Recent advances in understanding the roles of hypocretin/orexin in arousal, affect, and motivation

Sleep and wakefulness

Hcrt deficiency underlies the majority of cases of narcolepsy^17–20. Narcolepsy is characterized by unexpected sleep episodes during times of wakefulness, excessive daytime sleepiness, rapid eye movement (REM)-like episodes that can co-occur with conscious wakefulness, and disrupted nocturnal sleep^21,22. Further support for aberrant state boundaries in narcolepsy was recently published showing intrusions of REM sleep during wakefulness as well as intrusions of non-REM (NREM) sleep during wakefulness²³. While it is established that Hcrt neuron degeneration contributes to the etiology of narcolepsy in many cases, recent evidence has characterized how sleep and wakefulness are impacted through the progression of Hcrt cell loss^17,18,24. Studies in mice at different stages of Hcrt neuron degeneration found that loss of these neurons reduces the likelihood of long wake bouts but increases the likelihood of short wake bouts (that is, wakefulness is fragmented) as a result of waking primarily during the first 30 seconds of NREM sleep and a reduced likelihood of returning to sleep within the first 60 seconds of wakefulness²⁴.

While early observations demonstrated that Hcrt deficiency underlies narcolepsy, a causal role for Hcrt in sleep-to-wake transitions was shown only in 2007²⁵. Optogenetic manipulations of Hcrt circuitry revealed that activation of this neuronal population induces wakefulness in mice while optogenetic inhibition promotes NREM sleep^25,26. Likewise, chemogenetic studies targeting Hcrt neural activity have shown that injections of CNO in mice expressing excitatory (Gq) DREADDs promote wakefulness but that engagement of inhibitory (Gi) DREADDs decreases wakefulness and increases time in NREM sleep²⁷. Thus, Hcrt clearly plays a critical role in the regulation of sleep-to-wake transitions, but its various effects on these processes are regulated by the many brain regions and neurotransmitter systems with which it interacts. Indeed, research has demonstrated important interactions between Hcrt and histaminergic neurons within the tuberomammillary nucleus (TMN), cholinergic and GABAergic neurons of the basal forebrain (BF), dopamine (DA) neurons within the ventral tegmental area (VTA), and norepinephrine (NE) neurons of the locus coeruleus (LC), among others²⁸ (Figure 1). Recent advances in our understanding of the roles of these regions in sleep/wake regulation and their possible interactions with the Hcrt system are outlined below.

Figure 1. Hypocretin arousal network.

Research of the past three years has found evidence of hypocretin-associated arousal in the illustrated circuits. Solid lines denote excitatory projections, and dashed lines denote inhibitory projections. 5-HT, serotonin; ACh, acetylcholine; AMY, amygdala; BF, basal forebrain; DA, dopamine; DRN, dorsal raphe nucleus; GABA, gamma aminobutyric acid; HA, histamine; Hcrt, hypocretin; LC, locus coeruleus; LH, lateral hypothalamus; NA, noradrenergic system; NAc, nucleus accumbens; NE, norepinephrine; NREM, non-rapid eye movement; PV, parvalbumin; REM, rapid eye movement; SOM, somatostatin; TMN, tuberomammillary nucleus; VTA, ventral tegmental area.

As we discuss below, histaminergic neurons of the TMN play a role in arousal, but the ways in which Hcrt influences TMN-mediated arousal are not clear. TMN histaminergic neurons become active during wake onset and are silent during sleep^29,30. Optogenetic silencing of histaminergic TMN neurons induces NREM sleep and inhibits wakefulness³¹. Hcrt activates TMN neurons and increases histamine release at their terminals, suggesting that Hcrt activation of TMN neurons supports wakefulness^32–34. However, mice and zebrafish that lack the rate-limiting enzyme in histamine synthesis (histamine decarboxylase) show normal sleep-to-wake transitions upon optogenetic stimulation of Hcrt neurons^35,36. These data suggest that histaminergic signaling in the TMN may serve a redundant function in Hcrt-mediated arousal. Recent findings also show that histaminergic regulation of wakefulness within the TMN may be via co-transmission of GABA. Small interfering RNA (siRNA)-mediated knockdown of the vesicular GABA transporter (VGAT) or genetic knockout of the VGAT gene in histaminergic neurons results in hyperactivity and sustained wakefulness³⁷. Future studies should characterize how manipulations of GABA transmission in the TMN impacts Hcrt-induced wakefulness specifically.

