 23 , 9687–9695 (2003). D. P. Windred, A. C. Burns, J. M. Lane, R. Saxena, M. K. Rutter, S. W. Cain, A. J. K. Phillips, Sleep regularity is a stronger predictor of mortality risk than sleep duration: A prospective cohort study. Sleep 47 , zsad253 (2024). D. A. Kalmbach, J. R. Anderson, C. L. Drake, The impact of stress on sleep: Pathogenic sleep reactivity as a vulnerability to insomnia and circadian disorders. J Sleep Res 27 , e12710 (2018). S. J. Riede, V. van der Vinne, R. A. Hut, The flexible clock: predictive and reactive homeostasis, energy balance and the circadian regulation of sleep-wake timing. J Exp Biol 220 , 738–749 (2017). R. Rothhaas, S. Chung, Role of the Preoptic Area in Sleep and Thermoregulation. Frontiers in Neuroscience 15 , 664781 (2021). M. H. Schmidt, The energy allocation function of sleep: a unifying theory of sleep, torpor, and continuous wakefulness. Neurosci Biobehav Rev 47 , 122–153 (2014). H. Fong, J. Zheng, D. Kurrasch, The structural and functional complexity of the integrative hypothalamus. Science 382 , 388–394 (2023). A. Tauffenberger, P. J. Magistretti, Reactive Oxygen Species: Beyond Their Reactive Behavior. Neurochem Res 46 , 77–87 (2021). D. Zada, Y. Sela, N. Matosevich, A. Monsonego, T. Lerer-Goldshtein, Y. Nir, L. Appelbaum, Parp1 promotes sleep, which enhances DNA repair in neurons. Mol Cell 81 , 4979-4993.e7 (2021). F. Longordo, J. Fan, T. Steimer, C. Kopp, A. Lüthi, Do Mice Habituate to “Gentle Handling?” A Comparison of Resting Behavior, Corticosterone Levels and Synaptic Function in Handled and Undisturbed C57BL/6J Mice. Sleep 34 , 679–681 (2011). A. Lemons, R. M. Saré, C. Beebe Smith, Chronic Sleep Deprivation in Mouse Pups by Means of Gentle Handling. J Vis Exp , 58150 (2018). R. Boyce, S. D. Glasgow, S. Williams, A. Adamantidis, Sleep research: Causal evidence for the role of REM sleep theta rhythm in contextual memory consolidation. Science 352 , 812–816 (2016). A. Mensen, B. Riedner, G. Tononi, Optimizing detection and analysis of slow waves in sleep EEG. Journal of Neuroscience Methods 274 , 1–12 (2016). M. Panagiotou, V. V. Vyazovskiy, J. H. Meijer, T. Deboer, Differences in electroencephalographic non-rapid-eye movement sleep slow-wave characteristics between young and old mice. Sci Rep 7 , 43656 (2017). Additional Declarations There is NO Competing Interest. Supplementary Files ExtendedMoviemov1.avi Movie 1. Representation of LH-Vgat neuronal activity dynamics across vigilance states ExtendedMoviemov2.avi Movie 2. Representation of LH-Vgat neuronal selectivity dynamics across vigilance states Cite Share Download PDF Status: Under Review Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6960125","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":483707934,"identity":"57bbdb67-138a-4035-97c8-caaa73f141f3","order_by":0,"name":"Antoine Adamantidis","email":"data:image/png;base64,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","orcid":"https://orcid.org/0000-0003-2531-5175","institution":"Inselspital University Hospital Bern, University of Bern","correspondingAuthor":true,"prefix":"","firstName":"Antoine","middleName":"","lastName":"Adamantidis","suffix":""},{"id":483707935,"identity":"bec6eff5-dfa2-43f2-b6d1-6526a3e2a329","order_by":1,"name":"Yudong Yan","email":"","orcid":"","institution":"Bern University","correspondingAuthor":false,"prefix":"","firstName":"Yudong","middleName":"","lastName":"Yan","suffix":""},{"id":483707936,"identity":"fa3f22b6-e558-417c-bcaa-73e8225541c0","order_by":2,"name":"Niccolo Calcini","email":"","orcid":"","institution":"Department of Neurophysiology, Donders Institute for Brain, Cognition and Behaviour, Radboud University","correspondingAuthor":false,"prefix":"","firstName":"Niccolo","middleName":"","lastName":"Calcini","suffix":""},{"id":483707937,"identity":"a2abbd0b-7b78-4764-9d22-cbbdef8c9d98","order_by":3,"name":"Thomas Rusterholz","email":"","orcid":"","institution":"University Hospital of Bern","correspondingAuthor":false,"prefix":"","firstName":"Thomas","middleName":"","lastName":"Rusterholz","suffix":""},{"id":483707938,"identity":"ff80274b-92b4-49d7-a2bb-c9518e36ba52","order_by":4,"name":"Carolina Gutierrez Herrera","email":"","orcid":"https://orcid.org/0000-0002-7916-3341","institution":"Zentrum für Experimentelle Neurologie, Department of Neurology, Inselspital University Hospital Bern, University of Bern","correspondingAuthor":false,"prefix":"","firstName":"Carolina","middleName":"Gutierrez","lastName":"Herrera","suffix":""}],"badges":[],"createdAt":"2025-06-23 23:25:30","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6960125/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6960125/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":88414375,"identity":"0ba1e3be-e764-4d5b-a8ae-18f3516aa8f7","added_by":"auto","created_at":"2025-08-06 08:46:01","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":3598442,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eLH subgroups exhibit state-dependent activity across sleep-wake cycles in baseline recordings. (A)\u003c/strong\u003e a. The anatomical structure shows four dominant neuronal populations within the lateral hypothalamic area (LH), including glutamatergic, GABAergic, Orexins/Hypocretins (OX/Hcrt), and melanin-concentrating hormone (MCH) neurons. Note that GABAergic and glutamatergic neurons can co-express OX/Hcrt or MCH.\u0026nbsp; b. Representative field of view from calcium imaging of LH-Vgat neurons using a miniature microscope. Bright cellular structures and dark blood vessels are visible. Cells were manually identified, and single-cell activity was longitudinally recorded across multiple days. Scale bar: 50 µm. c. Example of expressing the Ca\u003csup\u003e2+\u003c/sup\u003e indicator GCaMP6s virus in GABAergic cells in Vgat-Cre mice. Following Cre-dependent local virus transfection, the GCaMP6s cassette is flipped in LH-Vgat neurons, allowing transcription and long-term expression in the LH. A GRIN lens was implanted at the same site for in vivo calcium imaging. Scale bar: 200 µm. The right panel shows the animal setup created in BioRender, including EEG/EMG electrodes and a miniature fluorescence microscope. \u003cstrong\u003e(B)\u003c/strong\u003e Timeline of experimental surgeries and longitudinal calcium imaging protocol during sleep-wake cycles. \u003cstrong\u003e(C)\u003c/strong\u003e Representative recording of EEG (squares show magnified traces of wake, NREM, and REM sleep), EMG, and Ca\u003csup\u003e2+\u003c/sup\u003e transients from GCaMP6-expressing LH-Vgat neurons across sleep-wake states. From top to bottom: EEG, EMG, Hypnogram, color-coded and raw single-cell ΔF/F0 traces from 78 GCaMP6s-expressing neurons. \u003cstrong\u003e(D)\u003c/strong\u003e Average percentages of LH-Vgat neurons (n = 509 neurons, m = 6 mice) that are inactive (grey) or have maximal Ca²⁺ event frequency during wake (blue), NREM (green), or REM (orange) sleep across 3-day baseline recordings. \u003cstrong\u003e(E)\u003c/strong\u003e Session Mean ± SEM Ca²⁺ event frequency (Hz) of LH-Vgat neurons during wake, NREM, and REM sleep of 3-day (9 sessions) baseline recordings. **P \u0026lt; 0.01; one-way ANOVA with Tukey post hoc test. \u003cstrong\u003e(F)\u003c/strong\u003e Mean ± SEM Ca²⁺ event frequency (Hz) of LH-Vgat neurons during wake, NREM, and REM sleep across 3-day (9 sessions) baseline recordings. \u003cstrong\u003e(G, L, Q)\u003c/strong\u003e Representative hypnogram, color-coded, and raw ΔF/F0 traces of LH-Vglut neurons (G), LH-OX/Hcrt neurons (L), and LH-MCH neurons (Q). \u003cstrong\u003e(H, M, R)\u003c/strong\u003e Fields of view from calcium imaging of LH-Vglut, LH-OX/Hcrt, and LH-MCH neurons recorded with miniature microscope after preprocessing. \u003cstrong\u003e(I, N, S)\u003c/strong\u003e Average percentages of LH-Vglut neurons (n = 257 neurons, m = 6 mice), LH-OX/Hcrt neurons (n = 425 neurons, m = 6 mice), and LH-MCH neurons (n = 28 neurons, m = 7 mice) that are inactive or have maximal activity during wake, NREM, or REM sleep, respectively. \u003cstrong\u003e(J, O, T) \u003c/strong\u003eSession Mean ± SEM Ca²⁺ event frequency (Hz) of LH-Vglut, LH-OX/Hcrt, and LH-MCH neurons during wake, NREM, and REM sleep. **P \u0026lt; 0.01; one-way ANOVA with Tukey post hoc test. \u003cstrong\u003e(K, P, U) \u003c/strong\u003eMean ± SEM Ca²⁺ event frequency (Hz) of LH-Vglut, LH-OX/Hcrt, and LH-MCH neurons during wake, NREM, and REM sleep across 3-day (9 sessions) baseline recordings.\u0026nbsp;\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-6960125/v1/0c14e9948fabb4c86cbd2036.png"},{"id":88414372,"identity":"90fd4e29-b7bb-4815-9e1a-db2d7d266528","added_by":"auto","created_at":"2025-08-06 08:46:01","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":979279,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSingle-neuron sleep state specificity decreases over time. (A-D)\u003c/strong\u003e Heatmaps showing changes in sleep state specificity of individual neurons over 7 days (21 recording sessions) in LH-Vgat, LH-Vglut2, LH-OX/Hcrt, and LH-MCH mice, respectively. Neurons are sorted based on their state preference in the first recording session, determined by their maximal event frequency during wake (blue), NREM (green), or REM (orange). The selectivity of these neurons is tracked over time. Neurons that did not exhibit sufficient activity in any of the three states (wake, NREM, or REM) were labeled as inactive (black). \u0026nbsp;\u003cstrong\u003e(E-H)\u003c/strong\u003e Survival curves showing the retention of sleep-state selectivity in wake, NREM, REM, or inactive neuron subclusters in LH-Vgat, LH-Vglut2, LH-OX/Hcrt, and LH-MCH mice, respectively. The number of neurons that retained their initial state-dependent classification (wake, NREM, REM, or inactive) was monitored across all 21 recording sessions. \u003cstrong\u003e(I-L) \u003c/strong\u003eDistribution of neurons based on the frequency of changes in their sleep-state selectivity across 21 experimental sessions in LH-Vgat, LH-Vglut2, LH-OX/Hcrt, and LH-MCH mice, respectively. Each bar represents the frequency of selectivity shifts, and the height of each bar corresponds to the number of neurons exhibiting that frequency. μ indicates the mean of the frequency, and σ² represents the variance. \u003cstrong\u003e(M-P)\u003c/strong\u003e Average frequencies of selectivity shift in inactive, wake, NREM, and REM neuronal clusters. Neurons are classified according to their state selectivity in the first recording session. *P \u0026lt; 0.05, **P \u0026lt; 0.01; one-way ANOVA with Tukey post hoc test.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-6960125/v1/d054d3960179c1a2bbae7c60.png"},{"id":88414371,"identity":"99cb7753-5d47-40da-9486-b64e79c1f189","added_by":"auto","created_at":"2025-08-06 08:46:01","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1821119,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSleep state-dependent neuron populations within the LH remain stable over time. (A-D)\u003c/strong\u003e River plots show the evolution of state selectivity in each neuronal group across baseline recordings in LH-Vgat, LH-Vglut2, LH-OX/Hcrt, and LH-MCH mice. Each column represents the proportion of inactive cells (dark gray) or active during wake (blue), NREM (green), or REM (orange) sleep. Light gray flows indicate selectivity shifts within neuron subclusters over time. \u003cstrong\u003e(E-H)\u003c/strong\u003e Aggregate percentage of cells in each state-selective cluster across all 21 recording sessions in LH-Vgat, LH-Vglut2, LH-OX/Hcrt, and LH-MCH mice, irrespective of individual neuron selectivity. \u003cstrong\u003e(I)\u003c/strong\u003e Schematic illustration of the concept of betweenness and betweenness variation. Betweenness centrality reflects how often a neuron acts as a bridge between other neurons within a network, representing its role in facilitating communication. Neurons with more and stronger connections, indicated by thicker lines. The color gradient shows the range of betweenness values from high to low. A neuron with betweenness = 0 is connected to fewer than two other neurons. Betweenness variation reflects the stability of these connections over time. At the population level, changes in the overall mean betweenness (e.g., from 1 to 2) or shifts in the distribution pattern of betweenness (e.g., from 3 to 4) can both contribute to increased betweenness variation, suggesting shifts in the network's structure or function. \u003cstrong\u003e(J-L)\u003c/strong\u003e Average normalized betweenness value of active neurons in wake, NREM, and REM state-specific networks, quantifying network correlation across sleep states in LH-Vgat, LH-Vglut2 and LH-OX/Hcrt mice. The solid line and dashed line represent the mean ± SEM of the expected chance level based on shuffled datasets. \u003cstrong\u003e(M-O)\u003c/strong\u003e Betweenness variation in wake, NREM, and REM state-specific networks, assessing network stability across sleep states in LH-Vgat, LH-Vglut2, and LH-OX/Hcrt mice. The solid and dashed lines represent the mean ± SEM of the expected chance levels, calculated using shuffled datasets.