From Headache Group, Institute of Neurology, The National Hospital for Neurology and Neurosurgery, London, UK (Drs. Holland and Goadsby); Department of Neurology, University of California, San Francisco, San Francisco CA, USA (Dr. Goadsby).
Address all correspondence to Dr. Peter J. Goadsby, Institute of Neurology, Queen Square, London, WC1N 3BG, UK.
The primary headaches are a group of distinct individually characterized attack forms, which although varying in presentation, share some common anatomical basis responsible for the pain component of the attack. The hypothalamus is known to modulate a multitude of functions and has been shown to be involved in the pathophysiology of a variety of primary headaches including cluster headache and chronic migraine. It seems likely that it may be involved in other primary headache disorders due to their episodic nature and may underlie many of their diverse symptoms. We discuss the hypothalamic involvement in the modulation of trigeminovascular processing and examine the involvement of the hypothalamic orexinergic system as a key regulator of this function.
The primary headaches are a group of distinct individually characterized attack forms including migraine, tension type headache (TTH), cluster headache (CH), and other trigeminal autonomic cephalalgias (TAC's).1 It is now widely believed that the primary headaches, although varying in presentation, share some common anatomical basis responsible for the pain component of the attack. Ray and Wolff initially identified that a variety of stimuli could illicit a nociceptive response from intracranial structures including the dura mater and dural blood vessels.2 The trigeminal nerve gives rise to the majority of afferent fibers innervating the head, face, and dural vasculature and for this reason it is of great importance to primary headaches.3,4
The rich innervation of the vasculature and meninges of the brain provides a dense plexus of mainly unmyelinated fibers that arise from the trigeminal ganglion and to a lesser extent the upper cervical dorsal roots. The pharmacology of the trigeminovascular system is complex in an overall sense, but can be understood in terms of the anatomy and physiology of the pain-producing structures. The peripheral branch consisting of the cranial circulation and dura mater receives sympathetic, parasympathetic, and sensory nerve fibers, all containing their own characteristic neurotransmitters (Fig. 1). Sympathetic nerve fibers arising from the superior cervical ganglion supply the cranial vasculature with neuropeptide Y (NPY), noradrenaline (NA), and adenosine triphosphate (ATP). Parasympathetic nerve fibers arising from the sphenopalatine and otic ganglia as well as the carotid miniganglia, supply the cranial vasculature with vasoactive intestinal peptide (VIP), peptide histidine isoleucine (PHI), acetylcholinesterase (AChE), peptide histidine methionine 27 (PHM, human version), pituitary cyclase-activating peptide (PACAP), and other VIP-related peptides.5 Sensory nerve fibers arising from the trigeminal ganglion supply the cranial vasculature with substance P (SP), calcitonin gene-related peptide (CGRP), neurokinin A (NKA), and PACAP. Bipolar trigeminovascular afferents innervating the cranial structures project centrally and synapse on second order neurons in the trigeminal nucleus caudalis (TNC), which is the key relay center for transmission of nociceptive information to higher brain structures.6
ASCENDING NOCICEPTIVE PROJECTIONS FROM THE TNC
Nociceptive stimulation activates neurons in the TNC, which project to multiple brainstem, thalamic, hypothalamic, and telencephalic sites that in sequence distribute sensory information to multiple cortical regions.7 These pathways can be either monosynaptic, as with the trigeminohypothalamic and trigeminothalamic tracts, or polysnaptic, like the spinomesencephalic tract, projecting to other brain regions en route to higher centers. Together they play a major role in the modulation and experience of pain, integrating both sensory discriminative and affective-cognitive aspects of nociceptive processing.8
The hypothalamus contributes only 0.5% of the brain,9 but is involved in a multitude of brain functions. It is known to be involved in hormone synthesis, regulation of the autonomic nervous system, thermoregulation, determining biological rhythms, emotional behavior, arousal, and the cardiovascular system.10 Afferent and efferent nerve fibers connect the hypothalamus to a variety of structures including the cerebral cortex, thalamus, hippocampus, amygdala, septum, periaqueductal gray (PAG), and the spinal cord. The hypothalamus also receives a dense blood supply and individual hypothalamic nuclei can be influenced by a wide variety of chemical messengers from the blood and cerebrospinal fluid, and neurotransmitters from other neurons.
