View all hidden authors and organizations

Edited by Joseph Takahashi, Department of HHMI and Neuroscience, University of Texas Southwestern Medical Center; received January 25, 2021; accepted October 21, 2021

Our current understanding of how sleep is regulated is based on the sleep homeostasis model, which defines a variable called Process S as a measure of sleep demand, and a so-called “trigger” model of state switching, which is based on subcortical sleep Facilitate and awaken the concept of antagonistic relationships between the promotion circuits. The neurobiological substrate of the interaction between sleep switch and process S is unknown. Our research has determined the previously unrecognized role of the hypothalamic circuit in regulating brain activity or arousal levels in the state, which in turn determines the steady-state drive of sleep.

Sleep and wakefulness are not simple, homogeneous all-or-nothing states, but represent a series of sub-states, distinguished by behavior, arousal levels, and brain activity at the local and global levels. So far, the role of the hypothalamic circuit in sleep-wake control is mainly to study its contribution to rapid state transition. In contrast, whether the hypothalamus regulates intra-state dynamics (state "quality") and its functional significance remain to be explored. Here, we show that the light activation of inhibitory neurons in the lateral preoptic region (LPO) of the hypothalamus in adult male and female laboratory mice not only triggers awakening from sleep, but the resulting wakefulness is also an activated brain The electrogram (EEG) is a characteristic pattern, indicating an increase in arousal level. This is related to the faster accumulation of sleep pressure, which is reflected in the higher EEG slow wave activity (SWA) during subsequent sleep. In contrast, photoinhibition of inhibitory LPO neurons does not result in changes in alertness, but is associated with a continuous increase in EEG SWA during spontaneous sleep. These findings indicate the role of LPO in regulating the level of wakefulness, and we recommend it as a key variable in shaping the daily structure of sleep-wakefulness.

The interspecies difference in daily sleep amount is strongly influenced by genetic factors (1). However, individuals also have amazing abilities to adjust the time and duration of awakening and sleep according to various internal and external factors (2). The key regulatory factors of "adaptive sleep architecture" include 1) steady-state sleep requirements, 2) endogenous biological clocks, and 3) the need to meet other physiological and behavioral needs, such as eating or avoiding danger (3⇓ –5). It is unclear how and in what form these numerous signals are integrated into the neural circuit, resulting in a rapid and stable transition between sleep and wakefulness.

For decades, brain state switching has been the main focus of circuit-oriented sleep research. Early studies identified the preoptic hypothalamus as the main candidate for the hypothetical "key sleep center" (6⇓ –8), and subsequent studies confirmed the presence of sleep-active neurons in the ventrolateral and median preoptic areas (VLPO and MPO) . Hypothalamus (9⇓ –11). Combining orexin/hypocretin neurons are necessary to maintain awake results (12, 13), and proposed a model in which sleep/wake promotion circuit acts as a trigger switch (14). This model can explain the rapid and complete transition between sleep and wakefulness, and prevent the occurrence of unstable states (15) or mixed mixed alert states (16). Over the past decade, our knowledge of the subcortical nucleus that controls sleep has steadily expanded, leading to the specialization of recognition functions in the sleep-wake control network, while highlighting previously underestimated complexity (17⇓ ⇓ ⇓ ⇓ ⇓ ⇓ ⇓ ⇓ ⇓ ⇓ ⇓ ⇓ ⇓ ⇓ ⇓ ⇓ ⇓ –35).

A key issue that arises is how to express and integrate the signals that regulate sleep-wake structure in the hypothalamic state switching circuit, so as to maximize ecological health (36). Although sleep homeostasis is considered to be an important factor that affects the sleep/wake transition (37⇓ ⇓ –40), relatively few studies have explored whether and how sleep-wake control brain regions overlap with brain regions involved in homeostatic sleep regulation ( 26) Or the underlying neurophysiological mechanism (41⇓ ⇓ –44). A recent study pointed out the important role of galanin neurons in the lateral anterior hypothalamus, as demonstrated by selective ablation, which eliminates electroencephalogram (EEG) slow wave activity (SWA; EEG power density) (Between 0.5 and 4 Hz) rebound after sleep deprivation (26). Other studies have shown that although the steady-state sleep pressure reflected in SWA is established as a function of overall wakefulness duration, it is also locally regulated by specific activities during wakefulness (45, 46). The flexibility of sleep and wakefulness as attributes of the brain state at the global and local levels indicates additional complexity, which is difficult to coordinate with the existence of a single center responsible for the complete sleep-wake transition (47). For example, there is evidence that the "strength" of arousal helps to establish overall steady-state sleep requirements (48⇓ ⇓ -51), and the balance between internal and external arousal promotion and sleep promotion signals ultimately determines the probability of state switching And degree (3, 52).

Here, we studied the role of the hypothalamus in the two-way interaction between sleep-wake transition, wakefulness and sleep homeostasis. First, we performed optogenetic stimulation on glutamate decarboxylase 2 (GAD2) neurons in the lateral preoptic region (LPO) of mice (17) and found that the light activation of LPO during sleep leads to rapid arousal induction, but in the structure The surrounding LPO is irritated. Unexpectedly, GAD2LPO neuron stimulation not only triggers wakefulness, but the wakefulness produced by this stimulation is characterized by increased EEG theta activity—an established measure of arousal (53, 54). In turn, subsequent sleep is associated with elevated EEG SWA levels, indicating a higher steady-state sleep pressure (45). In contrast, unilateral inhibition of GAD2LPO neurons reduces the drive to arousal, which is reflected in the continuous increase in non-rapid eye movement (NREM) EEG SWA throughout the day. In summary, our experiments proved that GAD2LPO neurons play an important role not only in controlling state transitions but also in linking arousal to sleep homeostasis. We found that the response kinetics to photoactivation and photoinhibition are different, so they may come from different mechanisms, and at the same time converge on the dynamic adjustment of the arousal level, and finally shape the daily structure of the sleep-wake state.

We used the previously described protocol to optically stimulate the GAD2-expressing neurons in the hypothalamic LPO, resulting in a rapid transition from sleep to wakefulness (17). In order to achieve the expression of channel rhodopsin 2 (ChR2), we injected adeno-associated virus (AAV) fused with cre-dependent expression of ChR2 and enhanced yellow fluorescent protein (AAV-DIO-ChR2-EYFP) and implanted Targeting the preoptic hypothalamus of GAD2-cre mice (n = 14) (Figure 1A and SI appendix, Figure S1 A and B). Subsequent histology confirmed the expression of the virus in a wide range of areas, including the preoptic and adjacent areas of the hypothalamus. Specifically, of the 14 mice expressing the virus, 8 had the fiber tip located in the LPO (LPO group), while in the remaining 6 mice, the fiber tip was located in the other adjacent hypothalamic area behind the LPO. LPO (non-LPO group) (Figure 1B) (55). Since light stimulation may affect the sleep and behavior of animals due to local heating (56) or direct effects of light (57), we also analyzed a group of animals as a control group, and these animals received Cre-dependent AAV injections Express enhanced green fluorescent protein (AAV-DIO-EGFP) and implant optical fibers in the same area (GFP group). Electroencephalography and electromyography (EMG) electrodes are also surgically implanted, as previously described (58), to identify the sleep-wake state of the animal.