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-6960125/v1/af1956cc5105c72fceb9863b.png"},{"id":88415824,"identity":"6054c76b-10d2-4b68-a4cf-8784bbe29f70","added_by":"auto","created_at":"2025-08-06 08:54:01","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":2639991,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePOA-Vgat neurons exhibit similar stability patterns across longitudinal sleep-wake cycles. (A) \u003c/strong\u003eLeft: Expression of calcium indicator GCaMP6s virus in the preoptic area (POA) in Vgat-Cre mice and GRIN lens implantation in the POA. Scale bar: 1 mm. Right: Representative field of view from calcium imaging of POA-Vgat neurons using a miniature microscope. \u003cstrong\u003e(B)\u003c/strong\u003e Representative hypnogram, color-coded and raw ΔF/F0 traces of POA-Vgat mouse. \u003cstrong\u003e(C)\u003c/strong\u003e Average percentages of POA-Vgat neurons (n = 442 neurons, m = 6 mice) that are inactive (grey) or have maximal average event rate during wake (blue), NREM (green), or REM (orange) sleep across 3-day baseline recordings. \u003cstrong\u003e(D) \u003c/strong\u003eMean ± SEM Ca\u003csup\u003e2+\u003c/sup\u003e event frequency (Hz) of POA-Vgat neurons during wake, NREM, and REM sleep. **P \u0026lt; 0.01; one-way ANOVA with Tukey post hoc test. \u003cstrong\u003e(E) \u003c/strong\u003eMean ± SEM Ca\u003csup\u003e2+\u003c/sup\u003e event frequency (Hz) of POA-Vgat neurons during wake, NREM, and REM sleep across 3-day (9 sessions) baseline recordings. \u003cstrong\u003e(F)\u003c/strong\u003e Tracking of sleep state specificity of neurons over 7 days (21 sessions) in POA-Vgat mice. Neurons are sorted by their initial state preference (wake, NREM, REM), with inactive neurons labeled in black. \u003cstrong\u003e(G)\u003c/strong\u003e Survival curves depicting the retention of state selectivity (wake, NREM, REM, or inactive) across 21 sessions in POA-Vgat mice. \u003cstrong\u003e(H) \u003c/strong\u003eFrequency distribution of sleep-state selectivity shifts in POA-Vgat across 21 sessions. μ represents the mean frequency, and σ² indicates the variance. \u003cstrong\u003e(I) \u003c/strong\u003eThe evolution of state selectivity in POA-Vgat mice across baseline recordings. Each column represents the proportion of inactive and active neurons during wake, NREM, or REM. Light gray flows represent selectivity shifts over time. \u003cstrong\u003e(J)\u003c/strong\u003e Aggregate percentage of neurons in each state-selective cluster across all recording sessions, regardless of individual neuron selectivity, in POA-Vgat mice.\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-6960125/v1/359f253306b193ee499a128e.png"},{"id":88414373,"identity":"2834edf8-8fbd-4f04-a23e-efa872084e63","added_by":"auto","created_at":"2025-08-06 08:46:01","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":1188514,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDiazepam (DZP) injection at ZT12 increases POA-Vgat neuronal activity and selectivity during NREM sleep. (A)\u003c/strong\u003e Experimental schematic. Saline or DZP (5 mg/kg) is injected intraperitoneally at ZT12 (20:00), and calcium imaging recordings begin 1 hour post-injection. \u003cstrong\u003e(B)\u003c/strong\u003e Time course of wake, NREM, and REM sleep over 6 hours following saline or DZP treatment. \u003cstrong\u003e(C)\u003c/strong\u003e Cumulative duration of wake, NREM, and REM sleep within 6 hours post-saline or -DZP treatment. *P \u0026lt; 0.05, **P \u0026lt; 0.01; two-way ANOVA with Sidak's multiple comparisons test (m = 4 mice in NaCl, m = 3 mice in DZP). \u003cstrong\u003e(D) \u003c/strong\u003eAverage frequency of slow waves during NREM sleep in saline and DZP conditions (m = 5 mice). \u003cstrong\u003e(E) \u003c/strong\u003ePeak to peak amplitude of slow waves in saline and DZP conditions. **P \u0026lt; 0.01; paired t-test for repeated measures. \u003cstrong\u003e(F)\u003c/strong\u003e Examples of hypnograms and color-coded ΔF/F0 traces in saline and DZP-treated groups. \u003cstrong\u003e(G)\u003c/strong\u003e Composition and shifts of state-selective clusters following saline and DZP treatment (n = 365 neurons, m = 5 mice). \u003cstrong\u003e(H) \u003c/strong\u003eQuantification of the percentage of cells in each state-selective group. *P \u0026lt; 0.05; two-way ANOVA with Sidak's multiple comparisons test. \u003cstrong\u003e(I)\u003c/strong\u003e Average event frequency (Hz) in each sleep-wake state. *P \u0026lt; 0.05; two-way ANOVA with Sidak's multiple comparisons test. \u003cstrong\u003e(J)\u003c/strong\u003e Average Ca\u003csup\u003e2+\u003c/sup\u003e traces at state transitions: from NREM to wake and from wake to NREM (-10 to +10 seconds around the transition).\u0026nbsp;\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-6960125/v1/bff85c29deade7565db4e338.png"},{"id":88416700,"identity":"140cd272-fca9-48ee-a995-96d28882b205","added_by":"auto","created_at":"2025-08-06 09:02:08","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":13862928,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6960125/v1/74e0bbef-f948-415a-8fab-38f3628a99b3.pdf"},{"id":88414377,"identity":"55307220-5f91-43c6-81fd-d011b21ed706","added_by":"auto","created_at":"2025-08-06 08:46:02","extension":"avi","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":30306536,"visible":true,"origin":"","legend":"\u003cp\u003eMovie 1. Representation of LH-Vgat neuronal activity dynamics across vigilance states\u003c/p\u003e","description":"","filename":"ExtendedMoviemov1.avi","url":"https://assets-eu.researchsquare.com/files/rs-6960125/v1/a0c3289ae9e90fa4ddd309a3.avi"},{"id":88414376,"identity":"f5fa8807-fb5b-4cef-898f-b37579951e77","added_by":"auto","created_at":"2025-08-06 08:46:01","extension":"avi","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":30306536,"visible":true,"origin":"","legend":"\u003cp\u003eMovie 2. Representation of LH-Vgat neuronal selectivity dynamics across vigilance states\u003c/p\u003e","description":"","filename":"ExtendedMoviemov2.avi","url":"https://assets-eu.researchsquare.com/files/rs-6960125/v1/c2be912b38ca56a9fdda380c.avi"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Temporal drift of sleep-wake representations in hypothalamic neuronal ensembles","fulltext":[{"header":"Introduction","content":"\u003cp\u003eThe hypothalamus is a core brain structure involved in maintaining physiological functions including food intake (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e), reproduction (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e), fight-or-flight responses (\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e) and sleep-wake states (\u003cspan additionalcitationids=\"CR8\" citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e). The lateral hypothalamic area (LH) encompasses diverse neuron populations characterized by distinct biological and chemical markers, including vesicular glutamate transporters 2 (Vglut2, a marker for glutamatergic neurons), γ-Aminobutyric acid (GABA, Vgat), as well as neuropeptides orexins/hypocretins (OX/Hcrt) and melanin-concentrating hormone (MCH) (\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e). Inhibitory LH-Vgat neurons and excitatory LH-Vglut2 neurons show high discharge rate during both wakefulness and REM sleep as compared to NREM sleep, and their activation promotes wakefulness (\u003cspan additionalcitationids=\"CR12 CR13\" citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e). Similarly, the activity of LH-OX/Hcrt neurons, which co-release glutamate (\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e), correlate with sensory inputs, emotional responses, locomotion, and sleep-to-wake transitions (\u003cspan additionalcitationids=\"CR18\" citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e), while their optogenetic activation causally induced awakening from both NREM and REM sleep (\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e). In contrast, the firing rate of LH-MCH neurons is highest during REM sleep and relatively low during both wakefulness and NREM sleep (\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e). Inhibitory neurons in the preoptic area (POA) are predominantly active during NREM, and to a lesser extent REM sleep (\u003cspan additionalcitationids=\"CR23\" citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e), and have been implicated in the regulation of sleep homeostasis (\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThe prevailing model is that hypothalamic neurons that govern sleep-wake states are clustered into wake-, NREMs- or REMs-active groups consistent with classical c-fos staining (\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e), electrophysiology (\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e), or population-based calcium activity recordings (\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e, \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e), suggesting that neurons within these functional clusters are highly stable across repeated sleep-wake transitions. To test this hypothesis, we conducted longitudinal in vivo single-cell calcium (Ca\u0026sup2;⁺) imaging of genetically targeted wake- and sleep-promoting neuronal subpopulations in the LH and POA (LH-Vgat, LH-Vglut2, LH-Hcrt/OX, LH-MCH, and POA-Vgat) across successive sleep-wake state transitions in freely behaving mice.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cb\u003eLH sub-populations show different sleep state-dependent activity.\u003c/b\u003e\u003c/p\u003e\u003cp\u003eWe first characterized the single-cell activity of GABAergic (LH-Vgat), glutamatergic (LH-Vglut2), orexins/hypocretins (LH-OX/Hcrt), and melanin-concentrating hormone (LH-MCH) neurons in the LH across sleep-wake states using longitudinal one-photon Ca\u0026sup2;⁺ imaging combined with simultaneous electroencephalogram (EEG) and electromyogram (EMG) recordings in freely-behaving mice \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, A \u003cb\u003eand B; Extended data\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e \u003cb\u003eand\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e\u003cb\u003e).\u003c/b\u003e\u003c/p\u003e\u003cp\u003eTo characterize the dynamics of single cell LH-Vgat neurons, we stereotactically injected a Cre-dependent AAV-Syn-flex-GCaMP6 virus into the LH of Vgat-Cre mice, resulting in specific expression in LH-Vgat neurons \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA; \u003cb\u003eand Extended data\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA\u003cb\u003e)\u003c/b\u003e. In 3-day baseline sleep recordings, LH-Vgat neurons exhibited a significant increase of the frequency of Ca\u0026sup2;⁺ event during REM sleep as compared to NREM sleep and wakefulness \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, C \u003cb\u003eand E)\u003c/b\u003e. We classified LH-Vgat neurons into wake-max, NREM-max, and REM-max populations based on their peak activity during each vigilance state. The activity of each population prior to state transitions confirmed the selectivity of the clustering method \u003cb\u003e(Extended data\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e\u003cb\u003e)\u003c/b\u003e. LH-Vgat neurons were preferentially activated during either REM sleep (30%) or wakefulness (45%), while a minor subpopulation of cells remained active during NREM sleep (8%) or showed no activity modulation across states (17%; Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eD\u003cb\u003e)\u003c/b\u003e. These activity patterns were consistent across each of the baseline recording sessions \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eF\u003cb\u003e)\u003c/b\u003e.\u003c/p\u003e\u003cp\u003eWe next examined the activity of LH-Vglut2 neurons by expressing Cre-dependent AAV-Syn-flex-GCaMP6 virus in the LH of Vglut2-Cre mice \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, G \u003cb\u003eand H; and Extended data\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA\u003cb\u003e)\u003c/b\u003e. We found that the majority of neurons were preferentially active during wakefulness (59%) or REM (21%) sleep, while a small subset of cells was active during NREM sleep (6%) or inactive (14%) during baseline sleep \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eI\u003cb\u003e)\u003c/b\u003e. Similar to LH-Vgat neurons, LH-Vglut2 neurons showed the highest frequency of Ca\u0026sup2;\u003csup\u003e+\u003c/sup\u003e events during REM sleep as compared to NREM sleep and wakefulness. Their activity during wakefulness was greater than during NREM sleep, though lower than during REM sleep \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, G \u003cb\u003eand J)\u003c/b\u003e. This pattern of neuronal activity was consistent across all baseline recording sessions \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eK\u003cb\u003e)\u003c/b\u003e.