THE HYPOTHALAMUS AND NOCICEPTIVE PROCESSING
The hypothalamus is not traditionally associated with nociceptive processing; however, it does have direct ascending and descending connections with the dorsal horn.11,12 It has afferent and efferent connections with many structures, including the nucleus tractus solitarius (NST), rostroventromedial medulla (RVM), PAG, and nucleus raphe magnus (NRM) as well as some corticolimbic structures implicated in the affective and cognitive aspects of pain.12 To date numerous hypothalamic nuclei have been implicated in the descending modulation of pain and nociceptive processing. The medial preoptic nucleus (MPO) has clear projections to the PAG, NRM, and the RVM,13,14 all areas involved in nociceptive processing. This nucleus is known to play a key role in the autonomic response to pain, and stimulation of the MPO inhibits the responses of spinal neurons to noxious stimuli.15 Stimulation of the medial hypothalamus inhibits the responses of spinal cord neurons to noxious peripheral stimuli,16–18 as does lateral hypothalamic stimulation.16,19,20 Electrical stimulation of the paraventricular nucleus (PVN) has also demonstrated antinociceptive effects.21 Stimulation of the anterior hypothalamus is also known to suppress the response of wide dynamic range neurons in the dorsal horn to noxious stimulation16,22 and opioid injections into the posterior, preoptic, and arcuate nuclei of the hypothalamus elicits behavioral antinociception.23,24
THE HYPOTHALAMUS AND THE AUTONOMIC NERVOUS SYSTEM
The autonomic nervous system is regulated by nuclei of the hypothalamus and limbic system via the NTS onto the brain stem nuclei and spinal autonomic ganglia. The brain stem nuclei and spinal autonomic ganglia receive afferent input from the periphery, which then project to the spinal cord and brainstem nuclei, en route to the hypothalamus and limbic system. The integration of this autonomic system (facial nerve, parasympathetic outflow) and the trigeminal nerve gives rise to the trigeminal-autonomic reflex, which is considered to underlie the cranial autonomic features displayed in certain TAC's including rhinorrhoea, conjunctival infection, lacrimation, and nasal congestion.25–27
THE HYPOTHALAMUS AND CIRCADIAN RHYTHMS
The hypothalamus is intrinsically linked with the control of body rhythms and metabolic function. The most obvious circadian rhythm is the sleep wake cycle, however, many other body systems demonstrate a circadian pattern, including feeding behavior, temperature, and hormonal regulation.28 The central control center of biological rhythms is the suprachiasmatic nucleus (SCN), which orchestrates the many physiological systems influenced by the biological clock.29 The SCN uses both direct neuronal connections and diffusible signals to influence other regions of the brain and direct projections have been identified to the lateral hypothalamic, PVN, preoptic, and supraoptic nuclei.30 The hypothalamic projections of the SCN can in turn further influence endocrine and autonomic function via monosynaptic or polysynaptic connections as demonstrated by the rapid decrease in plasma corticosterone initiated by a light pulse in rats.31
THE HYPOTHALAMUS AND PRIMARY HEADACHES
Hypothalamus and Cluster Headache.— The fact that the hypothalamus plays a major role in cluster headache is now widely accepted. Initial observations indicating a seasonal and circadian rythmicity of attacks point firmly to the involvement of the suprachiasmatic nucleus.32 Further evidence is gained from the close correlation with sleep33 and the presence of cranial autonomic symptoms. PET and fMRI studies have identified hypothalamic activation during spontaneous, as well as triggered cluster attacks and short-lasting unilateral neuralgiform headache with conjunctival injection and tearing (SUNCT),34–37 accompanied by the presence of permanent subtle structural abnormalities.38 In response to the imaging evidence the use of deep brain stimulation in the posterior hypothalamus for the treatment of chronic cluster headache has proved a successful intervention strategy providing strong evidence for the involvement of the hypothalamus in cluster headache.39–41
Hypothalamus and Migraine.— As with cluster headache, clinical evidence indicating endocrine abnormalities in chronic migraine implicate the hypothalamus in its pathophysiology.42 The presence of premonitory symptoms up to 48 hours preceding the onset of an attack indicate an underlying hypothalamic dysfunction43,44 as does recent research indicating the presence of similar symptoms following glyceryl trinitrate triggered attacks.45 Akin to cluster headache, migraine attacks demonstrate a striking circadian rythmicity46–48 and are linked to hormonal fluctuation,49 further implicating the hypothalamus.