Light activation of GAD2 neurons in the LPO and surrounding hypothalamic area induces rapid awakening. (A) (Top) Schematic diagram of the implant. (Bottom) Representative brain slices from animals in the LPO group and corresponding illustrations from the mouse brain atlas (55). The dotted line depicts the position of the fiber tip. (B) Schematic diagram of the position of the fiber tip of all animals expressing ChR2. The center of the fiber tip is shown as a dot, and the estimated stimulus coverage area based on the fiber diameter (400 µm) is shown as a circle of a single animal. Blue, LPO group; green, non-LPO group. The upper right illustration shows a three-dimensional map of the region of interest (ROI) hypothalamic region constructed by Allen Brain Explorer (beta). The LPO is shown in green, and the other hypothalamic nuclei in the ROI are shown in red. (C) Representative EEG spectrogram and corresponding hypnosis diagram, EMG and EEG SWA. Blue shade, light stimulation. Frequency, frequency. Color scale: the spectral power in the common logarithmic value. Hypnogram and SWA are color-coded according to the alert status (blue, awakened; green, NREM; yellow, REM). (D) EEG and EMG traces in a typical light stimulation test in a single animal. From top to bottom: frontal lobe EEG, occipital lobe EEG, EMG and light stimulation time. (EG) are displayed in the LPO area and outside the LPO area (non-LPO) and GFP control. Blue shade, light stimulation. Mean ± SEM. (H) Delayed awakening. Data points represent individual mice. The box represents the 25/75 percentile, with the middle value in red. The P value is calculated by the non-parametric two-sided Wilcoxon signed rank test. (I) The representative EMG variance distribution (median ± 25/75 percentile) averaged relative to the start of stimulation of one mouse. (J) Delayed awakening of stimuli delivered during spontaneous NREM sleep and REM sleep in LPO. The data points represent individual mice, the boxes represent the 25/75 percentiles of mice, and the red bars represent the median. **P = 0.007, Wilcoxon signed rank test on both sides. LPO, n = 8; non-LPO, n = 6; GFP control, n = 8. 3V, third ventricle; ac, anterior commissure; AH, anterior hypothalamus; AP, anterior and posterior; BST, striatal bed nucleus DV, dorsoventral; HDB, diagonal horizontal limb nucleus; LH, lateral hypothalamus; ML, middle; MPA: medial preoptic area; SI, unnamed substance; VLPO, ventrolateral preoptic nucleus; VMH, hypothalamus Venomedial nucleus; ZT, Zeitgeber time.

In the first set of experiments, we performed photoactivation in the same way as Chung et al. (16), including a 2 minute light pulse sequence provided at a frequency of 10 Hz every 20 minutes for a full 24 hours of the day, regardless of the behavioral state (Figure 1 C and D and SI appendix, Figure S1C). The 2-minute stimulation period of GAD2-ChR2 animals implanted with two LPOs was completely dominated by awakeness [Figure. 1E, movie S1; the probability of arousal in the second minute of light activation: stimulus, 99.9 ± 0.1%; sham, 45.3±2.0%; P <0.0001 in paired t test] and non-LPO position (Figure 1F Probability of arousal during stimulation: 99.1 ± 0.5%, sham operation: 50.8 ± 2.2%, P <0.0001), but not in the GFP control (Figure 1G); probability of arousal during stimulation: 50.7 ± 2.35, sham operation: 48.1 ± 3.0% , P> 0.1). The average wake-up delay time calculated based on the EMG level in the LPO group (Figure 1D and SI Appendix, Figure S1D) averaged 4.8 seconds from the start of stimulation (±0.94 SEM, n = 8), and in the non-LPO group (10.3 ± 3.4 s, n = 6, Figure 1H, P = 0.059, U = 61, two-tailed Wilcoxon rank sum test). Interestingly, when stimuli were provided during REM sleep, compared with NREM sleep, LPO (12.58 ± 1.9 s; P = 0.007; Figure 1 I and J) and non-LPO group (SI appendix, Figure S1E) delayed wakefulness ). If the animal is stimulated when the animal is already awake, the sleep latency after stopping the stimulation will be significantly prolonged (SI appendix, Figure S1F; arousal: 9.05 ± 0.59 minutes, NREM: 4.47 ± 0.42 minutes, REM: 5.22 ± 0.70 minutes, P < 0.0001, one-way analysis of variance), and compared with the LPO group in the dark period, the non-LPO group took longer (SI appendix, Figure S1G). Therefore, stimulation of hypothalamic GAD2 neurons during wakefulness will not induce sleep, but will lead to sustained wakefulness.

Two recent observations of the hypothalamic arousal circuit may indicate the underlying cellular mechanisms we observe. First, high-frequency stimulation can cause some hypothalamic cells to enter the conduction block, thereby transforming optogenetic activation into de facto neuron inhibition (25). Second, although the GAD2 promoter is mainly targeted at suppressor cells, studies have shown that GAD2 cells can also be excitatory (59).

In order to assess whether conduction block occurs only at high stimulation frequencies, 1-Hz, 2-Hz, and 5-Hz stimulations were applied to a group of animals in sequence, and were compared with the arousal effects observed after stimulation frequencies of 10 Hz and 20 Hz. compare. SI appendix, Figure S2 AC). This indicates that the stimulation frequency has a significant effect on the probability of arousal (SI appendix, Figure S2D; repeated measures analysis of variance [RM-ANOVA], P <0.01, n = 7) and the delay between the start of stimulation and arousal (SI appendix , Figure S2E; RM-ANOVA, P <0.001, n = 7). However, compared to higher frequency stimuli, frequencies below 10 Hz do not seem to cause the effect to be reversed. On the contrary, the arousal promotion effect depends on the stimulation frequency (SI appendix, Figure S2 AC). Consistent with this, the latency of arousal is significantly affected by the stimulation frequency (SI appendix, Figure S2E; P <0.05 Wilcoxon signed rank test, n = 7). Similarly, the probability of arousal occurring within 2 minutes after the start of the stimulation is 75.0% (±5.1) at 2 Hz, increases to 90.6% at 5 Hz, and reaches a peak of 100% at 10 Hz (SI Appendix, Figure S2D; P <0.05 Wilcoxon signed order test, n = 7). Therefore, although it cannot be ruled out that some neurons transfected by ChR2 in our study may enter conduction block, there is no response of different nature after high-frequency stimulation, indicating that it may only affect a few cells. To support this conclusion, galaninergic neurons in LPO (which have been previously shown to block conduction) only form a few LPO neurons (25).