\u003c/p\u003e\u003cp\u003eThe expression of GCaMP6s in LH-OX/Hcrt neurons was achieved by stereotactic injection of AAV1/2-hORX-GCaMP6s-hGHp(A) into the LH of C57BL/6 mice \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, L \u003cb\u003eand M; and Extended data\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB\u003cb\u003e)\u003c/b\u003e. The majority of LH-OX/Hcrt neurons were preferentially active during wakefulness (60%) while few remained active during REM sleep (18%; \u003cb\u003eExtended data\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, C, F \u003cb\u003eand H\u003c/b\u003e) or NREM sleep (3%; Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eN\u003cb\u003e)\u003c/b\u003e. The frequency of Ca\u0026sup2;⁺ event in LH-OX/Hcrt neurons was similar during wakefulness and REM sleep, and both were significantly higher than during NREM sleep \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, L \u003cb\u003eand O)\u003c/b\u003e. This activity pattern was consistent across all baseline sessions \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eP\u003cb\u003e)\u003c/b\u003e.\u003c/p\u003e\u003cp\u003eTo image the activity of LH-MCH neurons, we stereotactically injected AAV-Syn-flex-GCaMP6 in the LH of MCH-Cre mice \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, Q \u003cb\u003eand R; and Extended data\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB\u003cb\u003e)\u003c/b\u003e. Our results showed that a large subset of MCH neurons were predominantly active during REM sleep (40.5%), while others were maximally active during wakefulness (36.5%) and to a lesser extent NREM sleep (7.5%; Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eS\u003cb\u003e)\u003c/b\u003e. LH-MCH neurons displayed significantly elevated Ca\u0026sup2;⁺ event frequencies at the transition to, and during REM sleep \u003cb\u003e(Extended data\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eD\u003cb\u003e)\u003c/b\u003e, while event frequencies remained very low during both wakefulness and NREM sleep across all baseline recording sessions \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, Q, T \u003cb\u003eand U)\u003c/b\u003e.\u003c/p\u003e\u003cp\u003e\u003cb\u003eSleep state specificity of single neurons shifts over time.\u003c/b\u003e\u003c/p\u003e\u003cp\u003eAlthough the overall event frequencies in each vigilant state within LH subpopulations remained stable over time, we investigated whether individual neurons maintained consistent state-specific activity or exhibited dynamic shifts. To address this, we assessed the stability and selectivity of single LH neuron activity within each cluster over 7-days longitudinal imaging \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB \u003cb\u003eand Extended data\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e\u003cb\u003e)\u003c/b\u003e. Interestingly, we observed that 58.0% of wake-active, 14.3% of NREM-active, and 57.9% of REM-active neurons remained in the same cluster between the first and second imaging sessions, however, a large majority of individual LH-Vgat neurons did not. Instead, their peak activity shifted to a different cluster between recording sessions, indicating a dynamic reorganization of state-dependent single cell dynamics over time \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA; \u003cb\u003eand movie S1)\u003c/b\u003e. This was also true for the LH-Vglut2, LH-OX/Hcrt and LH-MCH neurons \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, B \u003cb\u003eto D)\u003c/b\u003e. This shift was further confirmed by quantifying the state selectivity of each neuron, defined by their Ca\u0026sup2;⁺ event rates across all vigilance states \u003cb\u003e(Extended data\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e; \u003cb\u003eand movie S2)\u003c/b\u003e. Strikingly, a survival analysis revealed that only a small fraction of neurons (4/509 in LH-Vgat, 1/257 in LH-Vglut2, 6/425 in LH-OX/Hcrt and 0/28 in LH-MCH) retained their initial state-dependent activity after 7 days of recording, suggesting a rapid reorganization of state-specific dynamics across all LH cell types recorded \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, E \u003cb\u003eto H)\u003c/b\u003e. We found that each neuron shifted its predominant state-specific activity an average of 10 times across the 21 recording sessions (LH-Vgat: 11.06, LH-Vglut2: 9.13, LH-OX/Hcrt: 9.53 and LH-MCH: 10.30; Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, I \u003cb\u003eto L\u003c/b\u003e). Interestingly, we observed that neurons initially identified as wake-active, particularly those from the LH-Vgat, LH-Vglut2, and LH-OX/Hcrt populations, exhibited greater stability and underwent less frequent shifts in state-dependent activity compared to neurons classified as REM- or NREM sleep-active \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, M \u003cb\u003eto O)\u003c/b\u003e. In contrast, neurons amongst the LH-MCH population originally clustered as REM sleep-active showed a significantly higher stability as compared to wake or NREM sleep-active cells \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eP\u003cb\u003e)\u003c/b\u003e. This is consistent with the observation that cells that belong to large clusters tend to remain within this cluster for a longer period of time \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, E \u003cb\u003eto H)\u003c/b\u003e.\u003c/p\u003e\u003cp\u003e\u003cb\u003eSleep state-dependent neuronal populations remain stable over time.\u003c/b\u003e\u003c/p\u003e\u003cp\u003eNext, we examined whether the distribution of wake-, NREM-, and REM sleep-active neurons within each recorded LH neuronal type remained stable over time at the population level, despite the rapid shifts observed at the single-cell level. Strikingly, although individual neurons frequently shifted their predominant activity between wake, NREM, and REM states \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA\u003cb\u003e)\u003c/b\u003e, we found that the overall distribution of LH-Vgat neurons that are preferentially active during wakefulness (44.3\u0026thinsp;\u0026plusmn;\u0026thinsp;4.4%; mean\u0026thinsp;\u0026plusmn;\u0026thinsp;SD), NREM (8.6\u0026thinsp;\u0026plusmn;\u0026thinsp;3.2%), and REM sleep (27.4\u0026thinsp;\u0026plusmn;\u0026thinsp;5.8%) remained remarkably stable across all 21 recorded sessions \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eE\u003cb\u003e).\u003c/b\u003e This population-level stability was also true for other LH subpopulations, including LH-Vglut2, LH-OX/Hcrt, and LH-MCH neurons \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, B \u003cb\u003eto D and F to H)\u003c/b\u003e, indicating that stable ensemble-level organization persists despite dynamic shifts in individual neuronal activity.\u003c/p\u003e\u003cp\u003eTo further characterize the structure and variability of these neuronal activity patterns across sleep-wake states, we analyzed their functional interconnectivity between hypothalamic cell clusters by calculating the betweenness centrality across different sleep states. Betweenness centrality is a metric used to assess the importance of a neuron within a network by measuring how often it acts as a \u0026lsquo;bridge\u0026rsquo; between other neurons (\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e). In our analysis, hypothalamic neurons with higher betweenness centrality showed a stronger functional correlation with other neurons \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eI\u003cb\u003e)\u003c/b\u003e. A network with a high betweenness indicates a more interconnected and structured organization, where specific neurons play a key role in facilitating communication. Additionally, betweenness variation reflects the stability of these connections over time; shifts in overall network betweenness \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eI, \u003cb\u003estates 1 to 2)\u003c/b\u003e or specific connectivity patterns \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eI, \u003cb\u003estates 3 to 4)\u003c/b\u003e can both lead to increased variation, suggesting fluctuations in the hypothalamus network structure.\u003c/p\u003e\u003cp\u003eWhen computed for LH-Vgat, LH-Vglut2, and LH-OX/Hcrt neuron activities across sleep-wake states, we found that betweenness centrality remained relatively stable across wake, NREM, and REM sleep for these hypothalamic networks \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, J \u003cb\u003eto L)\u003c/b\u003e. Notably, LH-Vglut2 neurons during wake and REM sleep, as well as LH-Vgat and LH-OX/Hcrt neurons during wake, exhibited sharper-tailed centrality, indicating the presence of highly \u0026lsquo;central\u0026rsquo; neurons within a structured functional network. However, most neurons showed betweenness values that were not significantly different from chance \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, J \u003cb\u003eto L)\u003c/b\u003e, suggesting a limited degree of interconnectedness within these populations and an overall lack of persistent network structure across sleep-wake states in line with our experimental data. Furthermore, betweenness variation in LH-Vgat, LH-Vglut2, and LH-OX/Hcrt neuron populations also displays a uniform pattern across wake, NREM, and REM sleep states \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, M \u003cb\u003eto O)\u003c/b\u003e. In all three LH subpopulations, the mean betweenness variation exceeded chance level, reflecting dynamic shifts of state-dependent activity over time, regardless of the behavioral state. These findings are consistent with our experimental results \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e\u003cb\u003e)\u003c/b\u003e and further indicate a high degree of flexibility of state-dependent activity patterns across recurrent sleep-wake states.\u003c/p\u003e\u003cp\u003eTo determine whether the neuronal activity of LH subpopulations is affected by sleep homeostasis, we subjected animals to a 4-hour total sleep deprivation (SD) protocol \u003cb\u003e(see Methods)\u003c/b\u003e. In vivo imaging of single-cell activity was performed during the subsequent sleep rebound period \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB \u003cb\u003eand Extended data\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e, A \u003cb\u003eand B)\u003c/b\u003e. In LH-Vgat mice, SD led to increased NREM sleep duration and reduced wakefulness during the recovery period \u003cb\u003e(Extended data\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eC\u003cb\u003e)\u003c/b\u003e, along with enhanced slow wave activity (SWA; 0.5\u0026ndash;4 Hz), evidenced by a greater number of slow waves \u003cb\u003e(Extended data\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eD\u003cb\u003e)\u003c/b\u003e and increased peak-to-peak amplitude \u003cb\u003e(Extended data\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eE\u003cb\u003e)\u003c/b\u003e. Importantly, we found that the rate of Ca\u0026sup2;⁺ transients recorded from individual LH-Vgat neurons across sleep-wake states remained largely unchanged following SD \u003cb\u003e(Extended data\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eF\u003cb\u003e)\u003c/b\u003e. Additionally, the distribution of neurons that active across sleep-wake states remained largely unchanged \u003cb\u003e(Extended data\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eG\u003cb\u003e)\u003c/b\u003e. This pattern was similarly observed in LH-Vglut2, LH-OX/Hcrt and LH-MCH neurons \u003cb\u003e(Extended data\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e\u003cb\u003e)\u003c/b\u003e. Furthermore, the amplitude of Ca\u0026sup2;⁺ transients remained unchanged after SD, except for a reduction of MCH neuron activity during NREM sleep \u003cb\u003e(Extended data\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eH\u003cb\u003e)\u003c/b\u003e. These findings suggested that the neuronal activity of LH subpopulations remains robustly stable despite acute disruptions in sleep homeostasis.\u003c/p\u003e\u003cp\u003e\u003cb\u003ePOA GABAergic neurons showed similar dynamics as LH neurons.