The BF is an attention- and arousal-sustaining structure containing cholinergic, GABAergic, and glutamatergic cells that are depolarized by Hcrt³⁸. Similarly, the region expresses both Hcrt receptors, and there is a higher density of OX₂R than OX₁R³⁹. This difference may be meaningful, as studies in organotypic slice cultures show that Hcrt depolarizes cholinergic cells of the BF via actions at OX₂R but not OX₁R³⁸. However, injections of Ox-A into the BF of rats resulted in wakefulness in regions of the BF that show stronger expression of OX₁R⁴⁰. Chemogenetic studies demonstrate that activation of cholinergic neurons of the BF decreases electroencephalogram (EEG) delta power (specifically during NREM sleep) and promotes cortical desynchronization without behavioral wakefulness⁴¹. In contrast, activation of GABAergic neurons in this region produces sustained wakefulness whereas inhibition increases NREM sleep⁴². Further genetic targeting studies show that subsets of GABAergic neurons in the region exhibit a diversity of responses across arousal states^43–45. For example, parvalbumin-positive (PV⁺) GABAergic neurons are more active during wakefulness and REM sleep than during NREM sleep whereas somatostatin-positive (SOM⁺) GABAergic neurons are reciprocally silent during wakefulness. Predictably, optogenetic activation of PV⁺ GABA neurons powerfully induces wakefulness whereas activation of SOM⁺ GABAergic neurons promotes NREM sleep^46–49. Modern genetic tools will continue to allow more detailed examinations of the impact of neuronal heterogeneity within regions in the context of Hcrt-mediated arousal.

The BF receives projections from midbrain DA neurons which may underlie the coupling of motivation to arousal states. Indeed, Hcrt axons project to midbrain DA neurons, and DA cell bodies express Hcrt receptors^13,50,51. In vitro electrophysiological recordings show that Hcrt1 and Hcrt2 treatment increases VTA DA neural firing⁵². Hcrt1 injections into the VTA increase time awake and levels of DA at axonal terminals in the prefrontal cortex^53,54. Although Hcrt neurons project to systems for all the monoamines and drugs that increase DA transmission increase wakefulness, DA was thought not to be involved in normal sleep/wake regulation until recently^55–60. Work from our laboratory has shown a role for VTA DA neurons in promoting arousal and the initiation of sleep-preparatory behaviors⁶¹. Optogenetic activation of VTA DA neurons induces emergence from anesthesia, and chemogenetic activation of the VTA induces and consolidates wakefulness^62,63. Further manipulations have demonstrated that VTA effects on wakefulness are through a D₂ receptor-mediated mechanism^62,63. Future work using projection-specific manipulations of Hcrt fibers within the VTA should better characterize their role in VTA-mediated arousal.

Noradrenergic neurons of the LC are strong promoters of arousal^64,65. Direct administration of Hcrt1 into the LC increases firing rates while optogenetic silencing of these neurons with concurrent excitation of Hcrt cells prevents Hcrt-evoked sleep-to-wake transitions^66–68. Additional studies have shown that noradrenergic activity is required to promote wakefulness and Hcrt-induced arousal in zebrafish. Using DA b-hydroxylase (dbh) (the rate-limiting enzyme in NE synthesis) mutant zebrafish, researchers found that these animals had dramatically increased sleep yet lower arousal thresholds⁶⁹. Additionally, wakefulness induced by genetic overexpression of Hcrt and optogenetic activation of Hcrt neurons is blocked by the inhibition or knocking out of NE in zebrafish larvae⁶⁹. However, further investigations have shown that overexpression of Hcrt or activation of Hcrt neurons has no significant effect in dbh mutant zebrafish³⁵. Thus, future work should continue to parse out the roles in which NE functions in sleep/wake regulation and how it may serve specifically within the Hcrt circuit to help regulate wakefulness in particular.