HYPOTHALAMUS AND SHORT-LASTING UNILATERAL NEURALGIFORM HEADACHE WITH CONJUNCTIVAL INJECTION AND TEARING
SUNCT is a member of the larger group of TAC's with cluster headache and paroxysmal hemicrania; like the other members SUNCT demonstrates clear autonomic symptoms pointing to a disinhibition of the trigeminal-autonomic reflex. Ipsilateral hypothalamic activation has also been reported during attacks50 in the same area that was activated in cluster headache attacks, pointing to a similar mechanism.
EXPERIMENTAL EVIDENCE FOR HYPOTHALAMIC INVOLVEMENT IN PRIMARY HEADACHES
Although the wide array of accompanying symptoms seen in primary headache disorders underline a differential pathophysiology, it is clear that the trigeminal nerve and its peripheral and central connections are of great importance to the pain component of attacks. To date no acceptable animal model has been developed to investigate the many sensory and autonomic symptoms, which are so characteristic of primary headaches, such as migraine and cluster headache. However, modeling the nociceptive component remains an achievable goal with current experimental models.
Experimental evidence for the involvement of the hypothalamus in trigeminovascular nociception has been provided from a variety of studies. Malick et al. demonstrated that stimulation of the dura mater in the rat produced Fos expression in the ventromedial, PVN, and dorsomedial hypothalamic nuclei, resulting in suppression of appetite and increases in arterial blood pressure.51 The findings indicate a complex relationship between nociceptive trigeminal neuronal activation and suppression of food intake at the level of the hypothalamus, consistent with clinical observations of loss of appetite in migraine.51 Stimulation of the superior sagittal sinus in the cat has also demonstrated hypothalamic activation with upregulation of Fos protein-like immunoreactivity in the supraoptic and posterior hypothalamic nucleus, consistent with a role for hypothalamic structures in modulation of nociceptive processing.52
The orexins are 2 neuropeptides derived from the same 130 amino acid precursor peptide, prepro-orexin, by proteolytic cleavage. A single gene located on chromosome 17q21 in humans is responsible for encoding prepro-orexin.53,54 Following proteolytic cleavage both orexin peptides are posttransationally modified, with the addition of N-terminal pyroglutamyl cyclisation and C-terminal amidation for orexin A and only C-terminal amidation for orexin B (Fig. 2).55 Mature orexin A is a 33-residue neuropeptide containing 2 disulfide bridges, Cys6-Cys12 and Cys7-Cys14,56 which are essential for functional potency57,58 as are the 19 C-terminal residues.58 Mature orexin B is a 28-residue neuropeptide containing 2 α-helices linked by a flexible loop.55 Orexin A and B are 46% homologous, and the sequence of orexin A is fully conserved across rat, mouse, pig, cow, and human.54,59,60 Orexin B shows a slight variation across species, with the rat/mouse isoforms differing by 2 amino acids compared with the human sequence60 and porcine orexin B has a single serine-to-proline substitution.59
The orexins bind to 2 G-protein coupled receptors, termed OX1R and OX2R.77 The 2 receptors are 64% homologous and are most closely related (26%) to the NPY2 receptor.54 The rat and human versions of the OX1R and OX2R demonstrate a 94% and 95% homology, respectively, suggesting a high level of conservation across mammalian species.54 Orexin A has equal affinity for both the OX1R and OX2R, with orexin B demonstrating a 10-fold higher affinity for the OX2R than the OX1R. Activation of either OX1R or OX2R results in elevated levels of intracellular Ca2+ concentrations,78,79 which results in the enhancement of the Gq-mediated stimulation of phospholipase C.