\u003c/b\u003e\u003c/p\u003e\u003cp\u003eWe next tested whether similar single-cell and population dynamic apply to sleep-promoting neurons located in the anterior hypothalamus. To this end, we stereotactically injected AAV-Syn-flex-GCaMP6 into the preoptic area (POA) of Vgat-Cre (POA-Vgat) mice \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA; \u003cb\u003eand Extended data\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eC\u003cb\u003e)\u003c/b\u003e. Ca\u003csup\u003e2+\u003c/sup\u003e transients and EEG/EMG were monitored similarly as described in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB; \u003cb\u003esee Methods)\u003c/b\u003e. Lateral POA-Vgat neurons were predominantly wake-active, with 53% of neurons displaying maximal event rate during wakefulness; 26.4% were selectively active during REM sleep, 8.6% during NREM sleep, and 12% remained inactive \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC\u003cb\u003e)\u003c/b\u003e. These neurons displayed significantly higher Ca\u003csup\u003e2+\u003c/sup\u003e event frequency during REM sleep, while remaining low during wakefulness and NREM sleep\u0026mdash;with wake-related activity being higher than that during NREM. \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e, D \u003cb\u003eand E)\u003c/b\u003e. Similar to LH neurons, we found that lateral POA-Vgat neurons exhibited a rapid loss of initial state-selective activity, with only a small fraction of wake-selective neurons (12/442 cells) remaining in this cluster after seven days of recording \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e, F \u003cb\u003eand G; and Extended data\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e\u003cb\u003e)\u003c/b\u003e. Similar to LH neurons, individual lateral POA-Vgat neurons shifted on average 9.7 times their state-selectivity over the 21 recording sessions \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eH\u003cb\u003e)\u003c/b\u003e. Despite this dynamic shift, the overall distribution of neuronal activity profile across sleep-wake states remained stable at the population level over time \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e, I \u003cb\u003eand J)\u003c/b\u003e. These results suggested that the observed drift of neuronal state-selective activity is independent of the subregions and cell types in the hypothalamus.\u003c/p\u003e\u003cp\u003e\u003cb\u003eDZP increased POA-Vgat neuron activity and their selectivity for NREM sleep.\u003c/b\u003e\u003c/p\u003e\u003cp\u003eTo determine whether pharmacologically-induced changes in sleep architecture alter the state-dependent activity of lateral POA-Vgat neurons, we next recorded their response to Diazepam (DZP), a GABAergic sedative known to enhance NREM sleep (\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e). As expected, a single dose of intraperitoneal (i.p) DZP (5 mg/kg) at ZT12, when sleep pressure is minimal, extended the duration of NREM sleep at the expense of wakefulness \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e, A-C\u003cb\u003e)\u003c/b\u003e and significantly reduced the peak-to-peak amplitude of slow waves without altering their number, compared to saline-injected control conditions \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e, D \u003cb\u003eand E)\u003c/b\u003e. Importantly, we found that DZP significantly increased the number of lateral POA-Vgat neurons activated during NREM sleep as compared to the control group, by either recruiting inactive neurons or switching a large part of the wake-active into NREM-active neurons \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e, G \u003cb\u003eand H)\u003c/b\u003e. In addition, we observed that the average rate of Ca\u003csup\u003e2+\u003c/sup\u003e transients in lateral POA-Vgat neurons was significantly increased during NREM sleep upon DZP treatment \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e, F \u003cb\u003eand I)\u003c/b\u003e and was suppressed during the transition from NREM sleep to wakefulness \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eJ\u003cb\u003e)\u003c/b\u003e. Of note, DZP minimally altered NREM sleep, Ca\u003csup\u003e2+\u003c/sup\u003e transients or the distribution of wake, NREM and REM sleep active neurons when administered at ZT0 \u003cb\u003e(Extended data Fig.\u0026nbsp;6)\u003c/b\u003e.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eThese results indicate that while individual hypothalamic neurons exhibit dynamic, state-dependent shifts of their encoding of sleep-wake states, the overall organization of neural populations remains stable. This shift is similar to the ‘representational drift (\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e), described in highly plastic brain structures including the hippocampus, visual and auditory cortex, where individual neuron responses to specific stimuli or behavioural tasks fluctuate over time, while the overall population-level pattern remains consistent (\u003cspan additionalcitationids=\"CR35 CR36 CR37\" citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e–\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e). Such mechanism is thought to support functional stability through redundancy and adaptability during learning or sensory perception (\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e, \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e). Our findings extend this mechanism to innate behavior, possibly to ensure the stability of sleep-wake states upon external or internal challenges. This is consistent with the relatively minor impact of experimental ablation of the hypothalamic system on sleep-wake architecture (\u003cspan additionalcitationids=\"CR42\" citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e–\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThis shift in the cellular substrates of sleep and wakefulness likely reflects the heterogeneous inputs to Vgat, Vglut2, OX/Hcrt and MCH neurons (\u003cspan additionalcitationids=\"CR45\" citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e–\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e)), as well as the molecular heterogeneity of hypothalamic neuronal populations recorded (\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e, \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e). Consistent with this, amassing studies suggested a higher functional heterogeneity amongst hypothalamic circuits, in particular LH-OX/Hcrt cells (\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e), than previously thought (\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e, \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e). Moreover, their activity prior to sleep onset may shape subsequent selective patterns of activity (\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e), while the sparse intrinsic connectivity among LH neurons (\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e) could further support this functional flexibility in sleep regulation.\u003c/p\u003e\u003cp\u003eImportantly, the relatively rapid timescale of shifts in state-selective activity patterns suggested that these changes are not solely input-driven and may include synapse and dendritic spine plasticity (\u003cspan additionalcitationids=\"CR52\" citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e–\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e), specific gene expression (\u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e, \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e), protein translation and phosphorylation (\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e, \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e) or axonal bouton dynamics (\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e, \u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e59\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eWhile both sleep deprivation (SD) and diazepam (DZP) increased NREM sleep, only DZP significantly altered the cellular encoding of sleep through the recruitment of NREM-active cells and the shifting of the excitation-inhibition balance in hypothalamic circuits (\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e, \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e). The DZP effect was most prominent under low sleep pressure, a condition typical of insomnia, when endogenous sleep drive is low and sleep-promoting systems are presumably less active (\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e, \u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eAltogether, these findings highlight the delicate balance between genetically encoded stability (or rigidity) and functional flexibility in the regulation of sleep–wake states. Such flexibility likely supports the brain’s ability to maintain predictable and restorative sleep while adapting to changing internal and external demands (\u003cspan additionalcitationids=\"CR62\" citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e–\u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e63\u003c/span\u003e). In this context, a flexible encoding strategy may allow sleep-regulating circuits to respond to changes, such as hormonal fluctuations, stress, temperature, or metabolic demands (\u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e64\u003c/span\u003e, \u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e65\u003c/span\u003e), and to support multifunctional roles such as thermoregulation and energy balance (\u003cspan additionalcitationids=\"CR67\" citationid=\"CR66\" class=\"CitationRef\"\u003e66\u003c/span\u003e–\u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e68\u003c/span\u003e). On a broader scale, our observation of shifts in the cellular encoding of sleep-wake states suggested a network-level resilience in case of neuronal damage or dysfunction, that preserve essential homeostatic functions such as sleep (\u003cspan citationid=\"CR69\" class=\"CitationRef\"\u003e69\u003c/span\u003e, \u003cspan citationid=\"CR70\" class=\"CitationRef\"\u003e70\u003c/span\u003e).\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cb\u003eAnimal subjects\u003c/b\u003e\u003c/p\u003e\u003cp\u003eC57Bl6JRj, Vgat-ires-Cre (Strain #016962), Vglut2-ires-Cre (Strain #:028863) and MCH-ires-Cre (Strain #014099) adult male mice, aged 8–24 weeks, were used in this study. Before surgeries, the mice were group-housed in ICV cages (Green Line, Techniplast), while post-surgery was single-housed in custom, open polyacryl cages (18 x 29 cm). In both conditions, animals stayed at constant temperature (22 ± 1°C) and humidity (30–50%), under a circadian cycle of a 12-hour light-dark cycle with lights on at 08:00. Food and water were available \u003cem\u003ead libitum\u003c/em\u003e. All procedures were conducted following protocols and guidelines approved by the Veterinary Office of the Canton of Bern, Switzerland (license number BE 129/2020 and BE122/2023).\u003c/p\u003e\u003cp\u003e\u003cb\u003eViral targeting\u003c/b\u003e\u003c/p\u003e\u003cp\u003eIn this study, we utilized male mice, aged between 8 to 12 weeks, for \u003cem\u003ein vivo\u003c/em\u003e calcium imaging to examine specific neuron sub-populations in the brain. The mice were initially anesthetized using 5% isoflurane and subsequently maintained at a level of 1.5-2% throughout the procedure. During the experiments, each mouse was fixed in a digital stereotactic frame, and body temperature was maintained at 37℃ using a feedback-coupled heating device (Panlab/Harvard Apparatus). Eye hydration was preserved with Bepanthen ointment (Bayer), and pain management was handled using 100 µl of lidocaine (1%) locally and a subcutaneous injection of Carprofen (10 mg/kg) administered 30 minutes before surgery. The preparation for surgery included shaving the hair and disinfecting the skin over the skull with 70% modified ethanol and betadine. Following a precise skin incision, alignment of anatomical landmarks (bregma and lambda) was performed. A small hole was then drilled using a pneumatic dental drill (Foredom, K1030 Portable) for targeted injections into the brain.\u003c/p\u003e\u003cp\u003eSpecific targeting involved unilateral injections into the lateral hypothalamus (LH) and the preoptic area (POA) to reach various neuronal types. To target GABAergic, glutamatergic, and MCH neurons in the LH, we administered unilateral injections of 0.6 µl of recombinant AAV5-Syn-flex-GCaMP6s (Addgene, 100845) into Vgat-Cre, Vglut2-Cre, and MCH-Cre mice, respectively. For targeting Orexins/Hypocretins neurons in the LH, 0.6 µl of AAV1/2-hORX-GCamp6s-hGHP(A) (VVF ETHZ) was injected into C57Bl6JRj mice. The coordinates used for LH were from Bregma anterior-posterior (AP): -1.35 mm, medial-lateral (ML): ±0.9 mm, dorso-ventral (DV): -5.30 mm. Additionally, to target GABAergic neurons in the preoptic area (POA), Vgat-Cre mice received 0.6 µl injections of AAV5-Syn-flex-GCaMP6s (Addgene, 100845) at AP: 0 mm, ML: +0.7 mm, DV: -5.2 mm. The viral infusions were performed at a maximum rate of 100 nl/min using a micro-infusion pump (PHD Ultra, Harvard Apparatus) through a 28 G stainless steel cannula (Plastic One).