Motor tone

Despite evidence demonstrating innervation of motor control systems by the Hcrt neurons, the coupling of arousal states with motor control is poorly understood⁷⁰. Indeed, measures of muscle tone along with cortical activity are the most common endpoints for characterizing various arousal states. A hallmark of waking is low-amplitude, high-frequency EEG activity with high muscle activity. REM sleep, also known as paradoxical sleep, is characterized by a near complete loss of skeletal muscle activity and an EEG resembling wakefulness. Hcrt-deficient narcoleptics show cataplexy (a loss of muscle tone during wakefulness that can result in postural collapse and can be triggered by strong emotions such as happiness and fear)^22,71–74. Similarly, individuals with REM sleep behavior disorder (RBD) show muscle tone problems. Under normal conditions, REM sleep is devoid of skeletal muscle tone; however, in RBD, an individual acts out their dreams by moving their limbs or talking, which can be dangerous for the individual enacting their dreams as well as anyone in their surroundings⁷⁵. Noradrenergic activity is necessary for motor behavior⁷⁶. Indeed, NE depletion has been shown to have a stronger motor-impairing effect than dopaminergic lesions with MPTP (1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine) infusions of NE-induced hyperactivity, and loss of NE neurons is associated with motor learning deficits in aged rats^77–80. Likewise, increasing noradrenergic tone has been shown to reduce cataplectic episodes⁸¹. As discussed above, noradrenergic neurons of the LC are powerfully regulated by Hcrt; Hcrt dysfunction predictably alters both arousal and motor control. Moreover, Hcrt neurons project to dorsal raphe nucleus (DRN) serotonergic neurons where they may further influence motor behavior. Indeed, restoration of OX₂R into serotonergic DRN neurons of dual Hcrt receptor knockout mice suppresses cataplexy-like episodes yet has no effect on sleep/wake fragmentation. Likewise, optogenetic stimulation of serotonergic DRN terminals in the amygdala (AMY) suppresses cataplexy-like arrests in Hcrt-deficient mice, and optogenetic inhibition blocks the cataplexy-reducing effect of Hcrt receptor restoration in serotonergic DRN neurons⁸². Additional chemogenetic manipulations of this amygdalar circuit show that GABAergic populations of the central AMY are responsible for the production of cataplexy in mice but may not be the only circuit that can drive emotionally driven cataplexy⁸³. Together, these findings demonstrate a key role for amygdalar circuits in the production of cataplexy; however, they do not rule out other nuclei or circuits that may influence emotionally driven cataplexy. Indeed, the neural infrastructure exists for Hcrt activity to modulate AMY activity via its connections from the LC and DRN, and future studies should characterize the influence of Hcrt in emotion-driven cataplexy.

Stress and anxiety

Hcrt plays a role in the coordination of stress responses. Plasticity in the Hcrt system is thought to contribute to long-term dysregulation of arousal seen in certain psychiatric disorders^85,86. This may be an adaptive response to repeated stress, where heightened arousal and vigilance are needed under conditions of instability or high threat⁸⁷. Recent literature has supported the idea that activation of OX₁R promotes anxiety-like behavior. For example, in rodent models of panic, an extreme form of anxiety, animals with panic vulnerability treated with the OX₁R antagonist compound 56 reduced panic-like behaviors in a sodium lactate model of panic induction⁸⁸. Similarly, treatment with the OX₁R antagonist JNJ-54717793 attenuates panic-like behavior and cardiovascular responses in both the sodium lactate model of panic and a carbon dioxide (CO₂) model of panic provocation⁸⁹. Additional studies within the CO₂ model that screened selective Hcrt receptor antagonists (SORAs) and dual Hcrt receptor antagonists (DORAs) found that both a SORA1 (compound 56) and a DORA-12 attenuate anxiety-like behaviors but that a SORA2 did not⁹⁰. Importantly, these data provide a promising treatment route, as animals treated with SORA1 and DORA-12 showed no significant changes in sleep⁹⁰. Currently, the levels of benzodiazepines needed to achieve anxiolytic effects are also sedating; as discussed here, OX₁R antagonists can have anxiolytic effects without impacting sleep⁹⁰.