The broad projections of the orexinergic system (Fig. 3) have led to its implication in a variety of functions including feeding, sleep wake cycle, cardiovascular function, neuroendocrine, and autonomic function80–83 as well as a more recent implication in the modulation of nociceptive processing.84–87 The orexinergic system projects to many areas involved in nociceptive processing and autonomic regulation, including the hypothalamus, PAG, ventrolateral reticular formation (VLM), PVN, parabrachial nucleus (PBN), NTS, and lamina I, II, and X of the spinal and trigeminal dorsal horns.76,88,89 In general the orexin receptor expression is in agreement with the orexinergic fiber projections.63,90 Destruction of the posterior hypothalamus in rats results in transient hyperalgesia indicating a possible hypothalamic role in the maintenance of a basal nociceptive threshold.91 Bingham et al. postulated a novel descending orexinergic inhibitory system in the rat raising the possibility that this maintenance of basal nociceptive threshold may be orexin driven.84
Experimental evidence on a possible role of the orexins in modulating nociceptive processing is both complimentary and contradictory. Bingham et al. reported that orexin A is antinociceptive in the thermal and visceral nociceptive tests in the rat and mouse and antihyperalgesic in the mouse carrageenan-induced thermal hyperalgesia test when given intravenously or intracerebroventricularly.84 N-(2-Methyl-6-benzoxazolyl)-N-1,5-naphthyridin-4-yl urea (SB-334867) a selective nonpeptide OX1R antagonist reversed the orexin A effects as well as demonstrating hyperalgesic activity when given alone, suggesting the presence of a descending orexinergic inhibitory system that is activated under inflammatory conditions. The responses observed were independent of the endogenous opiate system, despite having similar effects as morphine.84
Yamamoto et al. also demonstrated the analgesic effects of orexin A and proposed a spinal cord mechanism.92 Orexin B had no effect on any of the aspects studied and SB-334867 was again able to antagonize the effects of orexin A, however, no change was seen when given alone. Thus it is possible that the descending inhibitory system proposed by Bingham et al. is only activated during inflammatory conditions, such as the mouse carrageenan-induced thermal hyperalgesia test and not under resting conditions or acute nociceptive stimuli such as the rat formalin or hot plate test.92
Further discrepancies in the roles of orexin A and B in different models of nociception exist. Suyama et al. demonstrated that orexin A but not orexin B inhibits heat-evoked hyperalgesia in the rat when given intrathecally.93 Orexin A also reduces mechanical allodynia when given intrathecally or intracerebroventricularly via a spinal or supraspinal mechanism that did not involve any peripheral mechanisms.93 Clear evidence for a state dependant role of the orexins has been demonstrated in prepro-orexin knockout mice.94 Under normal baseline conditions there was no difference in pain thresholds between knockout and wild type mice. However, following the induction of peripheral inflammation prepro-orexin knock out mice demonstrated a greater degree of hyperalgesia and lowered stress-induced analgesia than their wild type counterparts. It would appear that the orexinergic system facilitates an increase in pain thresholds under inflammatory conditions and this is likely to be mediated within the CNS.94 Peripheral inflammation is known to activate ascending pain pathways, which project to higher structures including the hypothalamus. Watanabe et al.94 demonstrated that under inflammatory conditions orexinergic hypothalamic neurons are activated; however, this may be a result of the ascending nociceptive input or as a result of the activation of endogenous descending inhibitory pathways.95 The distribution of the orexinergic system from the hypothalamus to the PAG and spinal cord present the possibility for either the modulation of descending inhibitory pathways or the direct release of orexin at the spinal cord. Thus suppressing nociceptive inputs at the second order relay neurons as demonstrated by the antinociceptive effects of spinal as well as intracerebroventricular administration.84,85,87,92,96,97
The orexins have only very recently been linked with a possible role in primary headache disorders and the evidence as such is circumstantial. An obvious generic link is the anatomical distribution of the orexins and the discovery that they could influence nociceptive processing. A link between both migraine and cluster headache with sleep has long been established. Attacks are common during rapid eye movement sleep periods and many patients find sleeping to be the preferred method to abort an attack. Narcolepsy is a disorder of the sleep–wake cycle with a prevalence of approximately 0.05% in the general population.98 It is characterized by excessive daytime somnolence, overwhelming episodes of sleep, disturbed nocturnal sleep, hypnagogic hallucinations, sleep paralysis, and cataplexy.98 The orexinergic system is known to play a key role in the pathophysiology of narcolepsy, via a loss of orexinergic cells, possibly as a result of an autoimmune disorder.99
It has been shown that narcoleptic patients have a greatly increased prevalence of migraine. In a study of 68 narcoleptic patients 64% of women and 35% of men were found to suffer from migraine attacks100 fulfilling all the International Headache Society (IHS) criteria.1 In contrast, the life time prevalence of migraine in the general population is in the range of 16% to 25% for woman and 7% to 8% in men.101–106 The increased prevalence of migraine in narcoleptic patients was later confirmed in a larger study, with 44% of women and 23% of men suffering from migraine.107
One interesting finding is the late onset of migraine in narcoleptic patients, demonstrating that narcolepsy appears to put people at an increased risk for migraine compared to the general population.108 The neuroanatomical base of both disorders may involve the brainstem, lesions in the region of the PAG, and dorsal raphe nucleus (DRN) combined with activation of the serotonergic DRN and noradrenergic locus coeruleus, all point to a role of the brainstem in migraine.109–112 Cells in these regions play a critical role in rapid eye movement sleep, a regulatory disturbance of which is thought to underline narcolepsy.98,113 Further evidence for a possible role of the orexinergic system in the pathophysiology of episodic brain disorders is provided by genetic association studies in cluster headache. The 1246 G > A polymorphism of the OX2R gene (HCRT2) is significantly associated with an increased risk of cluster headache. Patients homozygous for the G allele, in comparison to the remaining genotypes, are 5-fold more likely to develop the disorder.114 The hypothalamus is clearly implicated in the pathophysiology of cluster headache, paroxysmal hemicrania, and SUNCT35,36,50,115 and the same may be true of migraine116 and other primary headaches. The orexinergic system provides a novel sexually dimorphic117,118 descending pathway through which the hypothalamus could modulate numerous processes including sensory processing of pain at various levels exerting both excitatory and inhibitory effects.
