Table 1 Verisimilitude calculations for proposed migraine and cluster headache mechanisms (for details, see text)

Vasodilation of large cranial arteries is involved in migraine pain

 Vasodilation of large arteries during migraine is undecided (0)

Increased temporal pulsations during migraine and the effect of ergotamine [15]. Decreased blood velocity in MCA during migraine measured with TCD [17–19] (1.0)

No change in MCA velocity measured with TCD [20, 21] No vasodilatation measured directly with MCA in MMA and MCA during NTG-induced migraine [22] (1.0)

Calcitonin gene-related peptide (CGRP) is increased in the external jugular vein (EJV) during migraine

 CGRP increase in EJV during migraine remains undecided (0)

CGRP was increased in EJV in two studies [23, 24]. In one study, sumatriptan treatment normalized CGRP levels [24] (1.0)

CGRP was unchanged in two studies on spontaneous migraine [26, 27] and one study in nitroglycerin-induced migraine [27] (1.0)

Migraine is a dysfunction of the sensory modulatory network with the dominant disturbance affecting abnormal processing of essentially normal neuronal traffic [7]

 Whether a migraine attack is a pure neuronal process without vascular components being involved is unresolved (0). See text

A migraine attack must start in the brain to cause the prodromes and aura. Persistent activation in the brain stem is observed by PET during migraine attack [4, 33]. Few cases of symptomatic migraine are caused by brain stem lesions [35, 36]. Other migraine symptoms, photo- and phonophobia, have no peripheral cause [7]. The β-blockers used in migraine prophylaxis probably exert their effect in the CNS [37]. Valproate and topiramate also most likely work in the CNS [29] (1.0)

C-fos expression and “evoked potential” are observed in TNC after superior sagittal sinus stimulation in the cat model of migraine. [42]. There may be a peripheral source for activity. No other part of the body experiences pain without nociceptive input, except thalamic pain and other neuronal lesions with sensory sign [129]. A pure neuronal disorder does not explain the co-morbidities of migraine with aura and stroke and ischemic heart disease [48, 49, 129]; a vascular or systemic factor must be involved. A central theory would not explain possible CGRP increases in EJV [23, 24]. Systemic endothelial dysfunction present in migraine [127] (1.0)

Does aura trigger headache in migraine attacks?

 Aura is likely to trigger a migraine attack (+0.25)

Clinically, the headache in migraine is contralateral to aura in 92% [58] Experimentally, CSD activates trigeminal afferents and evokes a series of cortical meningeal and brain stem events consistent with headache development in rats [44]. CSD activates matrix metalloproteinase, which opens the blood–brain barrier [62] (0.75)

Clinically, there are well-documented cases of headache ipsilateral to aura [57]. Patients with aura but no headache challenge the notion that aura causes headache. [57]. Aura does not necessarily precede headache [57]. Experimentally, no correlation between CSD and neurogenic inflammation and nociception in rats. [66] (0.5)

Brain stem activation occurs during spontaneous and provoked migraine attacks

 Brain stem most likely activated during migraine, but lateralization doubtful; pathophysiological implications somewhat unclear. (+0.75)

Two PET studies in spontaneous [33, 34] and one in NTG-induced migraine [4], showed brain stem activation which persisted after sumatriptan treatment [4, 33, 34] (1.0)

Lateralization of activation and pain is inconsistent. In one study, PET activation was ipsilateral [4], in two others contralateral [33, 34] or bilateral [34] to pain (0.25)

Regional cerebral blood flow (rCBF) is normal in migraine without aura

 No firm conclusions (0)

rCBF measurements were normal in one SPECT study [68]. Brain stem activated in migraine but no occipital hypoperfusion observed by PET [4, 34]. Normal rCBF measured with PWI [69] (1.0)

Occipital hypoperfusion was observed with PET (n = 6) [73]. Spreading oligemia observed with PET in one case [74]. A SPECT study showed focal hypoperfusion in 74% of patients [70]. Patchy hypoperfusion was observed [71]. Small general reduction of CBF [72] (1.0)

NO is involved in migraine pathophysiology. iNOS inhibitors will be effective migraine prophylactics

 NO is likely involved in migraine (+0.5)

Nitroglycerin induces genuine migraine attacks [22, 75–83]. L-NMMA is effective in migraine [84] (1.0)

INOS inhibitors (GW273629, GW274150) were ineffective in treating migraine attacks [85, 86]. GW274150 was ineffective as a prophylactic agent [86, 87] (0.25)

Dural neurogenic inflammation (NI) is involved in migraine, predicting effectiveness of NI inhibitors in migraine

 NI unlikely to have a pivotal role in migraine pain (–0.5)

Endothelin and NK-1 receptor antagonists effectively inhibit NI in animal studies [94, 95]. In addition, triptans and ergot alkaloids inhibits NI [90, 91] (0.5)

Randomized clinical trials show no effect of substance P, neurokinin-1 antagonists [96–98], neurosteroid ganaxolone [99], endothelin antagonist [100], or specific NI blockers [101, 102] (1.0)

Aura is common in cluster headache patients [Schürks-et al-2006]^a

 Aura must be rare in cluster headache (−0.5)

Aura occurred in 4% [106],14% [103], 23% [104], and 28% [105] of cluster headache patients (0.5)

None of 554 cluster headache patients experienced aura [111] (1.0)

Hypothalamic activation is specific for cluster headache and other trigeminal autonomic cephalalgia (TAC) [133]

 Hypothalamic activation is not cluster headache specific; the Popper falsification rule [8] was used

Activation in the posterior hypothalamus during nitroglycerin-induced cluster headache attacks was observed by PET [112, 113]. In migraine without aura, no hypothalamic activation was found in two PET studies [4, 33]. In two SUNCT patients, functional MRI identified hypothalamic activation [116, 117]

Activation was observed in both the hypothalamus and brain stem (n = 7) with PET [34]