Although the mechanism of action of the wake-promoting drug modafinil is mainly through activation of DA circuitry, it also activates Hcrt neurons and is used for the treatment of narcolepsy. Treatment with modafinil after a traumatic experience reduces the incidence of post-traumatic stress disorder (PTSD), a disorder characterized by anxiety and hyperarousal. The anxiolytic effect of this treatment may be due to its interference with normal sleep-dependent memory processes⁹¹. However, the benefits of modafinil treatment may go beyond this, as it has been shown to stimulate adaptive stress responses in an animal model of PTSD^92,93. In a model of orofacial pain-induced anxiety, rats given injections of capsaicin into the upper lip showed increased anxiety-like responses on the elevated plus maze. Administration of Hcrt exacerbates this response while treatment with OX₁R antagonists inhibits orofacial pain-associated anxiety⁹⁴. In another study, differential effects of OX₁R antagonism were observed. The OX₁R antagonist SB-334867 influenced arousal (mobility/immobility in an open field) but not anxiety-like behavior (center exploration) in conditions of mild stress in male rats⁹⁵. Yet Hcrt knockout mice show increased anxiety in the open-field test, light-dark box test, and predator scent avoidance test despite intact fear learning⁹⁶. Likewise, OX₁R receptor knockout mice show increased anxiety and reduced social interaction, increased startle responses, and altered depressive-like behavior⁹⁷. Although genetic knockout results do not completely contradict findings from pharmacological studies, they do showcase the necessity to use the newest genetic techniques to parse out the role of Hcrt in anxiety. Two points must be made with regard to these findings: first, knockout models may result in compensatory mechanisms that may explain how Hcrt-null or OX₁R-deficient mice display lower anxiety. Second, models of stress discussed here vary greatly, and the conclusions drawn from these works may reflect the differences in the circuits underlying different types of anxiety. Thus, findings must be interpreted within the context of pharmacological, genetic, and behavioral manipulations used in these studies.

Recent work is also characterizing how individual differences in baseline Hcrt activity may pose resilience or susceptibility to stress. Rats that show low expression of preprohypocretin mRNA are resilient to social stress, and further manipulations show that chemogenetic inhibition of Hcrt reduces depressive-like behavior in otherwise stress-susceptible rats⁹⁸. Together, these data suggest that the activity of Hcrt on stress may be context or stressor specific but additionally that individual differences at baseline may influence stress resilience.

Motivation and addiction

The mesolimbic DA system, which originates in the VTA and projects to the striatum, is a key region for the processing of reward and reinforcement^99,100. These processes necessitate and evoke arousal states to monitor reinforcers and facilitate learning¹⁰¹. Reciprocally, motivational states impact arousal so as to facilitate the seeking of rewards and the avoidance of punishments^102,103. As discussed above, LH-Hcrt neurons send excitatory projections to the VTA^13,50,51. Thus, the VTA may be an optimal region by which Hcrt can influence motivated arousal states. The majority of recent advances made in this field have investigated the effects of Hcrt manipulations on motivation for cocaine and ethanol (EtOH). To date, these studies suggest that Hcrt1 plays a role in motivation for drug reward, especially when drug presentation is dependent on effortful responses on the part of the animal. Here, we discuss the role of Hcrt in addiction and motivation, focusing on cocaine, alcohol, and opioids.

Hcrt knockdown attenuates cocaine self-administration under progressive ratio schedule (that is, Hcrt knockdown lowers cocaine breakpoint) but not under a fixed ratio schedule¹⁰⁴. Similarly, Hcrt-deficient mice show reduced cue-induced cocaine-seeking behavior following a period of abstinence, suggesting a role for Hcrt in relapse behavior¹⁰⁵. Additionally, these animals show blunted cocaine intake at the highest dose and reduced behavioral responses to cocaine after abstinence¹⁰⁵. Additional work from Navarro and colleagues further supports the role of Hcrt in relapse behavior¹⁰⁶. In particular, their work shows that cocaine acts at and alters activity of corticotropin-releasing factor receptor (CRF₁R)/OX₁R heterodimers within the VTA. Action of cocaine at these sites disrupts Hcrt/CRF crosstalk even 24 hours after a single systemic injection and may be a mechanism underlying stress-induced cocaine relapse¹⁰⁶.