Clinical evidence for a role of the orexinergic system in the modulation of primary headaches has not yet been investigated. However, in an animal model of trigeminovascular nociception, systemically administered orexin A significantly inhibits nociceptive responses of TNC neurons to electrical stimulation of the dura mater surrounding the middle meningeal artery.119 The orexinergic trigeminal modulatory effect is further evidenced, via the orexin A inhibition of neurogenic dural vasodilation.120 Activation of the OX1R and OX2R in the posterior hypothalamus has been shown to differentially modulate nociceptive dural inputs to the TNC.121 Activation of the OX1R elicits an antinociceptive effect whereas OX2R activation elicits a pronociceptive effect. Experimental evidence indicates that regulation of autonomic and neuroendocrine functions as well as nociceptive processing is closely coupled in the hypothalamus122 and recent data point to the involvement of orexinergic mechanisms.123 Thus, the orexinergic system is a posterior hypothalamic mechanism that is involved in central pain modulating of dural input and could be a possible link between the pain of primary neurovascular headaches and the symptomatology that is suggestive of a hypothalamic dysfunction.124 The first orally available orexinergic antagonist ACT-078537 is currently under trial for sleep disorders,125 the development of a specific OX1R agonist may prove beneficial in the treatment of a variety of conditions including the primary headaches and narcolepsy.
DIRECT ACTIONS OF A HYPOTHALAMIC OREXINERGIC DYSFUNCTION ON NOCICEPTIVE TRANSMISSION
Direct descending orexinergic projections have been demonstrated to terminate in the spinal and trigeminal dorsal horns.62,64,76,126 Although the orexins are on the whole excitatory, they have been shown to produce antinociceptive effects at the level of the spinal cord.97 The densest orexinergic projections to the dorsal horn terminate in lamina IIo and in the border region of laminae IIo/IIi, an area classically associated with inhibitory interneurons. It has been shown that as well as a direct excitatory projection on to second order relay neurons, the orexinergic fibers also form excitatory synapses with inhibitory interneurons.81 Thus a gated mechanism exists via which the hypothalamic orexinergic projections could facilitate or inhibit the activation of these ascending sensory neurons (Fig. 4).
Evidence from other structures has demonstrated a similar novel modulatory mechanism for the orexins, which allows a differential level of inhibition or excitation depending on the presence of other inputs on to the target cell. For example at low doses orexins excite serotonergic DRN neurons, however, at higher concentrations they begin to excite presynaptic GABAergic neurons, also resulting in inhibition.127 It is therefore possible, due to the nature of the orexinergic terminations, that their effect is dependant on the inputs into the inhibitory cell. For example, it is possible that at the level of the spinal cord the orexins are directly exciting the postsynaptic cell and simultaneously exciting a GABAergic/glycinergic cell that inhibits that postsynaptic cell. The inhibitory effects would be greatly enhanced if other descending pathways had activated excitatory inputs into the inhibitory cell. This form of direct and indirect opposing modulation allows the orexinergic system great flexibility over the responses it elicits and provides a mechanism by which relatively weak inputs can be amplified to produce a much larger effect in postsynaptic cells.
INDIRECT ACTION OF OREXINS VIA MODULATION OF OTHER NEUROTRANSMITTER FUNCTIONS
The majority of the orexinergic system's functions act via the modulation of other neurotransmitter systems throughout the CNS. Orexin A and B have been shown to act on a variety of neurotransmitter systems including the dopaminergic, histaminergic (Fig. 5), sympathetic, autonomic, noradrenergic, and serotonergic systems.68,80,128–136
It is important to realize that there are also other mechanisms via which the orexins could be exerting their effects. Bingham et al.84 postulated a possible peripheral mechanism via activation of peripheral OX1R, and based on the recent findings of the orexins ability to excite GABAergic neurons it is possible that their antinociceptive activities may be via activation of functional GABAergic inhibitory projections from the hypothalamus and PAG.134,136,137
HOW DOES THE OREXINERGIC SYSTEM FIT A CONSTRUCT OF THE DIVERSE SYMPTOMS OF THE PRIMARY HEADACHES?