\u003c/p\u003e\u003cp\u003e\u003cb\u003eGRIN lens and EEG/EMG electrodes implantation\u003c/b\u003e\u003c/p\u003e\u003cp\u003eTwo weeks after virus injection, mice were anesthetized with an intraperitoneal (i.p.) injection of a mix containing medetomidine (0.5 mg kg-1), midazolam (5 mg kg-1) and fentanyl (0.05 mg kg-1) in sterile NaCl 0.9% (MMF-mix). Analgesia was achieved by local application of 100 µl of lidocaine (lurocaine, 1%) and subcutaneous (s.c.) injection of Carprofen (10 mg kg-1). Dexamethasone (0.5 mg.kg-1) was injected intramuscularly into the quadriceps to prevent potential inflammation that would be caused by the drilling and implantation. The animals were then shaved and placed in the stereotactic frame as described above. The skull bone was cleaned with 4% hydrogen peroxide and saline to remove remaining tissue. To ensure precise and controlled placement, the GRIN lens (diameter: 0.6 mm; length: 7.3 mm, Inscopix) was descended into the LH (-1.35mm AP; +0.9 mm ML; -5.35mm DV) or POA (0 mm AP; +0.7 mm ML; -5.25 mm DV) at 8 µm.s-1 for the first two-thirds of DV and then 0.8 µm.s-1 for the the final third with the help of stereotaxic motorized manipulator (Scientifica IVM Single). The lens was then fixed to the skull with a three-component dental cement (C\u0026amp;B-Metabond, Parkell Inc.).\u003c/p\u003e\u003cp\u003eFor polysomnographic recordings, four EEG electrodes made of stainless-steel screws were placed into the skull to record EEG signals (bilaterally into the frontal regions and parietal cortices, avoiding lens implanted region) and two EMG bare-ended wire loops were sutured to the trapezius muscle of the neck to record muscle activity. The EEG/EMG implants were then anchored to the skull with dental cement (Paladur, Patterson dental) and a cement plateau was built around the lens. The top of the GRIN lens was cleaned and covered with a plastic lid. To end the anesthesia, mice were injected subcutaneously with an antagonist mix (ABF-mix) containing atipamezole (2.5 mg.kg-1), buprenorphine (0.1 mg.kg-1) and flumazenil (0.5 mg.kg-1). After the surgery, mice underwent a recovery period with three days of subcutaneous administration of analgesic buprenorphine (0.1 mg.kg-1).\u003c/p\u003e\u003cp\u003eAfter a minimum of two weeks of recovery, the animals were anesthetized again with isoflurane and placed in the stereotactic frame. Black nail polish was applied around the lens to prevent light propagation during the imaging. The optimal imaging focus plane for the micro-endoscope (nVista 3, Inscopix) was determined. The baseplate that attached to the micro-endoscope was fixed with UV glue and cement onto the existing cement layer, then black nail polish was applied around the cement again. The EEG/EMG cables were plugged into the EEG/EMG implants, taped with parafilm, and covered by aluminum foil. A tethered dummy micro-endoscope was connected to the baseplate for habituation to the recording device. After the procedures, the mice returned to their home cages. The cables from the dummy miniscope and EEG/EMG electrodes were securely taped to the top of the cage. This setup ensured that the animals could move freely within their environment.\u003c/p\u003e\u003cp\u003e\u003cb\u003eCalcium Imaging and Polysomnography Data Acquisition\u003c/b\u003e\u003c/p\u003e\u003cp\u003eAfter an additional week of recovery and habituation to the cables, the mice were restrained to remove the microscope dummy and attach the real miniature microscope to the baseplate 1–2 days before the start of the experiment. Mice were recorded using a longitudinal protocol consisting of three days of baseline (BL1-3), 4 hours of sleep deprivation followed by acute sleep recovery (R0), and three days of chronic sleep recovery (R1-3) \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB\u003cb\u003e)\u003c/b\u003e. Each day included three recording sessions, with each session lasting approximately 30 minutes. Before each recording session, mice were thoroughly acclimated to the experimenter and the recording environment to minimize stress.\u003c/p\u003e\u003cp\u003eAt the beginning of the longitudinal calcium imaging protocol, the focus, gain, and power settings of the microscope were carefully optimized to ensure clear visualization of neuronal activities. For each animal, the focus of the miniature microscope remained unchanged throughout the study, and the same field of view (FOV) was consistently imaged across all sessions to track the same neurons over time. Minor adjustments to the light power were occasionally made during later sessions to compensate for photo bleaching. The miniature microscope remained attached to the mouse throughout the entire recording period to ensure the stability of the FOV across all sessions.\u003c/p\u003e\u003cp\u003eCalcium imaging data were collected at a rate of 10 frames per second using Inscopix Data Acquisition Software (IDAS) with typical excitation light intensities ranging from 0.6–1.2 mW/mm² (30–60% of maximum power, nVista 3, Inscopix). Simultaneously, EEG and EMG signals were recorded using an AM Systems 3500 amplifier, sampled at 512 Hz, and digitized with a National Instruments USB X DAQ device. To synchronize the calcium imaging data with the polysomnography recordings, binary timestamps for each frame from the miniscope were integrated with the AM systems. Animal behavior was recorded as a reference using a FireFly MV camera (Point Grey Research) at 15 frames per second. All EEG, EMG, behavioral video data, and synchronization signals were compiled and analyzed using SleepScore software (ViewPoint).\u003c/p\u003e\u003cp\u003eDuring the baseline sleep phase (days 1–3) and the sleep deprivation recovery phase (days 5–7), the morning recording session began at around 10:00 AM (ZT2), followed by two additional sessions at 3:00 PM (ZT7) and 5:00 PM (ZT9). Sleep deprivation was carried out from 8:00 AM to 12:00 PM (ZT0-ZT4) for a total of 4 hours. The first recording session after sleep deprivation started as soon as the animal entered the first long NREM sleep bout, with subsequent sessions at 3:00 PM and 5:00 PM. Data acquisition was performed while the mice remained in their home cages. The experimenter entered the room at least 30 minutes before the start of each session to allow the mice to naturally settle. Recordings began once the animals reached a stable NREM sleep state. Due to the large amount of data streaming, individual imaging sessions were limited to approximately 30 minutes.\u003c/p\u003e\u003cp\u003e\u003cb\u003eSleep deprivation\u003c/b\u003e\u003c/p\u003e\u003cp\u003eTotal sleep deprivation was performed manually using a gentle handling technique. Mice remained in their home cages with real-time EEG and EMG monitoring. When the experimenter observed signs of the mouse entering NREM sleep, gentle stimulus such as touching the tail or body with a cotton swab, offering enrichment materials for exploration, or adding fresh bedding to prompt nesting behavior—were used to wake the mouse. The aim was to keep the mouse awake without causing stress. This approach has been widely employed in sleep studies and has been shown to have minimal impact on stress hormone levels in a short period, particularly corticosterone, thereby minimizing confounding stress effects during the deprivation process (\u003cspan citationid=\"CR71\" class=\"CitationRef\"\u003e71\u003c/span\u003e, \u003cspan citationid=\"CR72\" class=\"CitationRef\"\u003e72\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003cb\u003ePolysomnography data scoring\u003c/b\u003e\u003c/p\u003e\u003cp\u003eThree vigilance states—wakefulness, NREM sleep, and REM sleep—were identified using established criteria based on EEG and EMG characteristics, as outlined before (\u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e73\u003c/span\u003e). Briefly, wakefulness was characterized by low-amplitude, desynchronized EEG patterns and high-activity EMG signals featuring phasic bursts. NREM sleep was defined by synchronized EEG oscillations within the slow- and delta-frequency bands (0.5-4 Hz) that exhibited high amplitude and low frequency, coupled with significantly reduced EMG activity. REM sleep was identified by pronounced theta oscillations (6–9 Hz) in the EEG, or an increased presence of desynchronized, high-frequency oscillations, alongside an almost complete absence of EMG tone, except for brief muscle twitches. Micro-arousals during NREM sleep were determined by the presence of at least one second of cortical fast rhythms and EMG bursts. All recordings were analyzed at an epoch length of one second using the Matlab-based SlipAnalysis software (v.2.9.98, Lyon Neuroscience Research Center). The scoring data were compiled into a Hypnogram for further analysis, allowing alignment of neuronal activities with vigilance states.\u003c/p\u003e\u003cp\u003e\u003cb\u003eAutomatic single slow wave detection\u003c/b\u003e\u003c/p\u003e\u003cp\u003eIndividual slow waves were detected during NREM sleep epochs based on the single EEG channel of each mouse using the SWA-MATLAB toolbox (\u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e74\u003c/span\u003e), with detection parameters adjusted for mice according to the settings described by (\u003cspan citationid=\"CR75\" class=\"CitationRef\"\u003e75\u003c/span\u003e). Slow waves were identified in various conditions, including baseline sleep, sleep recovery after sleep deprivation, and sleep following NaCl or DZP treatment. Briefly, the EEG data underwent initial processing to compute the negative envelope, which was then filtered within the 0.5 to 4 Hz frequency band using a Chebyshev Type II filter to isolate slow wave frequencies. Subsequent analysis involved consecutive zero-crossing detection, where transitions from negative to positive and vice versa were identified to mark potential starts and ends of slow waves. For a waveform segment to be classified as a slow wave, it had to meet two criteria: the duration between its negative-going zero-crossing and subsequent upward zero-crossing needed to be between 100 milliseconds and 1 second, and the peak negative amplitude had to exceed a threshold of three standard deviations above the median amplitude of all negative peaks recorded. This approach provides a structured process for identifying slow wave events based on specific amplitude and duration parameters, helping to mitigate potential individual differences due to electrode reference type, distance to those references, and electrode depth, which could affect the recorded amplitude.\u003c/p\u003e\u003cp\u003e\u003cb\u003eCalcium traces extraction\u003c/b\u003e\u003c/p\u003e\u003cp\u003eStacks of raw calcium imaging recordings were preprocessed to crop the redundant imaging area, correct time-invariant pixels, or reconstruct frames lost during the acquisition, and were then adjusted for motion artifacts in the x-y plane using Inscopix Data Processing software 1.9.1 (Inscopix). Subsequently, the image stacks were exported to Fiji (ImageJ), where the imaging background was subtracted. Regions of interest (ROIs) corresponding to individual neurons were manually defined by scrolling through all the recording sessions for each mouse, and the average calcium intensity in each ROI from each recording session was extracted and exported using Fiji software, providing a single time-series of raw fluorescence.\u003c/p\u003e\u003cp\u003eRaw traces are further processed and normalized using custom-made scripts (Matlab, MathWorks) as described before (\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e). In brief, due to the potential bleaching of the GCaMP6 fluorescence within the session, an exponential curve was fitted and subtracted from the raw signal of each cell during each session to ensure consistency in the fluorescence measurements. Subsequently, F\u003csub\u003e0\u003c/sub\u003e was determined using Gaussian mixture modeling, where the observed intensity values from each cell in each session were approximated as a mixture of Gaussian distributions via an expectation-maximization procedure. The mean of the distribution containing the lowest values was identified as the F\u003csub\u003e0\u003c/sub\u003e for each neuron and session. ΔF/F\u003csub\u003e0\u003c/sub\u003e is then calculated based on bleaching-corrected calcium traces and estimated F\u003csub\u003e0\u003c/sub\u003e, and is normalized using envelope normalization. Normalized traces are referred to as ‘ΔF/F\u003csub\u003e0\u003c/sub\u003e’ throughout the paper.