Intracellular current clamp recordings from neurons in acute brain slices of GAD2LPO mice were used to further characterize the cell matrix behind the observed behavioral effects. Since this test is designed to evaluate the light response of ChR2 expressing cells and their postsynaptic targets, cells in LPO are targeted regardless of ChR2 expression. As expected, a large proportion of cells respond to negative current steps with hyperpolarization-induced droop and subsequent rebound low threshold spikes (LTS) and hyperpolarization-induced inward currents (47%, n = 85 cells; Figure 2A, B and F). Although almost all cells showed a spike or subthreshold response to light (90.6%), the delay between stimulation and response distinguished cells that were directly activated by ChR2 (<1 ms delay, 58%; Figure 2B) and those that received ChR2. The cell that touches the input is activated. In terms of the proportion of cells directly activated by ChR2, no difference was observed between LTS and non-LTS cells [Figure. 2B, bottom; χ2 (1, 3) = 1.22, P = 0.75, χ2 test]. Next, we directly tested whether 10 Hz GAD2LPO stimulation is related to conduction block and neuronal silencing. In the presumed ChR2-expressing cells, there was no sign of conduction block at 10 Hz during 2 minutes of stimulation, and the spike rate of high-fidelity light pulse entrainment was independent of the stimulation frequency (Figure 2C and D; P = 0.70, one-way analysis of variance). In fact, the average spike rate during stimulation was slightly higher than the stimulation frequency (Figure 2E). These results are consistent with our behavioral observations that the 10 Hz in vivo stimulation effect is similar in nature to the lower frequency effect.

Single-cell recordings in acute brain slices of GAD2LPO mice. (A) Representative electrophysiological characteristics of LTS neurons in the preoptic zone. (B) (Upper left corner) Example of membrane potential (Vm) response to ChR2 activation in a hypothetical ChR2 positive cell (zoom in trace shown in Figure C). Note that the depolarization starts immediately when the lighting starts. (Upper right) The histogram of the delay between the light stimulus and the Vm response, showing a bimodal distribution; a response with a delay of <1 ms is classified as ChR2+, and a response with a delay of >1 ms is classified as a synaptic response. (Bottom) Classification of cells based on ChR2 response characteristics. (C) Example trace of hypothetical ChR2-positive cells responding to a 10 Hz stimulus. (D) The hypothetical spike fidelity of ChR2 positive neurons, expressed as the ratio of light pulses followed by spikes. (E) The relationship between the stimulation frequency and the average spike rate during the stimulation of putative ChR2 positive neurons. (F) Representative electrophysiological characteristics of non-LTS neurons in the preoptic zone. (G) Representative trace and average evoked potential of ChR2 negative neurons (top right), showing polarized light response (bottom trace and bottom average evoked potential) at -70 mV and at slightly depolarized Vm Exhibits hyperpolarized response (upper trace and average evoked response in the middle). The highest average evoked potential showed the light response of bicuculline in the presence of a depolarizing membrane potential. Pay attention to the disclosure of excitement. (H) Changes in the average evoked response (under depolarized Vm) in response to the blocking of ionotropic GABA receptors by bicuculline. *P = 0.02, t = 3.34, paired t test. (I) In response to CNQX's blocking of ionotropic glutamate receptors, changes in the average evoked response (at rest Vm). *P = 0.04, t = -2.78, paired t test. (J) Changes in spontaneous spike rate (caused by injection of depolarizing current) during 2 minutes of light stimulation at 1, 2, 5, and 10 Hz. ** Indicates that the spike rate of the ChR2 group increased significantly [F(1,53) = 9.46, P = 0.003] but the stimulation frequency [F(3,53) = 0.74, P = 0.53) or the interaction [F] was not significant The impact (3,53) = 1.13, P = 0.35; two-way analysis of variance]. BIC, bicuculline.

Although the GAD2-Cre line was selected for ChR2 expression in inhibitory neurons, it has been suggested that GAD2 can also be expressed in excitatory neurons (59). In order to examine the post-synaptic effect of stimulating GAD2-ChR2 cells, the light response to 0.5 Hz stimulation was analyzed in the presumed non-ChR2 positive cells (cells with subthreshold light response and delayed onset> 1 ms) in the LPO. In order to distinguish GABAergic and glutamatergic inputs, light-evoked responses were recorded at different membrane potentials (Figure 2G). At rest, non-ChR2 neurons respond very little to light and depolarize potentials, which become hyperpolarized potentials at higher membrane potentials (Figure 2G). This indicates that light induces a strong GABAergic response, and the glutamatergic component is weaker. Consistent with this, pharmacological inhibition of the GABA-A receptor (10 µM bicuculline) eliminated hyperpolarization potentials and unmasked excitatory responses (Figure 2H; n = 6, P = 0.02, paired t-test). In contrast, the excitatory response is sensitive to blocking ionotropic glutamate receptors with 10 µM 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX), although this effect is weak (Figure 2I; n = 6, P = 0.04, paired t test). Strong inhibitory and weak excitatory currents indicate that GAD2LPO stimulation has an overall inhibitory effect. This explanation is tested by increasing the membrane potential of ChR2-negative cells to promote spontaneous discharge, and then providing 2 minutes of light stimulation at 1, 2, 5, and 10 Hz, mimicking our in vivo protocol. The spontaneous firing rate of ChR2-negative cells does not significantly depend on the stimulation frequency (P = 0.43, RM-ANOVA), although the spontaneous firing rate is suppressed by stimulation [T(35) = -3.58, P = 0.001, a sample t is 0 Perform inspection; Figure 2J, yellow trace]. But please note that our sample size may be too small to adequately resolve frequency dependence. In contrast, hypothetical ChR2 positive cells increased their firing rate in the same experimental paradigm (Figure 2J, blue trace). Therefore, the simplest explanation for our in vivo and brain slice experiments is that the effects of GAD2LPO stimulation reported in this study are mainly driven by the activation of inhibitory neurons and are not affected by conduction block.