The orexins are neuropeptides synthesized in the lateral, posterior, and periventricular hypothalamic nuclei, from here orexinergic projections extend throughout the entire length of the neuroaxis to varying degrees. It is crucial to remember that the primary headache disorders are multifactorial syndromes that are divided into a number of symptoms, including sensory disturbances, autonomic features, and pain. The orexins, as well as demonstrating the ability to modulate trigeminal nociceptive processing, are linked with the modulation of a variety of functions. They activate the hypothalamic-pituitary-adrenal axis and modulate the secretion of a variety of hormones.80 It has also been demonstrated via anatomical and experimental studies that they can modulate the autonomic nervous system.76,80
OTHER HYPOTHALAMIC PEPTIDE HORMONES
Somatostatin is another hypothalamic peptide hormone that has been shown to modulate neuroendocrine and metabolic function, as well as affecting neurotransmission via activation of G-protein coupled somatostatin receptors and inhibition of a variety of hormones. It is secreted by cells in the PVN and delta cells in the stomach, intestine, and pancreas. Upon release it can bind to 1 of 5 somatostatin receptors (sst1-5), which are distributed throughout the body138,139 including the hypothalamus, PAG, NRM, anterior cingulate cortex, and the spinal and trigeminal dorsal horns140–144 as well as nonneuronal tissue.
Clinical data for a variety of pain syndromes has provided evidence for somatostatins involvement in the generation of analgesia. Somatostatin or one of its analogs have demonstrated analgesic properties in postoperative pain,145 cancer pain,146 arthritis and the primary headaches migraine, and cluster headache.147,148 Experimental in-vivo and in-vitro evidence points to an inhibitory effect at the level of the spinal and trigeminal dorsal horns.149–151 Further antinociceptive responses are elicited from the dorsal PAG, NRM, and caudate putamen152,153 while intracerebroventricular administration of octreotide, an analog of somatostatin reduces Fos expression in the TNC following noxious corneal stimulation in the rat.154
Manipulation of the posterior hypothalamic somatostatin receptors modulates dural and facial trigeminal nociceptive transmission in the TNC. Microinjection of cyclo-somatostatin into the posterior hypothalamus inhibited both A- and C-fiber responses to dural electrical stimulation and decreased their response to noxious thermal facial stimulation.155
Somatostatin has proven trigeminovascular modulatory functions and is known to play a role in the hypothalamic regulation of metabolic, neuroendocrine, and autonomic functions. It is therefore a possible key component in the link between dural nociception and the trigeminal-autonomic reflex, which could explain the diverse symptomatology which accompanies the pain component of a variety of primary headache disorders.
The hypothalamus is critically important in the brain, only the cerebral cortex is involved in as many diverse functions. As such it is likely to play an important role in the pathophysiology of a variety functional disorders, including primary headache syndromes. The evidence for the involvement of the hypothalamus in certain subtypes including cluster headache, migraine, chronic migraine, and SUNCT is compelling. The exact role it plays has yet to be elucidated, that is does the hypothalamus actively contribute to the generation of attacks or is it activated in response to a remote “generator?” Evidence from the sleep disorder narcolepsy, which is clearly linked to hypothalamic orexinergic dysfunction, points to a gating or switching mechanism. It is possible that the hypothalamus has a similar role in primary headaches, acting as a balance between a variety of systems that could underlie the symptoms of the individual attack. Thus a dysfunction in a hypothalamic circuit, for example the orexinergic system, would allow the destabilization of the pro- and anti-nociceptive inputs on the trigeminovascular system accompanied by autonomic and sensory disturbances. This could be as a result of direct actions on the trigeminal-autonomic reflex or a combined action on the reflex and higher autonomic centers including the PVN. The episodic nature of many attack syndromes points to a crucial role of the SCN, which influences many other hypothalamic, thalamic, and cortical structures. It is possible that the hypothalamic dysfunction is orchestrated by the SCN, which would help to explain both the autonomic and endocrine abnormalities, which are characteristics of many primary headaches.