Indeed, Hcrt may play a unique role in cue-reward associations, as OX₁R antagonism via SB-334867 only decreases cocaine demand in the presence of cues. SB-334867 treatment also blocks cue-induced reinstatement of drug seeking—an effect most pronounced in high-demand animals (animals with the greatest cue-dependent behavior). This suggests that OX₁R increases the reinforcing efficacy of cocaine-associated cues but not of cocaine alone. This supports the notion that Hcrt plays a role in the ability of conditioned cues to elicit motivational responses¹⁰⁷. Recent in vivo measurements of DA activity are beginning to inform the mechanisms that may underlie these observed effects on cocaine reinforcement. For example, Hcrt knockdown within the VTA delays acquisition of cocaine self-administration and reduces motivation for cocaine under a progressive ratio schedule while reducing DA release in the ventral striatum, DA uptake, and cocaine-induced DA reuptake inhibition at striatal terminals¹⁰⁸. Similarly, OX₁R blockade with RTIOX-276 attenuates motivation for cocaine and reduces the number of DA transients, DA release evoked by cocaine cues, and cocaine-induced DA reuptake inhibition as measured by fast scan cyclic voltammetry (FSCV)¹⁰⁹. Suvorexant, a DORA, attenuates the motivational properties of cocaine as measured by progressive ratio and place conditioning. Additionally, treatment with Suvorexant also reduces the hedonic properties of cocaine as measured by ultrasonic vocalizations. Additionally, DORA treatment reduced cocaine-induced elevations in ventral striatal DA¹¹⁰. Work by Prince and colleagues suggests that effects of the DORA may be mediated by OX₁R, as blockade of OX₂R receptors alone has no effect on DA signaling or self-administration of cocaine¹¹¹. However, blocking of OX₁R or both OX₁R and OX₂R decreases motivation for cocaine as measured by self-administration under a progressive ratio schedule and reduces the effects of cocaine on DA signaling as measured by FSCV¹¹¹.

In the case of EtOH, Hcrt antagonism generally reduces EtOH consumption. In a voluntary EtOH intake model in zebrafish, it was seen that intake of EtOH increases Hcrt expression in the hypothalamus¹¹². OX₁R antagonism with SB-334867 reduces EtOH self-administration in alcohol-preferring rats¹¹³. Similarly, the OX₁R antagonist GSK1059865 reduces EtOH drinking in EtOH-dependent mice¹¹⁴. In a model of EtOH seeking and preference, activation of the LH is correlated with degree of seeking in context-induced reinstatement and degree of preference in home cage EtOH preference testing. Interestingly, cue-evoked reinstatement shows no correlation with Hcrt activation in any region. This suggests that there is a relationship between Hcrt activity in the LH and EtOH seeking and preference behavior but that cue-induced reinstatement for alcohol may be mediated by a different mechanism¹¹⁵. Interestingly, EtOH consumption increases OX₂R mRNA within the anterior paraventricular nucleus of the thalamus and local antagonism of OX₂R reduces total EtOH intake¹¹⁶.