\u003c/p\u003e\u003cp\u003e\u003cb\u003eArtifacts detection and removal\u003c/b\u003e\u003c/p\u003e\u003cp\u003eThe ΔF/F\u003csub\u003e0\u003c/sub\u003e traces were further pre-analyzed for the removal of any leftover movement artifacts. The measured raw EMG signal was used to determine sharp movements and ignore the corresponding frames. For this purpose, the EMG was thresholded on 5 times its mean + standard deviation, and the parts of the recordings that corresponded to peaks in absolute value higher than the forementioned threshold were marked as motion artifacts and ignored by the analysis. Furthermore, since sudden movements of the field of view with respect to the drawn ROI usually correspond to brief, sharp decreases in the ROI average intensities, causing sudden decreases in the signal, as a second step of motion artifact detection, the ROI traces were averaged to obtain an overall population signal. Then, a negative peaks detection using Matlab’s ‘findpeaks’ function was run on the average trace, and the frames corresponding to negative peaks, removed from the analysis.\u003c/p\u003e\u003cp\u003e\u003cb\u003eCa\u003c/b\u003e\u003csup\u003e\u003cb\u003e2+\u003c/b\u003e\u003c/sup\u003e \u003cb\u003eevent detections\u003c/b\u003e\u003c/p\u003e\u003cp\u003eCalcium transient events (‘events’ in short) were detected using custom software made in our group, with Matlab 2021b. Each ROI ΔF/F\u003csub\u003e0\u003c/sub\u003e trace was analyzed separately after motion correction and artifacts detection. Traces were first upsampled via interpolation to be smoother with cubic spline interpolation (spline function in Matlab). The baseline median ‘M’ and noise level ‘E’ for each trace was estimated by computing the median and standard deviation of the signal in three iterations, the points exceeding 2 times the standard deviation were excluded from the estimation of M and E in the next iteration. Candidate calcium events were identified via the use of the findpeaks function in Matlab, using twice the estimated M + E as threshold for minimum peak height, and M + E as minimum peak prominence. A window of 6 seconds was considered before every identified peak and the maximum of the second derivative of the signal taken as the starting point of the candidate Calcium Event. The end point of the candidate event was taken as the point after the peak where the signal would return to the ‘local baseline’ value, computed as the average of the signal over the 6 seconds preceding the event start. The characteristics of the candidate Events, such as duration, peak amplitude, half-width, integral, duration of rise and decay phase were computed. Each measured characteristic was then compared with minimum and maximum physiological characteristics, and if not respecting these, excluded. The candidate events were then filtered and confirmed as Ca\u003csup\u003e2+\u003c/sup\u003e events if their measured characteristics were satisfying some minimum physiological parameters. Physiological parameters used for filtering candidate events were: Minimum Duration of an Event = 1 [seconds]; Maximum Duration of an Event = 20 [seconds]; Maximum Ratio between rise and decay duration = ⅓; Minimum Integral = 1 [a.u. (ΔF/F) * seconds]; Minimum Peak Amplitude = 2 [a.u. (ΔF/F)] for OX/Hcrt and VGlut2 neurons, or = 1 for VGAT neurons.\u003c/p\u003e\u003cp\u003e\u003cb\u003eState Selectivity and Stability Analysis\u003c/b\u003e\u003c/p\u003e\u003cp\u003eThe calcium traces were analyzed for each mouse and each session separately, to determine their state selectivity and their changes in different recording sessions and across multiple days. Calcium transients were identified as described above and separated into each different wake-sleep cycle state (awake, non-REM, and REM), according to where the calcium peak would occur. States with duration shorter than 20 seconds were removed from the analysis, as considered unstable.\u003c/p\u003e\u003cp\u003eThe frequency of calcium transient events (frequency of events) was then computed per cell, for each state, as the number of events happened during all stable states of a certain type, divided by the total duration of the stable states of that type. Cells with events’ frequency in the lower 25 percentile of all cells, in all three states, were considered separately as ‘inactive’ or ‘low active’ cells, and given the lack of activity, their state preference could not be estimated. All remaining cells were assigned as having state preference for the state where their frequency of events is maximal.\u003c/p\u003e\u003cp\u003e\u003cb\u003eCalcium traces integrals and integrals frequency\u003c/b\u003e\u003c/p\u003e\u003cp\u003eSingle-cell calcium traces were baseline corrected (baseline subtraction), and the integral of the entire trace was computed as the simple sum of all data points belonging to a specific state. The value obtained was divided by the duration of the state to obtain the ‘integral frequency’. Only states with a duration of 20 seconds or longer were considered to exclude unstable states.\u003c/p\u003e\u003cp\u003e\u003cb\u003eNetwork and Betweenness analysis\u003c/b\u003e\u003c/p\u003e\u003cp\u003eThe betweenness is a measure of the centrality of a node in a network, based on an estimation of how ‘disruptive’ to the network the removal of that specific node would be. To compute it, we first make a correlation-based network analysis between each calcium trace. As a pre-processing step for the correlation analysis, white noise, of the same amplitude as the estimated noise of the calcium traces, was summed to the raw calcium traces, to minimize the influence of noise correlations. State-specific traces were built by isolating all the stable states (longer than 20 seconds) of the same type, and gluing them together into three single traces, one for each state, ending up with 3 x \u003cem\u003enumber of cells\u003c/em\u003e traces. These state-specific traces were then correlated with each other over a sliding time window of length = 60 frames (6 seconds). Shuffled traces were used to compute a threshold for minimal significant correlation, for each state: this threshold was then used to weed out any non-significant correlation, setting it to zero. The remaining significant correlations were used to construct the upper triangular adjacency matrix for the undirected weighted graph, built via the “graph” Matlab function. Autoconnections, inactive cells, and cells that were isolated were removed from the graph and not considered in the rest of the analysis. At this point, we computed the undirected betweenness centrality \u003cem\u003eC\u003c/em\u003e of a node \u003cem\u003eu\u003c/em\u003e as (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e)\u003c/p\u003e\u003cp\u003e(\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e). C(u) = ∑\u003csub\u003es,t≠u\u003c/sub\u003e (n\u003csub\u003est\u003c/sub\u003e(u) / N\u003csub\u003est\u003c/sub\u003e),\u003c/p\u003e\u003cp\u003ewhere n\u003csub\u003est\u003c/sub\u003e is the number of shortest paths from s to t that pass through node u, and N\u003csub\u003est\u003c/sub\u003e is the number of shortest paths from s to t.\u003c/p\u003e\u003cp\u003e\u003cb\u003ePharmacology treatment and recording\u003c/b\u003e\u003c/p\u003e\u003cp\u003eDiazepam solution (DZP, Valium, 5 mg/ml) was prepared by diluting it in NaCl to a concentration of 2.5 mg/ml. To mitigate the stress associated with intraperitoneal injections, mice underwent a three-day acclimatization prior to the commencement of the experiment, during which they received 0.1 ml of NaCl intraperitoneally each day. This pre-treatment aimed to familiarize the mice with the handling and injection procedures. To explore the effects of DZP on sleep phases in mice, either 0.1 ml NaCl or DZP, at the dose of 5 mg/kg, was administered 10 minutes before the onset of the light phase (ZT0, 8 am) for three consecutive days. EEG and EMG recordings were then continuously taken for 12 hours following the initial administrations to monitor changes in sleep patterns. Additionally, calcium imaging targeting POA-Vgat neurons was performed for 30 minutes at 1-, 4-, and 8-hours post-administration each day to evaluate the neuronal activities response to the treatments.\u003c/p\u003e\u003cp\u003eFollowing this phase, the animals were left undisturbed and handled minimally for at least 3 days to ensure complete metabolism of DZP before beginning treatments in the active phase. Subsequently, a single injection of either 0.1 ml NaCl or DZP (5 mg/kg) was administered 10 minutes before the start of the night phase (ZT12, 8 pm). EEG and EMG were again recorded for 12 hours to assess the sleep patterns post-treatment. Concurrently, calcium imaging was conducted 1-hour post-treatment for approximately 30 minutes to monitor the neuronal activities.\u003c/p\u003e\u003cp\u003e\u003cb\u003eImmunohistochemistry\u003c/b\u003e\u003c/p\u003e\u003cp\u003eTo assess viral vector expression and confirm accurate anatomical targeting of brain regions, brain slices were imaged after the conclusion of all experiments. Mice were deeply anesthetized using an intraperitoneal injection of pentobarbital (250 mg/kg, Streuli Pharma). Once fully anesthetized, transcardial perfusion was performed with 5–10 ml of cold saline (0.9% NaCl), followed by 20–25 ml of 4% formaldehyde (Grogg Chemie) to fix the tissues. After perfusion, the brains were removed and stored overnight in formaldehyde at 4°C for complete fixation. The following day, the brains were transferred to a phosphate-buffered saline (PBS) solution containing 30% sucrose and kept there until they sank, ensuring full cryoprotection before sectioning.\u003c/p\u003e\u003cp\u003eThe brains were sectioned into 40 µm slices using a cryostat (Hyrax C 25, Zeiss) and arranged in triplicate sets of 1:3 series, which were collected in PBS. To verify the co-expression of GCaMP6 with MCH or OX/Hcrt, sections were washed in PBST (PBS with 0.1% Triton X-100, Sigma-Aldrich) for 5 x 5 minutes and incubated for 45 minutes at room temperature in a blocking solution containing PBST and 4% bovine serum albumin (Sigma-Aldrich). The sections were then incubated with anti-pMCH (goat, 1:500, sc-14509, Santa Cruz Biotechnology) or anti-Orexin A (rabbit, 1:1000, H-003-30, Phoenix Pharmaceuticals) for 24–48 hours.\u003c/p\u003e\u003cp\u003eAfter incubation, the sections were washed again for 5 x 5 minutes and then incubated for 2 hours at room temperature with AlexaFluor555-conjugated secondary antibody (1:1000 dilution, Invitrogen, A21432 for MCH and OX/Hcrt staining). The sections were then washed for 3 x 10 minutes and mounted on glass slides. Cover slips were applied using Fluoromount (F4680, Sigma-Aldrich). The GCaMP6s fluorescence was easily detected, and no additional staining was necessary.\u003c/p\u003e\u003cp\u003eSlides were imaged using a Nikon-PLAN Fluor 10x/0.3NA objective on a Nikon Eclipse Ti-E fluorescence microscope controlled by Nikon NIS software. Excitation was achieved with a Solar LED lamp using Cy3 (mCherry) and FITC (eGFP) filters. For display purposes, image brightness and contrast were moderately adjusted in Fiji (ImageJ), and figures were compiled using Adobe Illustrator 2020.\u003c/p\u003e\u003ch2\u003eStatistical Analysis\u003c/h2\u003e\u003cp\u003eAll data are presented as means ± standard error of the mean (SEM), with a significance level (α) of 0.05 for all two-sided tests. Statistical analysis was performed using MATLAB (MathWorks) and GraphPad Prism 10, focusing on graphed data. Significant results are detailed in the text, and sample sizes are provided in the corresponding figure legends. No formal power calculations were performed, but sample sizes were consistent with those used in similar studies.\u003c/p\u003e\u003cp\u003eData were compared using t-tests for parametric data, one-way repeated measures ANOVA with Tukey’s post-hoc correction for multiple comparisons, or two-way RM ANOVA with Sidak’s multiple comparisons test. While data distribution was assumed to be normal, this was not formally tested. Experimenters were not blinded to the conditions during data acquisition or analysis.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eAcknowledgments:\u003c/h2\u003e\n\u003cp\u003eWe thank all the Tidis Laboratory members, M. Schmidt, M. Baud, and C. Bellone, B. Engelhardt, T. Korotkova for their insightful discussions.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e: This work was supported by the Inselspital University Hospital Bern, the Swiss National Science Foundation (A.A. 310030_188761), the China Scholarship Council (Y.Y), and the University of Bern (A.A.).