In order to further characterize the arousal promotion effect of GAD2 neurons in the preoptic hypothalamus, we performed light stimulation with dexmedetomidine (Dex) during sedation, which is an α2-adrenergic Receptor agonists are known to induce "deep sleep"-like EEG activity (SI appendix, Figure S3A) and decrease in body temperature and metabolic rate (SI appendix, Figure S3 BD) (20, 26). The initial and late stages of Dex sedation were analyzed every 15 minutes at a frequency of 10 Hz for 2 minutes, corresponding to the first four stimuli (0 to 1 hour) and the stimuli between 2 to 3 hours. Inject Dex separately. Within 5 minutes after the injection of Dex, the animal was sedated—not moving, usually lying prone outside the nest—and in all cases, the EEG was dominated by high-amplitude slow waves (SI appendix, Figure S3A) The peripheral body temperature and metabolic rate decreased at the same time (Figure 3A and SI appendix, Figure S3 BD). During the initial stage of sedation, rapid awakening was observed in the LPO and non-LPO stimulation groups (Figure 3B and C), while light stimulation was ineffective in GFP control animals (Figure 3B). Since we have never observed stimulation-induced arousal in control animals, the incubation period cannot be calculated, so it is not reported here. Interestingly, at the beginning of the stimulation but before the animals woke up, the decrease in EEG SWA relative to the pre-stimulation level was only observed in the LPO group (SI appendix, Figure S3 E and F). In addition, compared with non-LPO-stimulated animals, the latency of waking up from sedation in the LPO group tended to be shorter (median LPO was 36 seconds, non-LPO was 80 seconds; P = 0.092, Welch t test; Figure 3C) . It is worth noting that although compared with spontaneous NREM sleep, the latency of wakefulness during stimulation under sedation is still significantly higher (SI Appendix, Figure S3G; NREM: 5.3 ± 1.6 seconds, SEM, Dex: 31.4 ± 5.5 seconds, n = 6, P <0.05, two-tailed paired t-test), the animal showed basically normal behavior after waking up, including the exploration of new objects (movie S2). Once the 2-minute stimulus is terminated, the animal will return to a sedative state within a few minutes (Figure 3B) and re-establish a high-amplitude SWA (Figure 3A). Unexpectedly, during late sedation-2 to 3 hours after Dex injection, when the peripheral body temperature drops to near room temperature and the EEG amplitude is very low-the stimulation still leads to rapid awakening (Figure 3C and SI appendix, figure) S3H), and compared with the early test after Dex injection, the wakefulness latency was significantly shorter (P <0.0021, two-way analysis of variance, factor "0 to 1 hour and 2 to 3 hours", P = 0.0021; post-hoc test, LPO: P = 0.0273, non-LPO: P = 0.0174).

The effect of light activation of GAD2LPO neurons during sedation and HSP. (A) Representative frontal lobe EEG spectrogram, sleep map, peripheral body temperature, EMG and SWA (EEG power between 0.5 and 4 Hz) are color-coded according to the alert state, from a representative mouse, injected at Dex Before and after. Color scale: the spectral power in the common logarithmic value. SWA is color-coded according to alert status (blue, awake; green, NREM; yellow, REM; purple, sedation). (B) In the first 1-hour interval after Dex injection, the percentage of awake time before, during, and after light stimulation (shaded area), averaged LPO (n = 7), non-LPO (n = 6), and GFP (n = 8) Animals. Mean ± SEM. Please note that in GFP animals (grey, arrow), it is very rare that a similar arousal state occurs before or during the stimulation. (C) Calculate the initial sedation (the average of the four stimulation sessions within the first hour after Dex injection) and the late sedation period at low temperature (the average of the four stimuli provided between 2 and 3 hours) after the start of stimulation The incubation period of awakening from sedation after Dex injection) in LPO and non-LPO animals. Note that the latency of arousal from GFP sedation is not shown, as stimulation does not cause arousal. Red bar, median; box: ±25/75 percentile. P value, Welch t test. (D) Representative 12-hour curves of EEG SWA and EMG variance displayed for "low" (top) and "high" (bottom) sleep pressure conditions, respectively. SWA is color-coded according to the alert status (blue, awakened; green, NREM; yellow, REM). Stimulus, light stimulation. (E) The average EEG power spectrum during the first 4 hours of resuming sleep after SD and the corresponding baseline interval when sleep pressure (LSP) is low. Mean ± SEM. (Inset) Amplified spectrum below 5 Hz. The black bar at the bottom of the graph represents the frequency interval, in which the difference between LSP and HSP is significant in the post hoc uncorrected Fisher's least significant difference (LSD) test after one-way analysis of variance (P <0.05). (F) Under HSP and LSP conditions, the percentage of NREM sleep that occurred before, during, and after 2 minutes of light stimulation (shaded area). Mean ± SEM. (G) Average delay in wakefulness from stimuli provided under HSP and LSP conditions. Data points represent individual mice, and red bars represent the median of mice. ns, there is no significance in the Wilcoxon signed rank test on both sides. Box, ±25/75 percentile. Number of animals used in EG: LPO, n=8.

We next hypothesized that increased sleep stress levels would reduce the likelihood of waking up after light stimulation. To this end, mice were stimulated after a 4-hour sleep deprivation (SD) or undisturbed sleep period (Figure 3D). As expected, SWA during NREM sleep during the 4-hour recovery period after SD was higher than the SWA level during the corresponding time period after undisturbed sleep (Figure 3E). Therefore, light stimulation during sleep after SD or after an undisturbed sleep period is referred to as "high sleep pressure" (HSP) and "low sleep pressure" (LSP) conditions, respectively. Unexpectedly, once the stimulation was initiated under HSP, the animals still woke up almost immediately (Figure 3F), and the arousal latency between HSP and LSP conditions was indistinguishable (Figure 3G; P> 0.1, Wilcoxon signed rank test on both sides) ; SI appendix, Figure S4A). The absence of sleep stress is surprising because the SWA level during NREM sleep after SD is approximately 70% higher than the level during baseline sleep (Figure 3E). We further compared the latency from stimulus excursion to sleep, but there was no significant difference between HSP and LSP (SI Appendix, Figure S4B), although overall SWA was still high during HSP despite regular sleep interruptions ( SI appendix, Figure S4C). In summary, these results indicate that regardless of the steady-state sleep pressure, the activation of a large number of inhibitory neurons in the hypothalamus will induce awakening, but LPO activation will not eliminate the accumulated sleep pressure.

The arousal caused by light activation of GAD2 neurons is usually 4.69 ± 0.42 and 4.85 ± 0.54 minutes longer than the stimulation duration of LPO and non-LPO areas, respectively (P = 0.82, Welch t test). Given that the mice were awake most of the time during the stimulation, and since a total of 72 stimulations were performed in 24 hours, we expected to observe a significant lack of sleep. In fact, in LPO mice (86.3 ± 3%) and non-LPO mice (86.4 ± 3.3%), the amount of sleep on the stimulation day was lower than the corresponding value on the baseline day, but in the GFP group (100.4 ± 5.6%, Fig. 4A; RM-ANOVA, P = 0.035, t test for LPO 100%: P = 0.002, n = 8, non-LPO: P = 0.01, n = 6, GFP: P = 0.7, n = 8). However, although the inhibition of NREM sleep during the stimulation period occurred in the LPO and non-LPO groups (Figure 4B) and the total amount was also reduced during the light period (Figure 4A), we observed that the lack of NREM sleep remained only in the LPO at the end of 24 hours. In the stimulus group (SI appendix, Figure S5A). In addition, compared with non-rapid eye movement sleep, REM sleep is more affected by stimulation; in LPO and non-LPO (Figure 4A and SI appendix, Figure S5 A and B), the total amount during the light period is reduced, However, the proportion of REM sleep per total sleep time within 24 hours was significantly reduced only in the LPO group (Figure 4A; RM-ANOVA, P = 0.012, LPO post-test: P = 0.048, non-LPO: P = 0.098, GFP: P = 0.067). These observed changes in the amount of NREM and REM sleep fit an explanation, because repeated light activation always performs NREM sleep first, so REM sleep is particularly limited by our experimental paradigm.