The interactions of Hcrt with opioid rewards are particularly interesting, as the endogenous opioid dynorphin (Dyn) is expressed in 94% of Hcrt neurons and Hcrt and Dyn are thought to be co-released at Hcrt terminals within the VTA¹¹⁷. The interactions of these neurotransmitters are beyond the scope of this review; however, of major relevance is the point that these neurotransmitters have opposing yet complementary actions on VTA cellular excitability^117–121. OX₁R antagonism with SB-332867 modulates demand for the opioid drug remifentanil in low takers but not in high takers¹²². Additionally, intra-VTA injections of the OX₁R antagonist SB-334867 attenuate morphine conditioned place preference (CPP) acquisition and expression. Interestingly, in the case of opioid reward, OX₂R antagonism via TCS-OX2-29 also significantly attenuates morphine CPP acquisition and expression, suggesting that both receptors within the VTA are important for expression of morphine reward¹²³. Similarly, systemic treatment with the OX₂R antagonist NBI-80713 dose-dependently reduces heroin self-administration in a long-access paradigm. Long-access heroin self-administration paradigms are thought to mimic compulsive drug taking; thus, OX₂R antagonism may be particularly effective at influencing drug-associated compulsivity. Similar effects have been observed in the hippocampal dentate gyrus (DG), which receives Hcrt projections from the LH and interacts with the VTA to play an important role in the linking of drug reward with contextual cues¹²⁴. In a stress- and drug-induced model of morphine reinstatement, intra-DG administration of OX₁R and OX₂R antagonists attenuates drug priming-induced reinstatement dose-dependently with no effect on stress-induced reinstatement¹²⁵. Similarly, morphine CPP increases Hcrt1 release in the DG and OX₁R antagonism via SB-334867 ameliorates morphine CPP. These findings suggest that Hcrt actions at the DG may influence the learning of drug-context associations¹²⁶.

Finally, additional work has begun to delineate the effect of Hcrt on motivation at VTA terminal sites such as the nucleus accumbens (NAc)¹²⁷. Blomeley and colleagues used optogenetics and electrophysiology to characterize a direct Hcrt→DA D₂ excitatory circuit that is necessary for the expression of risk avoidance behavior in mice¹²⁷. Indeed, increased DA D₂ neuron activation caused animals to avoid risks such as crossing a predator-scented chamber to attain a food reward and chemogenetic silencing of accumbal DA D₂ cells inhibited Hcrt-mediated avoidance. Importantly, these data showcase how Hcrt can influence adaptive behavioral inhibition even in the presence of rewards. These data open up new opportunities of research, such as characterizing the effects of Hcrt on different subregions of the NAc, which is a heterogeneous structure with distinct electrophysiological properties^128,129. Additional lines of research should investigate how Hcrt-mediated motivation in the NAc is impacted by diurnal rhythms as well as sleep disturbance and how the Dyn system interacts in this region to modulate motivation^120,130.

Cognitive function and memory

Studies suggest that Hcrt deficiency is associated with memory deficits. Hcrt deficiencies negatively impact working memory as tested in a non-matching-to-place T-maze task¹³¹. Hcrt/ataxin-3 transgenic mice (a progressive model of narcolepsy), which become Hcrt deficient at 12 weeks old, show impaired avoidance memory in a two-way active avoidance paradigm in which an animal has to perform a specific motor response to avoid an aversive stimulus. Hcrt1 administration reverses memory deficits, suggesting that Hcrt plays a role in hippocampal-dependent consolidation of two-way active avoidance memory¹³². Chemogenetic activation of Hcrt neurons improves short-term memory for novel locations, a function that putatively supports foraging and exploration¹³³.

Pain negatively influences memory processing in ways that may be influenced by Hcrt. In the Morris water maze (MWM) (a test of spatial learning and memory), orofacial pain-induced memory impairments are exacerbated by the OX₁R antagonist SB-334867 whereas administration of Hcrt1 prevented these spatial memory deficits¹³⁴. Importantly, injections were directed at the trigeminal nucleus caudalis, which is a central relay for orofacial pain. Thus, the observed effect on memory may be via alterations in the experience of pain itself rather than the formation of a pain-associated memory¹³⁴. In a similar study by Raoof and colleagues, orofacial pain memory was mediated by Hcrt at the level of the hippocampus (HPC). Intra-hippocampal injections of Hcrt1 inhibit pain-induced memory impairments as measured by the MWM. However, treatment with the OX₁R antagonist SB-334867 had no effect on learning and memory¹³⁵. Indeed, the HPC is a critical region for memory function and Hcrt action at this site may influence memory processes via its influence on the induction of long-term potentiation (LTP). In vitro studies show that OX₁R antagonists significantly decrease the firing rates of hippocampal CA1 neurons, showing that the effect of Hcrt on these neurons is excitatory¹³⁶. Additional in vitro electrophysiology studies demonstrate that Hcrt1 may bidirectionally modulate HPC CA1 function. Specifically, moderate doses of Hcrt1 inhibit LTP while subnanomolar concentrations result in re-potentiation via OX₁R and OX₂R¹³⁷. It is important to note that the Hcrt manipulations discussed here may have influenced sleep and therefore resulting memory problems may be sleep dependent and thus only indirectly dependent on Hcrt.