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e: A.A. conceived the study. Y.Y performed the experiments. N.C developed the Ca\u003csup\u003e2+\u003c/sup\u003e imaging analysis tools. Y.Y, N.C. and T.R. analyzed the data. Y.Y. wrote the original draft and A.A, N.C, C.G.H reviewed \u0026amp; edited the manuscript. A.A. and C.G.H supervised the research.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e: Authors declare that they have no competing interests.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData and materials availability\u003c/strong\u003e: All presented data and analysis scripts, including mat-files and Matlab scripts and functions, are available on \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://github.com/ZENLabCode\u003c/span\u003e\u003c/span\u003e. In addition, all data that supports the findings of this study are available from the corresponding authors upon request.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eJ. K. Elmquist, C. F. Elias, C. B. Saper, From lesions to leptin: hypothalamic control of food intake and body weight. \u003cem\u003eNeuron\u003c/em\u003e \u003cstrong\u003e22\u003c/strong\u003e, 221\u0026ndash;232 (1999).\u003c/li\u003e\n\u003cli\u003eC. Roger, A. Lasbleiz, M. Guye, A. Dutour, B. Gaborit, J.-P. Ranjeva, The Role of the Human Hypothalamus in Food Intake Networks: An MRI Perspective. \u003cem\u003eFront Nutr\u003c/em\u003e \u003cstrong\u003e8\u003c/strong\u003e, 760914 (2021).\u003c/li\u003e\n\u003cli\u003eE. K. McIlwraith, D. D. Belsham, Hypothalamic reproductive neurons communicate through signal transduction to control reproduction. \u003cem\u003eMol Cell Endocrinol\u003c/em\u003e \u003cstrong\u003e518\u003c/strong\u003e, 110971 (2020).\u003c/li\u003e\n\u003cli\u003eL. Mei, R. Yan, L. Yin, R. M. Sullivan, D. Lin, Antagonistic circuits mediating infanticide and maternal care in female mice. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e618\u003c/strong\u003e, 1006\u0026ndash;1016 (2023).\u003c/li\u003e\n\u003cli\u003eJ. P. Herman, M. M. Ostrander, N. K. Mueller, H. Figueiredo, Limbic system mechanisms of stress regulation: hypothalamo-pituitary-adrenocortical axis. \u003cem\u003eProg Neuropsychopharmacol Biol Psychiatry\u003c/em\u003e \u003cstrong\u003e29\u003c/strong\u003e, 1201\u0026ndash;1213 (2005).\u003c/li\u003e\n\u003cli\u003eT. Kuwaki, Orexin (hypocretin) participates in central autonomic regulation during fight-or-flight response. \u003cem\u003ePeptides\u003c/em\u003e \u003cstrong\u003e139\u003c/strong\u003e, 170530 (2021).\u003c/li\u003e\n\u003cli\u003eC. Von Economo, Sleep as a problem of localization. \u003cem\u003eJournal of Nervous and Mental Disease\u003c/em\u003e \u003cstrong\u003e71\u003c/strong\u003e, 249\u0026ndash;259 (1930).\u003c/li\u003e\n\u003cli\u003eC. B. Saper, T. E. Scammell, J. Lu, Hypothalamic regulation of sleep and circadian rhythms. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e437\u003c/strong\u003e, 1257\u0026ndash;1263 (2005).\u003c/li\u003e\n\u003cli\u003eA. R. Adamantidis, L. de Lecea, Sleep and the hypothalamus, \u003cem\u003eScience\u003c/em\u003e. \u003cstrong\u003e282\u003c/strong\u003e (2023). https://doi.org/10.1126/science.adh8285.\u003c/li\u003e\n\u003cli\u003eP. Bonnavion, L. E. Mickelsen, A. Fujita, L. de Lecea, A. C. Jackson, Hubs and spokes of the lateral hypothalamus: cell types, circuits and behaviour. \u003cem\u003eJ Physiol\u003c/em\u003e \u003cstrong\u003e594\u003c/strong\u003e, 6443\u0026ndash;6462 (2016).\u003c/li\u003e\n\u003cli\u003eO. K. Hassani, P. Henny, M. G. Lee, B. E. Jones, GABAergic neurons intermingled with orexin and MCH neurons in the lateral hypothalamus discharge maximally during sleep. \u003cem\u003eEur J Neurosci\u003c/em\u003e \u003cstrong\u003e32\u003c/strong\u003e, 448\u0026ndash;457 (2010).\u003c/li\u003e\n\u003cli\u003eA. Venner, C. Anaclet, R. Y. Broadhurst, C. B. Saper, P. M. Fuller, A Novel Population of Wake-Promoting GABAergic Neurons in the Ventral Lateral Hypothalamus. \u003cem\u003eCurr Biol\u003c/em\u003e \u003cstrong\u003e26\u003c/strong\u003e, 2137\u0026ndash;2143 (2016).\u003c/li\u003e\n\u003cli\u003eL. E. Mickelsen, M. Bolisetty, B. R. Chimileski, A. Fujita, E. J. Beltrami, J. T. Costanzo, J. R. Naparstek, P. Robson, A. C. Jackson, Single-cell transcriptomic analysis of the lateral hypothalamic area reveals molecularly distinct populations of inhibitory and excitatory neurons. \u003cem\u003eNat Neurosci\u003c/em\u003e \u003cstrong\u003e22\u003c/strong\u003e, 642\u0026ndash;656 (2019).\u003c/li\u003e\n\u003cli\u003eR.-F. Wang, H. Guo, S.-Y. Jiang, Z.-L. Liu, W.-M. Qu, Z.-L. Huang, L. Wang, Control of wakefulness by lateral hypothalamic glutamatergic neurons in male mice. \u003cem\u003eJ Neurosci Res\u003c/em\u003e \u003cstrong\u003e99\u003c/strong\u003e, 1689\u0026ndash;1703 (2021).\u003c/li\u003e\n\u003cli\u003eP. Henny, F. Brischoux, L. Mainville, T. Stroh, B. E. Jones, Immunohistochemical evidence for synaptic release of glutamate from orexin terminals in the locus coeruleus. \u003cem\u003eNeuroscience\u003c/em\u003e \u003cstrong\u003e169\u003c/strong\u003e, 1150\u0026ndash;1157 (2010).\u003c/li\u003e\n\u003cli\u003eC. Sch\u0026ouml;ne, J. Apergis-Schoute, T. Sakurai, A. Adamantidis, D. Burdakov, Coreleased Orexin and Glutamate Evoke Nonredundant Spike Outputs and Computations in Histamine Neurons. \u003cem\u003eCell Reports\u003c/em\u003e \u003cstrong\u003e7\u003c/strong\u003e, 697\u0026ndash;704 (2014).\u003c/li\u003e\n\u003cli\u003eB. Y. Mileykovskiy, L. I. Kiyashchenko, J. M. Siegel, Behavioral correlates of activity in identified hypocretin/orexin neurons. \u003cem\u003eNeuron\u003c/em\u003e \u003cstrong\u003e46\u003c/strong\u003e, 787\u0026ndash;798 (2005).\u003c/li\u003e\n\u003cli\u003eO. K. Hassani, M. R. Krause, L. Mainville, C. A. Cordova, B. E. Jones, Orexin Neurons Respond Differentially to Auditory Cues Associated with Appetitive versus Aversive Outcomes. \u003cem\u003eThe Journal of Neuroscience\u003c/em\u003e \u003cstrong\u003e36\u003c/strong\u003e, 1747 (2016).\u003c/li\u003e\n\u003cli\u003eC. Concetti, D. Burdakov, Orexin/Hypocretin and MCH Neurons: Cognitive and Motor Roles Beyond Arousal. \u003cem\u003eFront. Neurosci.\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e (2021).\u003c/li\u003e\n\u003cli\u003eA. R. Adamantidis, F. Zhang, A. M. Aravanis, K. Deisseroth, L. de Lecea, Neural substrates of awakening probed with optogenetic control of hypocretin neurons. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e450\u003c/strong\u003e, 420\u0026ndash;424 (2007).\u003c/li\u003e\n\u003cli\u003eO. K. Hassani, M. G. Lee, B. E. Jones, Melanin-concentrating hormone neurons discharge in a reciprocal manner to orexin neurons across the sleep-wake cycle. \u003cem\u003eProc Natl Acad Sci U S A\u003c/em\u003e \u003cstrong\u003e106\u003c/strong\u003e, 2418\u0026ndash;2422 (2009).\u003c/li\u003e\n\u003cli\u003eN. Suntsova, R. Szymusiak, N. Md. Alam, R. Guzman-Marin, D. McGinty, Sleep-waking discharge patterns of median preoptic nucleus neurons in rats. \u003cem\u003eThe Journal of Physiology\u003c/em\u003e \u003cstrong\u003e543\u003c/strong\u003e, 665\u0026ndash;677 (2002).\u003c/li\u003e\n\u003cli\u003eH. Gong, D. McGinty, R. Guzman-Marin, K.-T. Chew, D. Stewart, R. Szymusiak, Activation of c-fos in GABAergic neurones in the preoptic area during sleep and in response to sleep deprivation. \u003cem\u003eThe Journal of Physiology\u003c/em\u003e \u003cstrong\u003e556\u003c/strong\u003e, 935\u0026ndash;946 (2004).\u003c/li\u003e\n\u003cli\u003eG. Miracca, B. Anuncibay-Soto, K. Tossell, R. Yustos, A. L. Vyssotski, N. P. Franks, W. Wisden, NMDA Receptors in the Lateral Preoptic Hypothalamus Are Essential for Sustaining NREM and REM Sleep. \u003cem\u003eJ Neurosci\u003c/em\u003e \u003cstrong\u003e42\u003c/strong\u003e, 5389\u0026ndash;5409 (2022).\u003c/li\u003e\n\u003cli\u003eM. A. Alam, S. Kumar, D. McGinty, M. N. Alam, R. Szymusiak, Neuronal activity in the preoptic hypothalamus during sleep deprivation and recovery sleep. \u003cem\u003eJ Neurophysiol\u003c/em\u003e \u003cstrong\u003e111\u003c/strong\u003e, 287\u0026ndash;299 (2014).\u003c/li\u003e\n\u003cli\u003eN. L. S. Machado, W. D. Todd, S. Kaur, C. B. Saper, Median preoptic GABA and glutamate neurons exert differential control over sleep behavior. \u003cem\u003eCurr Biol\u003c/em\u003e \u003cstrong\u003e32\u003c/strong\u003e, 2011-2021.e3 (2022).\u003c/li\u003e\n\u003cli\u003eL. Verret, R. Goutagny, P. Fort, L. Cagnon, D. Salvert, L. L\u0026eacute;ger, R. Boissard, P. Salin, C. Peyron, P.-H. Luppi, A role of melanin-concentrating hormone producing neurons in the central regulation of paradoxical sleep. \u003cem\u003eBMC Neurosci\u003c/em\u003e \u003cstrong\u003e4\u003c/strong\u003e, 19 (2003).\u003c/li\u003e\n\u003cli\u003eM. N. Alam, H. Gong, T. Alam, R. Jaganath, D. McGinty, R. Szymusiak, Sleep-waking discharge patterns of neurons recorded in the rat perifornical lateral hypothalamic area. \u003cem\u003eJ Physiol\u003c/em\u003e \u003cstrong\u003e538\u003c/strong\u003e, 619\u0026ndash;631 (2002).\u003c/li\u003e\n\u003cli\u003eJ. C. Borniger, L. de Lecea, Peripheral Lipopolyssacharide Rapidly Silences REM-Active LHGABA Neurons. \u003cem\u003eFront Behav Neurosci\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 649428 (2021).\u003c/li\u003e\n\u003cli\u003eH. Ito, N. Fukatsu, S. M. Rahaman, Y. Mukai, S. Izawa, D. Ono, T. S. Kilduff, A. Yamanaka, Deficiency of orexin signaling during sleep is involved in abnormal REM sleep architecture in narcolepsy. \u003cem\u003eProceedings of the National Academy of Sciences of the United States of America\u003c/em\u003e \u003cstrong\u003e120\u003c/strong\u003e (2023).\u003c/li\u003e\n\u003cli\u003eL. C. Freeman, A set of measures of centrality based on betweenness. \u003cem\u003eSociometry\u003c/em\u003e \u003cstrong\u003e40\u003c/strong\u003e, 35\u0026ndash;41 (1977).\u003c/li\u003e\n\u003cli\u003eL. E. McKillop, S. P. Fisher, L. Milinski, L. B. Krone, V. V. Vyazovskiy, Diazepam effects on local cortical neural activity during sleep in mice. \u003cem\u003eBiochem Pharmacol\u003c/em\u003e \u003cstrong\u003e191\u003c/strong\u003e, 114515 (2021).\u003c/li\u003e\n\u003cli\u003eL. N. Driscoll, L. Duncker, C. D. Harvey, Representational drift: Emerging theories for continual learning and experimental future directions. \u003cem\u003eCurr Opin Neurobiol\u003c/em\u003e \u003cstrong\u003e76\u003c/strong\u003e, 102609 (2022).\u003c/li\u003e\n\u003cli\u003eT. D. Marks, M. J. Goard, Stimulus-dependent representational drift in primary visual cortex. \u003cem\u003eNat Commun\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, 5169 (2021).\u003c/li\u003e\n\u003cli\u003eC. E. Schoonover, S. N. Ohashi, R. Axel, A. J. P. Fink, Representational drift in primary olfactory cortex. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e594\u003c/strong\u003e, 541\u0026ndash;546 (2021).\u003c/li\u003e\n\u003cli\u003eY. Ziv, L. D. Burns, E. D. Cocker, E. O. Hamel, K. K. Ghosh, L. J. Kitch, A. E. Gamal, M. J. Schnitzer, Long-term dynamics of CA1 hippocampal place codes. \u003cem\u003eNat Neurosci\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 264\u0026ndash;266 (2013).\u003c/li\u003e\n\u003cli\u003eD. Deitch, A. Rubin, Y. Ziv, Representational drift in the mouse visual cortex. \u003cem\u003eCurr Biol\u003c/em\u003e \u003cstrong\u003e31\u003c/strong\u003e, 4327-4339.e6 (2021).\u003c/li\u003e\n\u003cli\u003eZ. N. Roth, E. P. Merriam, Representations in human primary visual cortex drift over time. \u003cem\u003eNat Commun\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e, 4422 (2023).\u003c/li\u003e\n\u003cli\u003eD. J. Cai, D. Aharoni, T. Shuman, J. Shobe, J. Biane, W. Song, B. Wei, M. Veshkini, M. La-Vu, J. Lou, S. E. Flores, I. Kim, Y. Sano, M. Zhou, K. Baumgaertel, A. Lavi, M. Kamata, M. Tuszynski, M. Mayford, P. Golshani, A. J. Silva, A shared neural ensemble links distinct contextual memories encoded close in time. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e534\u003c/strong\u003e, 115\u0026ndash;118 (2016).\u003c/li\u003e\n\u003cli\u003eM. E. Rule, A. R. Loback, D. V. Raman, L. N. Driscoll, C. D. Harvey, T. O\u0026rsquo;Leary, Stable task information from an unstable neural population. \u003cem\u003eElife\u003c/em\u003e \u003cstrong\u003e9\u003c/strong\u003e, e51121 (2020).\u003c/li\u003e\n\u003cli\u003eD. Gerashchenko, C. Blanco-Centurion, M. A. Greco, P. J. Shiromani, Effects of lateral hypothalamic lesion with the neurotoxin hypocretin-2\u0026ndash;saporin on sleep in Long\u0026ndash;Evans rats. \u003cem\u003eNeuroscience\u003c/em\u003e \u003cstrong\u003e116\u003c/strong\u003e, 223\u0026ndash;235 (2003).