During NREM sleep in GAD2LPO, light activation induces a rebound in SWA, but not in GAD2nonLPO animals. (A) The effect of light stimulus on the total amount of alertness during the light period is shown as a percentage change relative to the false stimulus day. The asterisk above the line indicates the significant difference of RM-ANOVA, and the asterisk and ns above the graph indicate the significance of the t-test to 100%. *P <0.05; ns, meaningless. Mean ± SEM. (B) The amount of sleep aligned with the stimulus offset within the 20-minute window, shown as a percentage of the average amount of sleep during the sham stimulation day. All experiments that occurred during the light period were averaged for LPO and non-LPO animals. Using two-way analysis of variance (P> 0.1) and the unadjusted Fisher LSD test after the fact, no significant difference was observed in the amount of sleep between the two groups. Mean ± SEM. (C) A representative 24-hour profile of the EEG SWA and EMG of an individual animal receiving LPO stimulation. SWA is drawn with a resolution of 4 seconds and is color-coded according to the alert status (green, NREM; blue, awakened; yellow, REM). Note the gradual increase in SWA during the 12-hour photoperiod (indicated by the white bar at the top). (D) Relative EEG power density during NREM sleep, shown as a 10-Hz 24-hour stimulus condition compared to a sham condition during the light period. Mean ± SEM. The black bar at the bottom of the figure represents the frequency interval, in which the difference between the experimental groups is significant (P <0.05, unpaired t-test). The shaded area indicates the SWA frequency range (0.5 to 4 Hz). (E) The time course of the EEG SWA within the 20-minute window is aligned with the stimulus offset (time 0) during the illumination period. SWA is shown as the percentage of SWA on false stimulation days during the light period. *P <0.05, **P <0.01, Fisher's LSD test was not adjusted for post hoc after RM-ANOVA. Mean ± SEM. (F) The time course of EEG SWA during NREM sleep within 24 hours of the day, with light stimulation. SWA is drawn at 2-hour intervals and expressed as a percentage of the average NREM SWA during the sham stimulation day (mean ± SEM; * or #P <0.05, ** P <0.01, Tukey multiple comparison test after RM-ANOVA) . * LPO and GFP; #LPO and non-LPO. (G) Cumulative EEG SWE within 24 hours. The asterisk at the top of the graph indicates the significance of the multiple paired t-test. ** P <0.01. The comparison of the SWE slopes of LPO and non-LPO during illumination is shown in the SI appendix, Figure S5D. Mean ± SEM. Number of animals in A: LPO, n = 8; non-LPO, n = 6; GFP, n = 8. Number of animals in DF: LPO, n = 8; non-LPO, n = 6; green fluorescent protein, n = 7.

Insufficient sleep can be compensated not only by the increase in sleep time, but also by changes in sleep intensity, as measured by the level of EEG SWA. Strikingly, although sleep is often fragmented, we noticed that the NREM sleep between stimulation sessions is characterized by increased SWA in the light and dark periods in the LPO group, which is not observed in the non-LPO group or the GFP control (Figure 4 CE and SI appendix, Figure S5C and S6 C and D). Since the non-LPO and LPO mice lost the same amount of sleep during the light period, but the non-LPO group partially recovered the loss in the subsequent dark phase, we next evaluated the distribution of NREM sleep SWA over the entire 24 hours. The time course of SWA in a day usually follows a U-shaped pattern, with SWA being high at the beginning and decreasing in the middle of 24 hours (Figure 4F). Unexpectedly, we found that there was no such pattern in the LPO group, but showed a continuous increase in the average SWA, and in some individuals, there was a trend of increasing SWA in the mild stage (Figure 4C and F). This change in the daily time of SWA was not observed in non-LPO or GFP control mice (Figure 4F; two-way analysis of variance, P = 0.0065; Sidak's multiple comparison test on LPO after the event, P = 0.0006; non-LPO, P > 0.9999; GFP, P = 0.9804). This is particularly interesting because the daily distribution of awakening and NREM amounts between LPO and non-LPO mice is comparable (SI appendix, Figure S5B). Consistently, compared to baseline or non-LPO mice, the increase in accumulated slow wave energy (SWE) in LPO mice during a day is characterized by a steeper slope during the stimulation day (Figure 4G and SI appendix, Figure S5D; both- way ANOVA, for the interaction of factors, P = 0.0406). Taken together, these results indicate that stimulation of GAD2 neurons in LPO not only leads to an increase in arousal, but the induced arousal state is characterized by faster accumulation or overall higher sleep stress levels that need to be compensated.

Previous studies have shown that the accumulation of sleep stress depends on the type of wakeful behavior, not just the duration of wakefulness (48⇓ –50, 60, 61). Therefore, we hypothesize that LPO and non-LPO stimulation have different effects on brain activity during wakefulness. To test this hypothesis, the EEG patterns obtained after spontaneous awakening during the baseline period were compared with the first EEG patterns during the awakening period caused by the stimulus. In LPO and non-LPO mice, stimulus-induced arousal is associated with a decrease in frontal lobe EEG power in the slow-wave range [Figure. 5 A and B; general linear mixed model, P <0.001, frequency main effect F(97) = 6.3, group and frequency × group P> 0.5; see Figure 5 for the post-event test]. However, compared with the non-LPO group, the occipital EEG θ frequency (6 to 9 Hz) power increased more in the LPO group, which is also a characteristic of stimulation-induced arousal [Figure. 5 A-C; general linear mixed model, P <0.01, F(97) = 1.91 frequency × group; see Figure 5B, post-hoc rank sum test for frontal and occipital lobe EEG]. Therefore, light activation of the LPO region not only triggers arousal, but the resulting state of arousal is also characterized by increased levels of arousal and alertness. We next hypothesized that this altered state of arousal might explain the difference in sleep pressure between LPO and non-LPO mice (Figure 4). To support this hypothesis, among animals, a strong positive correlation was observed between the effect of stimulation on the occipital theta frequency power (6 to 9 Hz) during wakefulness and the effect of stimulation on SWA (0.5 to 4 Hz) during NREM. Sleep in the LPO group (Figure 5C, lower; Pearson's R = 0.95, P = 0.0011). In the non-LPO group, there was only an insignificant weak positive correlation (R=0.42, P>0.1; the correlation coefficient difference between LPO and non-LPO, P<0.05), and there was no significant correlation between GFP animals (R = -0.05, P>0.1 ). In LPO, non-LPO or GFP animals, there was no significant correlation between the changes in frontal lobe SWA caused by stimulation during wakefulness and SWA during NREM sleep (Figure 5C, top panel; Pearson's R: LPO, 0.49; non-LPO, 0.34 : GFP, 0.06; all P values> 0.1), and LPO [T(4) = 3.28, P <0.05, the difference between the correlation coefficients] or the combination of LPO and non-LPO groups [T(10) = 3.06, P <0.01].