Models of hypocretin network in arousal

Current models have described Hcrt as functioning within a “flip/flop” model where it stabilizes wakefulness, preventing aberrant switches between mutually exclusive states¹³⁸. This model, however, cannot account for overlapping states of arousal such as those observed in narcolepsy or RBD in which REM sleep can co-occur with conscious awareness^139,140. Additionally, this model does not factor in the many systems that interact to influence arousal. These observations make it necessary to revise the binary nature of the flip/flop model. Studies have expanded the model by characterizing a circuit with hierarchical gating of additional neural circuits, feedback, and redundancy¹⁴¹. This hierarchical model provides a framework on which to add motivational influences on arousal states. Indeed, animals can adapt their sleep on the basis of internal and external variables such as migration or predator avoidance or to increase the likelihood of mating^142–144. Recently, an alternative has been proposed in which sleep-to-wake transitions are predicted on the basis of inputs with different “weights” onto an integrator neuron¹⁴⁵. An integrator neuron would continuously compute probabilities of wakefulness on the basis of functional connectivity of the system as well as physiological factors such as stress or circadian phase. Diversity of neuronal responses to stimuli can be integrated within this model to account for the heterogeneity of the system. In this vein, Schöne and Burdakov acknowledge the necessity of an adaptive behavioral control system that can respond to unpredictable changes in the environment¹⁴⁶. Thus, they propose a model of brain arousal control modules organized in a feedback loop by which Hcrt can gate relevant information on the basis of environmental and homeostatic needs¹⁴⁶. We look forward to the future advancement of this area of Hcrt research that will undoubtedly expand our understanding as an adaptable regulator of arousal.

Volume transmission

Volume transmission (VT) is a mechanism of neural signaling by which neurotransmitters can exert actions on cells in close proximity as well as distant targets. In VT, neurotransmitters signal via diffusion within extracellular fluid^147,148. This type of release is thought to allow for modulation of neural activity via long time courses and greater distances^147–149. VT may happen via cellular pores, diffusion through the plasma membrane, exocytosis, or reversal of transporter proteins¹⁴⁹. To date, actions of Hcrt at the dorsal lateral geniculate nucleus (DLG) and the DRN (aside from already-known synaptic actions) have been theorized to be exerted via VT^150,151. Observations of Hcrt1 immunoreactivity in many non-synaptic varicosities located far from synapses with axons forming asymmetric synapses suggest that DRN excitation via Hcrt1 may be via this mechanism¹⁵⁰. Indeed, the DRN plays an important role in the regulation of arousal and both synaptic and VT mechanisms may support long-term cortical arousal^25,36. In a separate set of findings, Hcrt was found to powerfully modulate neurons of the DLG despite only sparse expression of Hcrt nerve terminals in the region, suggesting that these actions are via VT¹⁵¹. Additionally, a recent study of melanin-concentrating hormone (MCH), a hypothalamic peptide important for the regulation of feeding, shows that MCH neurons project to ventricular regions where they increase MCH levels in the cerebrospinal fluid (CSF) and stimulate feeding¹⁵². MCH neurons are intermingled with Hcrt neurons in the LH, and the authors measure that 40% of Hcrt neurons also project to the CSF where they are poised to signal via VT to influence distal targets¹⁵². Further investigations should determine whether Hcrt acts via VT and, if so, how its activity is influenced by (1) temporal and spatial release dynamics, (2) diffusion and dilution parameters, and (3) transporter kinetics in order to characterize its effective radius.