\u003c/li\u003e\n\u003cli\u003eC. J. Hung, D. Ono, T. S. Kilduff, A. Yamanaka, Dual orexin and MCH neuron-ablated mice display severe sleep attacks and cataplexy. \u003cem\u003eeLife\u003c/em\u003e \u003cstrong\u003e9\u003c/strong\u003e, e54275 (2020).\u003c/li\u003e\n\u003cli\u003eD. Gerashchenko, T. C. Chou, C. A. Blanco-Centurion, C. B. Saper, P. J. Shiromani, Effects of lesions of the histaminergic tuberomammillary nucleus on spontaneous sleep in rats. \u003cem\u003eSleep\u003c/em\u003e \u003cstrong\u003e27\u003c/strong\u003e, 1275\u0026ndash;1281 (2004).\u003c/li\u003e\n\u003cli\u003eT. Sakurai, R. Nagata, A. Yamanaka, H. Kawamura, N. Tsujino, Y. Muraki, H. Kageyama, S. Kunita, S. Takahashi, K. Goto, Y. Koyama, S. Shioda, M. Yanagisawa, Input of orexin/hypocretin neurons revealed by a genetically encoded tracer in mice. \u003cem\u003eNeuron\u003c/em\u003e \u003cstrong\u003e46\u003c/strong\u003e, 297\u0026ndash;308 (2005).\u003c/li\u003e\n\u003cli\u003eC. Sch\u0026ouml;ne, D. Burdakov, Orexin/Hypocretin and Organizing Principles for a Diversity of Wake-Promoting Neurons in the Brain. \u003cem\u003eCurr Top Behav Neurosci\u003c/em\u003e \u003cstrong\u003e33\u003c/strong\u003e, 51\u0026ndash;74 (2017).\u003c/li\u003e\n\u003cli\u003eJ. A. Gonz\u0026aacute;lez, P. Iordanidou, M. Strom, A. Adamantidis, D. Burdakov, Awake dynamics and brain-wide direct inputs of hypothalamic MCH and orexin networks. \u003cem\u003eNat Commun\u003c/em\u003e \u003cstrong\u003e7\u003c/strong\u003e, 11395 (2016).\u003c/li\u003e\n\u003cli\u003eC. Blanco-Centurion, E. Bendell, B. Zou, Y. Sun, P. J. Shiromani, M. Liu, VGAT and VGLUT2 expression in MCH and orexin neurons in double transgenic reporter mice. \u003cem\u003eIBRO Rep\u003c/em\u003e \u003cstrong\u003e4\u003c/strong\u003e, 44\u0026ndash;49 (2018).\u003c/li\u003e\n\u003cli\u003eM. G. Lee, O. K. Hassani, B. E. Jones, Discharge of identified orexin/hypocretin neurons across the sleep-waking cycle. \u003cem\u003eJ Neurosci\u003c/em\u003e \u003cstrong\u003e25\u003c/strong\u003e, 6716\u0026ndash;6720 (2005).\u003c/li\u003e\n\u003cli\u003eL. T. Oesch, M. Gazea, T. C. Gent, M. Bandarabadi, C. G. Herrera, A. R. Adamantidis, REM sleep stabilizes hypothalamic representation of feeding behavior. \u003cem\u003eProceedings of the National Academy of Sciences of the United States of America\u003c/em\u003e \u003cstrong\u003e117\u003c/strong\u003e (2020).\u003c/li\u003e\n\u003cli\u003eD. Burdakov, M. M. Karnani, Ultra-sparse Connectivity within the Lateral Hypothalamus. \u003cem\u003eCurrent Biology\u003c/em\u003e \u003cstrong\u003e30\u003c/strong\u003e, 4063-4070.e2 (2020).\u003c/li\u003e\n\u003cli\u003eN. G. Hedrick, Z. Lu, E. Bushong, S. Singhi, P. Nguyen, Y. Maga\u0026ntilde;a, S. Jilani, B. K. Lim, M. Ellisman, T. Komiyama, Learning binds new inputs into functional synaptic clusters via spinogenesis. \u003cem\u003eNat Neurosci\u003c/em\u003e \u003cstrong\u003e25\u003c/strong\u003e, 726\u0026ndash;737 (2022).\u003c/li\u003e\n\u003cli\u003eG. Yang, C. S. W. Lai, J. Cichon, L. Ma, W. Li, W.-B. Gan, Sleep promotes branch-specific formation of dendritic spines after learning. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e344\u003c/strong\u003e, 1173\u0026ndash;1178 (2014).\u003c/li\u003e\n\u003cli\u003eM. N. Insanally, B. F. Albanna, J. Toth, B. DePasquale, S. S. Fadaei, T. Gupta, O. Lombardi, K. Kuchibhotla, K. Rajan, R. C. Froemke, Contributions of cortical neuron firing patterns, synaptic connectivity, and plasticity to task performance. \u003cem\u003eNat Commun\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 6023 (2024).\u003c/li\u003e\n\u003cli\u003eG. Tononi, C. Cirelli, Sleep and the Price of Plasticity: From Synaptic and Cellular Homeostasis to Memory Consolidation and Integration. \u003cem\u003eNeuron\u003c/em\u003e \u003cstrong\u003e81\u003c/strong\u003e, 12\u0026ndash;34 (2014).\u003c/li\u003e\n\u003cli\u003eY. Vanrobaeys, Z. J. Peterson, E. N. Walsh, S. Chatterjee, L.-C. Lin, L. C. Lyons, T. Nickl-Jockschat, T. Abel, Spatial transcriptomics reveals unique gene expression changes in different brain regions after sleep deprivation. \u003cem\u003eNat Commun\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e, 7095 (2023).\u003c/li\u003e\n\u003cli\u003eY. Wang, S. Cao, D. Tone, H. Fujishima, R. G. Yamada, R.-I. Ohno, S. Shi, K. Matsuzawa, S. Yada, M. Kaneko, H. Sakamoto, T. Onishi, M. Ukai-Tadenuma, H. Ukai, C. Hanashima, K. Hirose, H. Kiyonari, K. Sumiyama, K. L. Ode, H. R. Ueda, Postsynaptic competition between calcineurin and PKA regulates mammalian sleep-wake cycles. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e636\u003c/strong\u003e, 412\u0026ndash;421 (2024).\u003c/li\u003e\n\u003cli\u003eF. Br\u0026uuml;ning, S. B. Noya, T. Bange, S. Koutsouli, J. D. Rudolph, S. K. Tyagarajan, J. Cox, M. Mann, S. A. Brown, M. S. Robles, Sleep-wake cycles drive daily dynamics of synaptic phosphorylation. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e366\u003c/strong\u003e, eaav3617 (2019).\u003c/li\u003e\n\u003cli\u003eL. De Vivo, M. Bellesi, W. Marshall, E. A. Bushong, M. H. Ellisman, G. Tononi, C. Cirelli, Ultrastructural evidence for synaptic scaling across the wake/sleep cycle. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e355\u003c/strong\u003e, 507\u0026ndash;510 (2017).\u003c/li\u003e\n\u003cli\u003eX.-F. Du, F.-N. Li, X.-L. Peng, B. Xu, Y. Zhang, G. Li, T. Liu, Y. Li, H. Wang, J. Yan, J.-L. Du, Circadian regulation of developmental synaptogenesis via the hypocretinergic system. \u003cem\u003eNat Commun\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e, 3195 (2023).\u003c/li\u003e\n\u003cli\u003eF. M. R. de Mendon\u0026ccedil;a, G. P. R. R. de Mendon\u0026ccedil;a, L. C. Souza, L. P. Galv\u0026atilde;o, H. S. Paiva, C. de Azevedo Marques P\u0026eacute;rico, J. Torales, A. Ventriglio, J. M. Castaldelli-Maia, A. Sousa Martins Silva, Benzodiazepines and Sleep Architecture: A Systematic Review. \u003cem\u003eCNS Neurol Disord Drug Targets\u003c/em\u003e \u003cstrong\u003e22\u003c/strong\u003e, 172\u0026ndash;179 (2023).\u003c/li\u003e\n\u003cli\u003eJ. Lim, D. F. Dinges, A meta-analysis of the impact of short-term sleep deprivation on cognitive variables. \u003cem\u003ePsychol Bull\u003c/em\u003e \u003cstrong\u003e136\u003c/strong\u003e, 375\u0026ndash;389 (2010).\u003c/li\u003e\n\u003cli\u003eC. M. McDermott, G. J. LaHoste, C. Chen, A. Musto, N. G. Bazan, J. C. Magee, Sleep deprivation causes behavioral, synaptic, and membrane excitability alterations in hippocampal neurons. \u003cem\u003eJ Neurosci\u003c/em\u003e \u003cstrong\u003e23\u003c/strong\u003e, 9687\u0026ndash;9695 (2003).\u003c/li\u003e\n\u003cli\u003eD. P. Windred, A. C. Burns, J. M. Lane, R. Saxena, M. K. Rutter, S. W. Cain, A. J. K. Phillips, Sleep regularity is a stronger predictor of mortality risk than sleep duration: A prospective cohort study. \u003cem\u003eSleep\u003c/em\u003e \u003cstrong\u003e47\u003c/strong\u003e, zsad253 (2024).\u003c/li\u003e\n\u003cli\u003eD. A. Kalmbach, J. R. Anderson, C. L. Drake, The impact of stress on sleep: Pathogenic sleep reactivity as a vulnerability to insomnia and circadian disorders. \u003cem\u003eJ Sleep Res\u003c/em\u003e \u003cstrong\u003e27\u003c/strong\u003e, e12710 (2018).\u003c/li\u003e\n\u003cli\u003eS. J. Riede, V. van der Vinne, R. A. Hut, The flexible clock: predictive and reactive homeostasis, energy balance and the circadian regulation of sleep-wake timing. \u003cem\u003eJ Exp Biol\u003c/em\u003e \u003cstrong\u003e220\u003c/strong\u003e, 738\u0026ndash;749 (2017).\u003c/li\u003e\n\u003cli\u003eR. Rothhaas, S. Chung, Role of the Preoptic Area in Sleep and Thermoregulation. \u003cem\u003eFrontiers in Neuroscience\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 664781 (2021).\u003c/li\u003e\n\u003cli\u003eM. H. Schmidt, The energy allocation function of sleep: a unifying theory of sleep, torpor, and continuous wakefulness. \u003cem\u003eNeurosci Biobehav Rev\u003c/em\u003e \u003cstrong\u003e47\u003c/strong\u003e, 122\u0026ndash;153 (2014).\u003c/li\u003e\n\u003cli\u003eH. Fong, J. Zheng, D. Kurrasch, The structural and functional complexity of the integrative hypothalamus. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e382\u003c/strong\u003e, 388\u0026ndash;394 (2023).\u003c/li\u003e\n\u003cli\u003eA. Tauffenberger, P. J. Magistretti, Reactive Oxygen Species: Beyond Their Reactive Behavior. \u003cem\u003eNeurochem Res\u003c/em\u003e \u003cstrong\u003e46\u003c/strong\u003e, 77\u0026ndash;87 (2021).\u003c/li\u003e\n\u003cli\u003eD. Zada, Y. Sela, N. Matosevich, A. Monsonego, T. Lerer-Goldshtein, Y. Nir, L. Appelbaum, Parp1 promotes sleep, which enhances DNA repair in neurons. \u003cem\u003eMol Cell\u003c/em\u003e \u003cstrong\u003e81\u003c/strong\u003e, 4979-4993.e7 (2021).\u003c/li\u003e\n\u003cli\u003eF. Longordo, J. Fan, T. Steimer, C. Kopp, A. L\u0026uuml;thi, Do Mice Habituate to \u0026ldquo;Gentle Handling?\u0026rdquo; A Comparison of Resting Behavior, Corticosterone Levels and Synaptic Function in Handled and Undisturbed C57BL/6J Mice. \u003cem\u003eSleep\u003c/em\u003e \u003cstrong\u003e34\u003c/strong\u003e, 679\u0026ndash;681 (2011).\u003c/li\u003e\n\u003cli\u003eA. Lemons, R. M. Sar\u0026eacute;, C. Beebe Smith, Chronic Sleep Deprivation in Mouse Pups by Means of Gentle Handling. \u003cem\u003eJ Vis Exp\u003c/em\u003e, 58150 (2018).\u003c/li\u003e\n\u003cli\u003eR. Boyce, S. D. Glasgow, S. Williams, A. Adamantidis, Sleep research: Causal evidence for the role of REM sleep theta rhythm in contextual memory consolidation. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e352\u003c/strong\u003e, 812\u0026ndash;816 (2016).\u003c/li\u003e\n\u003cli\u003eA. Mensen, B. Riedner, G. Tononi, Optimizing detection and analysis of slow waves in sleep EEG. \u003cem\u003eJournal of Neuroscience Methods\u003c/em\u003e \u003cstrong\u003e274\u003c/strong\u003e, 1\u0026ndash;12 (2016).\u003c/li\u003e\n\u003cli\u003eM. Panagiotou, V. V. Vyazovskiy, J. H. Meijer, T. Deboer, Differences in electroencephalographic non-rapid-eye movement sleep slow-wave characteristics between young and old mice. \u003cem\u003eSci Rep\u003c/em\u003e \u003cstrong\u003e7\u003c/strong\u003e, 43656 (2017).\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-6960125/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6960125/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eCorrelative and causal evidence implicate distinct genetically-defined and evolutionary-conserved hypothalamic neurons in regulating wakefulness, non-rapid eye movement (NREM), and rapid eye movement (REM) sleep. The prevailing view is that these circuits govern sleep-wake states by recruiting stable, invariant neuronal substrates, yet, this remains unknown. Here, we showed that inhibitory, excitatory, hypocretins/orexins-, and melanin concentrating hormone-expressing-neurons in hypothalamus did not exhibit stable state-specific activities using longitudinal single cell calcium imaging in freely-moving sleeping mice. Instead, their activity patterns shift across sleep-wake states over time, while the distribution of active neurons in each sleep state remained stable. While sleep deprivation minimally affected the selectivity of these activity patterns, we found that the sleep-promoting drug diazepam recruited NREM sleep-active cells that were previously inactive or wake-active. These findings indicate that while individual neurons exhibit dynamic, state-dependent shifts of their activity, the overall organization of sleep-wake neural populations remains stable.\u003c/p\u003e","manuscriptTitle":"Temporal drift of sleep-wake representations in hypothalamic neuronal ensembles","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-08-06 08:45:57","doi":"10.21203/rs.3.rs-6960125/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-communications","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"NCOMMS","sideBox":"Learn more about [Nature Communications](http://www.nature.com/ncomms/)","snPcode":"","submissionUrl":"https://mts-ncomms.nature.com/","title":"Nature Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Communications","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"51df7989-8f7a-4b3f-afd6-83e682f98f9f","owner":[],"postedDate":"August 6th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":51361054,"name":"Biological sciences/Systems biology/Single-cell imaging"},{"id":51361055,"name":"Biological sciences/Neuroscience/Circadian rhythms and sleep/Sleep"}],"tags":[],"updatedAt":"2025-08-06T08:45:57+00:00","versionOfRecord":[],"versionCreatedAt":"2025-08-06 08:45:57","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6960125","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6960125","identity":"rs-6960125","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}