The effect of LPO and non-LPO light activation on wakefulness and subsequent sleep. (A) Average arousal EEG spectra (top: LPO, bottom: non-LPO) recorded from the occipital bones of representative mice during spontaneous wakefulness on the baseline day and during light-activated wakefulness. Time 0 corresponds to the beginning of waking up. Color scale: the spectral power in the common logarithmic value. (B) The average EEG spectral power density during light activation-induced wakefulness, expressed as the percentage of power during spontaneous wakefulness during the baseline period. The bar at the bottom represents the significant difference of the post-rank sum test following the general linear mixed model (P <0.05); black, there is a significant difference between the LPO group and the non-LPO group; gray, there is no between the LPO and non-LPO group Significant difference, but the combination of LPO and non-LPO group is significantly different from 100%. Double diagonal line: The cut-off frequency of artifacts caused by stimulation at around 10 Hz and 20 Hz, from the frequency where the power increases sharply to the frequency where the power drops sharply. (C) The correlation between awakening SWA and NREM SWA (top) and the stimulus-related differences in the arousal theta frequency (6 to 9 Hz) and NREM SWA (bottom). R: Pearson correlation coefficient, with corresponding P value. Please note that only the correlation between wake-up theta frequency power and NREM SWA in the LPO group is statistically significant. (D) A representative hypnotic diagram illustrating the experimental design of stimulation during SD. (Top) 2-hour SD without stimulation (SD only). (Bottom) The 2-hour SD combined with light stimulation is shown as a blue bar "stimulus". The SWA drawn for the 4 s period is color-coded according to the alert status. The y-axis, the unit is μV2/0.25 Hz. (E) The effect of stimulation during wakefulness on the EEG spectrum during subsequent NREM sleep. The EEG power after SD + stimulation was calculated within the first 2 hours of resuming sleep and expressed as a percentage of "SD only" conditions in LPO and GFP animals. The bar at the bottom represents the frequency range where EEG power is significantly affected by the stimulus (P <0.05, t test); black, which is significant between GFP and LPO; gray, which is significant compared to 100% in LPO. Number of animals used in B and C: LPO, n = 7; non-LPO, n = 6; GFP, n = 6. Number of animals used in D and E: LPO, n=6; green fluorescent protein, n=7.

In order to directly solve whether GAD2LPO stimulation affects sleep homeostasis by adjusting the level of wakefulness during wakefulness, we then perform light stimulation while the animal is kept awake by the new object (Figure 5D; 20 seconds 10 Hz training, transmitted every 2 minutes ). We observed that compared with non-stimulated conditions (SD only), under SD + LPO light-activated conditions, the EEG SWA during subsequent NREM sleep increased, but not in the GFP control [Figure. 5E; LPO: P = 0.0312, n = 6, GFP: P = 0.6875, n = 7, Wilcoxon signed rank test of average SWA (0.5 to 4 Hz); LPO and GFP: P = 0.0290, unpaired t of average SWA test]. Photoactivation has a moderate effect on wakefulness and sleep after the end of SD, which is mainly reflected in the increase in wakefulness in the first 30 minutes after the animal is undisturbed, and this effect dissipates quickly (SI appendix, Figure S7). In order to examine whether the prolonged wakefulness induced by light activation after SD directly leads to the subsequent increase in SWA, we calculated the correlation between the difference in total arousal between the two conditions and the corresponding increase in SWA. Contrary to our hypothesis, no meaningful relationship was observed (SI Appendix, Figure S7D), which indicates that the difference in SWA during sleep recovery after GAD2LPO activation is more likely to be caused by arousal level (arousal "quality") rather than arousal duration.

In order to further evaluate whether long-term light activation will cause a further increase in brain activity during wakefulness or subsequent sleep, some animals were exposed to a long-term strong stimulus of 10 Hz, starting from light or dark, with an interval of 1 hour ( SI appendix, Figure S8). Unexpectedly, even though the animals are behaviorally awake, brain activity and EMG indicate a clear waking state, three-quarters of the animals showed a significant decrease in peripheral body temperature (SI appendix, Figure S8E), one of them This condition is about 4 hours longer than the duration of the stimulation period (SI Appendix, Figure S8B).

Finally, we started to investigate whether the photoinhibition of GAD2 neurons in the LPO area would lead to changes in sleep-wake structure and sleep intensity. Specifically, we predict that if the light activation of these neurons increases arousal drive, which in turn is related to increased sleep pressure, the result of inhibiting inhibitory neuronal activity in the LPO area should result in decreased arousal. To this end, we conducted an experiment in which green light-emitting diode (LED) stimulation (550 nm, 18.8 to 24.0 mW at the fiber tip) was applied to Arch-expressing mice (n = 6) and the LPO area of ​​the GFP control ( n = 4) (Figure 6A). First, in order to examine the effect of stimulation on wakefulness and subsequent sleep, a continuous stimulus was applied for 2 minutes every 5 minutes during SD and dropped at the excursion of the stimulus, as we did in the photoactivation experiment (Figure 5D). The intensity of the light stimulus is gradually reduced for about 30 seconds before the stimulus is cancelled to prevent the burst activity at the end of the pulse, thereby minimizing the explosive excitement of the neuron (62). The stimulation was stopped 10 minutes before the end of the 2-hour SD, and the animals received a total of 22 stimulations during the 2-hour continuous awake period. Compared with SD only, we only observed small changes in EEG during SD + stimulation (Figure 6 B and C), and as expected, there was no difference in subsequent sleep under different conditions. We hypothesized that the relatively weak one-sided photoinhibition during SD is insufficient to cover the high-level arousal of animals that continue to provide new objects and engage in exploratory behaviors.

Optical inhibition of GAD2LPO neurons does not change sleep pressure accumulation during SD, but increases SWA and SWE when transmitted during spontaneous sleep-wake state. (A) Representative histology of animals with Arch expression in LPO used for optical suppression. (Upper left) Schematic diagram of the optical suppression experiment. (Upper right) Schematic diagram of the position of the fiber tip of all animals with Arch expression. The center of each fiber tip is shown as a dot, and the estimated stimulus coverage area based on the fiber diameter (400 µm) is shown as a circle of a single animal. (Bottom left) DAPI image and corresponding atlas (55) and implantation diagram. The dashed square corresponds to the right panel. (Bottom right) The fluorescence image of the squared area in the left panel and the corresponding atlas (Bregma = 0.02). The dotted line depicts the pathology of the implanted fiber. (B and C) Average 2-hour arousal spectra during SD+ stimulation versus 2-hour SD. Frontal lobe EEG (B) and occipital of the arch and GFP control animals (n = 6 and n = 4) only in conditions of SD Electroencephalogram (C), respectively; mean±SEM). (D) The effect of stimulation during SD on the EEG spectrum during subsequent NREM sleep. The EEG power after SD + stimulation was calculated within the first 2 hours of resuming sleep and expressed as a percentage of SD only conditions. The bar at the bottom represents the frequency range where the EEG power is significantly affected by the stimulus (P <0.05). (E) The probability of waking up, NREM and REM sleep (shown as green bars and shaded areas) before, during, and after 5 minutes of photoinhibition are shown for the GFP control and Arch group, respectively. Note that the reflected light at the right end of the green bar drops. Mean ± SEM. (F) The effect of photoinhibition on the total amount of alert status during the photoperiod, shown as a percentage change relative to the false stimulation day. The ns above the line indicates that there is no difference in the unpaired t-test. The asterisk and ns in the figure indicate the significance level for 100% in the one-sample t-test. * P <0.05, ** P <0.01. Mean ± SEM. (G) The average EEG power spectrum during NREM sleep that day, and the photoinhibition is shown as a percentage of the baseline day. Mean ± SEM. The bar under the curve depicts the frequency interval, where the EEG power between GFP and Arch animals is significantly different (P <0.05). (H) Cumulative EEG SWE during NREM sleep in 24 hours, showing GFP and Arch animals under false and stimulating conditions, respectively. The asterisk at the top of the graph indicates the significance of the multiple paired t-test. * P <0.05, ** P <0.01, *** P <0.01. The comparison of the SWE slope during illumination between LPO and non-LPO is shown in the SI appendix, Figure S9D.

In order to solve the problem of photoinhibition in undisturbed animals that were not stimulated by new objects, we then performed repeated stimulation for 24 hours, including continuous pulses for 5 minutes every 30 minutes. Regardless of the alert state, the jitter was 10%. A total of 48 stimuli were provided. No direct effects of inhibition on state transitions were observed (Figure 6E and SI Appendix, Figure S9A), and no changes in total sleep and NREM sleep during the light period were found, while REM sleep was slightly reduced (Figure 6). 6F). In contrast, the EEG spectrum power during NREM sleep is characterized by higher SWA levels (Figure 6G and SI appendix, Figure S9B). In contrast to photoactivation (Figure 4F), the SWA time course of photoinhibition during the day is completely normal (SI appendix, Figure S9C). We speculate that in this case, the continued increase in SWA is due to the overall weakening of arousal drive rather than the gradual increase in sleep debt. By the end of 24 hours, SWE, which is a comprehensive indicator of sleep time and intensity, was about 20% higher in Arch-expressing animals, which was significantly higher than under false stimulation conditions or GFP control animals (Figure 6H and SI appendix, Figure S9D) .

Current research reports that the light activation of inhibitory neurons in the lateral preoptic region of the hypothalamus not only triggers the transition from sleep to wakefulness, but also increases the EEG index of arousal. In turn, the high arousal during the waking state is associated with a greater increase in sleep SWA, which indicates that LPO not only controls global state switching, but also helps to regulate brain activity and steady-state sleep-driven intra-state dynamics. The photoinhibition of the same neuron further supports this view because it continuously reduces arousal levels without switching the global state. These findings indicate that the level of arousal is higher than and beyond the duration of wakefulness, which is an important parameter that contributes to the accumulation rate of sleep requirements. Based on these findings, we propose to add a qualitative dimension to the influential sleep-wake control model (37), which is best expressed as a built-in "spring switch". We assume that when awakeness is particularly intense, "spring" becomes more compressed. As a result, once the awakening is over, there will be a stronger rebound in sleep to release the "stress."

Our observations of rapid awakening at the onset of GAD2LPO photoactivation confirm the early findings (17) and are consistent with the view that only a subset of preoptic neurons are sleep-active or sleep-promoting (25, 63). Previous studies have sometimes assumed that the same hypothalamic area is related to sleep promotion and sleep homeostasis. Another possibility is that if some strong arousal promotion areas are suppressed, the tendency to sleep may increase, which is consistent with the idea that sleep represents the default state of the neural network or the entire organism (46, 64). Therefore, even if the accumulation of sleep demand occurs globally in some form (51, 52, 65⇓⇓⇓⇓⇓⇓⇓⇓–74), state switching is likely to start from a relatively limited set of brain circuits. They can combine sleep-wake history-related signals with other ecological and homeostasis needs. Although the biological basis of global sleep homeostasis is still unclear, the question of which brain regions are involved in encoding wake or sleep time appears to be solvable (41, 50, 51).

Interestingly, our data shows that although LPO and non-LPO light activation cause similar sleep loss, only LPO stimulation can produce enhanced sleep pressure, which is manifested by an increase in SWA during subsequent sleep. In addition, during arousal induction, only LPO stimulation leads to an increase in EEG θ frequency activity, which is an established marker of behavioral arousal and alertness (49, 54, 75⇓ –77). In addition, compared to NREM sleep, the observed latency of wakefulness is longer when stimulation is provided during REM sleep. These observations object to the possibility that our stimulation regimen is non-physiological or non-specific. Consistent with the observed accumulation of sleep pressure during theta-dominated waking period, we observed that the increase in EEG theta activity during LPO stimulation is closely related to the increase in SWA during subsequent sleep, and during this period the light activation of GAD2LPO neurons spontaneously Wakefulness leads to a further increase in the intensity of subsequent sleep.

A recognized limitation of the method we use is that we may target heterogeneous neuronal populations because the hypothalamus contains many different cell types (35, 78). We want to emphasize that it is not yet clear how fine the spatial resolution should be in stimulus experiments like ours in order to gain meaningful insights into the brain matrix that globally controls sleep and wakefulness. More relevantly, the method we choose may result in the activation of circuits that are not directly or incompletely related to sleep-wake control. For example, LPO is adjacent to the lateral hypothalamic circuit and contains GABAergic neurons that promote a wide range of functions, from reward-seeking behaviors to eating (79⇓ -81). In addition, the preoptic area of ​​the hypothalamus contains inhibitory neurons that participate in parental behavior and various homeostatic processes, such as thermoregulation and fluid homeostasis (82⇓ -84). Importantly, LPO receives input from several marginal regions, including the septum, hypothalamus, subliminal cortex, amygdala, and brainstem (85, 86), which can transmit salient or disgusting signals and can therefore act as an alarm system Respond to important physiological drives or threats. In addition, a large number of GAD2 neurons in LPO project to the parietal and frontal cortex (17, 87), and it has been shown that the selective light activation of preoptic GAD2 neurons of the forehead projection moderately but significantly promotes wakefulness (17) Therefore, the connectivity of LPO, especially its inhibitory neurons, supports the view that this region plays a central role in integrating the basic physiological drives related to arousal regulation.

Our dexmedetomidine experiments show that the activation of GABAergic neurons in the hypothalamus under deep sedation is sufficient to generate arousal, which is consistent with the observation that sleep-active GAD neurons express α2A adrenergic receptors, and is selective Ablation of galaninergic neurons in LPO attenuates dexmedetomidine (26). Once again, compared to non-LPO stimulation, LPO stimulation in sedated animals is more effective in generating wakefulness and increasing cortical activation. Surprisingly, animals exhibited basically normal arousal behavior when stimulated during sedation, even if their peripheral body temperature dropped to a value close to room temperature. We found that continuous strong stimulation of LPO can simultaneously produce sustained wakefulness and hypothermia, which is interesting and further increases the complexity of the link between alert state regulation and temperature control (25, 26, 84, 88).

In order to further verify the hypothesis that the LPO region of the hypothalamus contributes to changes in the overall arousal level, we conducted a loss of function experiment, by using LEDs with sufficiently high light intensity to unilaterally photoinhibit the GAD2LPO neurons of eArch3.0 to achieve light stimulation . Interestingly, we found that the photoinhibition of active and alert animals during SD does not cause significant changes in the arousal EEG compared to the photoactivation of the same neurons. Not surprisingly, this therefore did not cause any changes in subsequent sleep. However, when an undisturbed animal undergoes photoinhibition during spontaneous wakefulness and sleep, this can lead to a decrease in the level of arousal, which is manifested as a continuously enhanced sleep SWA.

Combined with light activation experiments, we speculate that these neurons have the following effects. When the GAD2LPO neuron population is spontaneously active, this may be indirectly driven by external stimuli through input from other brain regions, such as hypothalamic secretin/orexin, norepinephrine, or cortical/striatal neurons (17, 41, 89, 90), the animal engages in a state of wakefulness characterized by increased activity and arousal levels, which leads to a subsequent increase in sleep intensity (49, 61). If the activity of these neurons is artificially enhanced by light stimulation, this situation will increase further in animals that are already awake, and rapid awakening will occur when stimulation is provided during sleep. In contrast, the relatively weak selective unilateral inhibition of GAD2LPO neurons during SD does not cause any impact, because their reduced activity is compensated by strong non-specific environmental stimuli, which effectively drives other stimuli Aroused neuron group (19). However, in the absence of strong external stimuli, continuous inhibition of GAD2LPO neurons during the baseline period does not cause immediate state transitions, but ultimately promotes sleep. This is reflected in the continuous decrease in the level of arousal, which is related to deep sleep and is characterized by an increase in SWA.

In conclusion, our research shows that the LPO of the hypothalamus not only participates in the overall sleep/wake transition, but also plays a more subtle role in connecting the level of arousal, intra-state dynamics, and sleep homeostasis. Our data illustrates the important role of the hypothalamus in the control of alertness and provides an important step for a more comprehensive characterization of the neural substrate of sleep regulation.

For detailed methods, please refer to the SI appendix. According to the Animals (Scientific Procedures) Act 1986 and Oxford University guidelines, all experimental procedures are carried out under the UK Home Office Project License #P828B64BC. This study used male and female Gad2-IRES-Cre mice (Jackson Laboratory 019022; B6N.Cg-Gad2tm2(cre)Zjh/J). The virus was obtained from the UNC vector core. Acquire and analyze chronic EEG and EMG recordings as described previously (58). Virus injection and light stimulation target the lateral preoptic area of ​​the hypothalamus. The electrophysiological data was obtained using the Tucker-Davis Technologies recording system. SD is performed as previously described (75). In the sedation experiment, Dex was diluted with saline and injected subcutaneously (100 µg/kg). Thermal imaging cameras Optris Xi 70 and Xi 80 (Optris GmbH) (91) were used to measure skin temperature. Use indirect calorimetry and CaloBox (PhenoSys) to measure metabolic rate. At the end of the experiment, the mice were deeply anesthetized and perfused with 0.1 M phosphate buffered saline (PBS) through the heart, followed by histology with 4% paraformaldehyde in PBS to confirm virus expression and fiber position. Acute slice preparation and recording were performed as previously described (92). Statistical analysis was performed in MATLAB (The MathWorks), SPSS (IBM) and GraphPad Prism (GraphPad).

The data of this research report is contained in the article, SI appendix and/or data set S1. The code generated during this research period was stored by TY on GitHub (November 27, 2021) and can be accessed at https://github.com/TomokoYamagataw/Yamagata_PNAS_HypLinkArousalSleepHomeostasis. Supplementary materials and methods, figures S1 to S9, tables S1 and legends to movies S1 to S2 can be found in the SI appendix.

We thank Professor Denis Burdakov, Dr. Christin Kosse, Dr. Hannah Alfonsa, and Dr. Adam Packer for their suggestions on experiments, data analysis and interpretation, and Professor Gerhard Heldmaier for suggestions on using CaloBox. We thank the members of VVV Lab and SCNi for their help in the experiment and comments on the manuscript. This work was awarded the Wellcome Trust Senior Investigator Award (106174/Z/14/Z, to RGF), Wellcome Trust Strategic Award (098461/Z/12/Z, to RGF), John Fell Oxford University Press Research Fund Grant (131 /032, to VVV), FP7-PEOPLE-CIG (PCIG11-GA-2012-322050, to VVV), Medical Research Council (MR/L003635/1 and MR/S01134X/1, to VVV), Royal Society (RG120466, To VVV), Naito Foundation (study abroad grant, to TY) and Uehara Memorial Foundation (postdoctoral scholarship for foreign studies, to TY).

Author contributions: TY, AJ, SNP, EOM, RGF and VVV design research; TY, MCK, JP-M., EM, MCCG, VvdV, Y.-GH, LEM and EOM research; TY contributed new Reagents/analysis tools; TY, MCK, M.Š. and EOM analysis data; TY, MCK, MCCG, VvdV, LEM and VVV wrote this paper.

The author declares no competing interests.

This article is directly contributed by PNAS.

This article contains online support information https://www.pnas.org/lookup/suppl/doi:10.1073/pnas.2101580118/-/DCSupplemental.

This open access article is distributed under the Creative Commons Attribution-Non-Commercial-No Derivative License 4.0 (CC BY-NC-ND).

Thank you for your interest in advertising on PNAS.

Note: We only ask you to provide your email address so that the people you recommend the page to know that you want them to see it and that it is not spam. We do not capture any email addresses.

Feedback privacy/legal

Copyright © 2021 National Academy of Sciences. Online ISSN 1091-6490. PNAS is a partner of CHORUS, COPE, CrossRef, ORCID and Research4Life.