Sage Journals: Discover world-class research

Abstract

Neuroimaging methods have been widely used in headache and migraine research. They have provided invaluable information on brain perfusion, metabolism and structure during and outside of migraine attacks, contributing to an improved understanding of the pathophysiology of the disorder. Human models of migraine attacks are indispensable tools in pathophysiological and therapeutic research. This review of neuroimaging methods and the attack-provoking nitroglycerin test is part an initiative by a task force within the EUROHEAD project (EU Strep LSHM-CT-2004-5044837-Workpackage 9) with the objective of critically evaluating neurophysiological tests used in migraine. The first part, presented in a companion paper, is devoted to electrophysiological methods, this second part to neuroimaging methods such as functional magnetic resonance imaging, positron emission tomography and voxel-based morphometry, as well as the nitroglycerin test. For each of these methods, we summarize the results, analyse the methodological limitations and propose recommendations for improved methodology and standardization of research protocols.

Keywords

Methodology migraine neuroimaging neurophysiology nitroglycerin test

Functional magnetic resonance

D Magis, PS Sandór, M Sanchez del Rio, GG Schoonman & J Schoenen

We will focus on the following magnetic resonance (MR) techniques that have been used in migraine: blood oxygen level dependent functional magnetic resonance imaging (BOLD-fMRI), MR spectroscopy, diffusion-weighted and perfusion-weighted MRI.

Generalities on optimal method, personal experience

Until now the number of studies has been too small to establish an optimal protocol. Inclusion criteria should be identical with electrophysiological studies (see above), and as strict as possible, e.g. the timing of the study is critical, as neurophysiological patterns are likely to be different during premonitory, postdromal and interictal phases; a comprehensive discussion of these differences can be found in the electrophysiological paper.

There are no studies concerning the repeatability of functional MR techniques in migraine, and there are no studies correlating results with those obtained from, for example, electrophysiological techniques. Therefore, functional MR techniques are, until proven otherwise by means of proper methodological data, not to be regarded as diagnostic tests for individuals. The few studies on repeatability of electrophysiological methods in migraine (1, 2), suggest that, due to the weak signal-to-noise ratio, their results are of interest for the study only of groups, not of individuals. This might be due to the fact that the electrophysiological signal is obtained from the scalp surface. In contrast, fMRI techniques look non-invasively ‘into’ the brain and there is no such limitation; the probability is therefore higher that methodological improvements will result in a signal-to-noise ratio strong enough to be useful for diagnostic purposes.

BOLD-fMRI

Definition

The BOLD-fMRI signal, a surrogate marker of neuronal activation, was first described by Ogawa et al. (3). Changes in blood flow and the proportion of oxy- and deoxyhaemoglobin are responsible for local magnetic changes which can be detected (3). The signal increase in activated areas is thought to be related to the relative hyperaemia in these regions of neuronal activation, resulting from blood flow increase proportional to tissue oxygen demand. This technique is widely used to understand how the human brain responds to various conditions: when a brain region is used (or not), the blood flow is relatively increased (decreased) in this area.

Literature data

The literature concerning the BOLD technique is focused on the migraine attack and limited to a small number of studies (Table 1). Most of them have explored the visual cortex [visually triggered migraine (4, 5) and migraine with aura (MA) (6, 7)].

Table 1

Main fMRI studies

First Author	Year	Material			Methods	Main findings in migraineurs
First Author	Year	Nb patients & CTRL	Mean age (SD/Range)	Timing of recordings	Methods	Main findings in migraineurs
Cao	1999	2 MO 10 MA 6 CTRL	?	Visually triggered migraine—early phase of an attack	BOLD fMRI during repeated checkerboard visual stimulation (alternated red and green) Study of the occipital cortex activation	Visually triggered headache and visual change in patients with migraine is accompanied by spreading suppression of initial neuronal activation and increased occipital cortex oxygenation.
Hadjikhani	2001	3 MA	?	During visual aura (cortical spreading depression)	High-field functional MRI with near-continuous recording. BOLD signal evaluated during CSD	Data suggest that an electrophysiological event such as CSD generates the aura in human visual cortex.
Cao	2002	3 MO 23 MA 10 CTRL	?	Visually triggered migraine	BOLD fMRI during repeated checkerboard visual stimulation. Three axial image sections covering the occipital cortex and brainstem. Visual stimulation: red-green checkerboard spanned an 11.3° radius with a 2° check size, alternated colors (red and green) at 9 Hz. 3.0-T scanner. T2∗-weighted images (128 × 64 matrix over a 24-cm field of view)	Activation (hyperoxia and blood volume increase) of the red nucleus and substantia nigra in association with visually triggered symptoms of migraine.
Rocca	2003	15 MO (Min 4 MRI lesions) 15 CTRL	38 (16–60) 37.3 (26–59)	Mean time after last attack = 4.5 weeks (range 2–12)	Subjects scanned during a simple motor task (flexion-extension of 4 fingers—right hand). 1.5 T scanner, T2∗ weighted single-shot echo-planar imaging sequence. fMRI analysis using SPM99 and cluster detection. DT MRI also performed.	Functional cortical changes in patients with migraine and brain MRI abnormalities. Might be secondary to the extent of subcortical structural damage.
Vincent	2003	5 MA 5 CTRL	34 (25–51) 36.4 (28–48)	Interictally (2 to 6 months after last attack)	Visual incongruent line stimulation developped to obtain functional magnetic resonance images (fMRI). 1.5 T scanner. Functional images covering the whole brain acquired across 16 slices parallel to the plane of the anterior and posterior commissures with echoplanar BOLD techniques.	Enhanced interictal reactivity of the visual cortex in migraineurs

Legend: MO = migraine without aura, MA = migraine with aura, CTRL = healthy controls.

The studies performed during spontaneous migraine attacks showed BOLD changes compatible with a brief activation followed by prolonged neuronal depression spreading over the cortical surface at a speed of 1–4 mm/min suggestive of the phenomenon of ‘cortical spreading depression’ (6) and, correlating with the migraine aura, brainstem activation for visually triggered MA or migraine without aura (MoA) (5). Interictally, Vincent et al. (7) have found enhanced reactivity of the visual cortex compared with the control population. Rocca et al. (8) have demonstrated in patients with MoA an extended area of cortical activation during a motor task, which was correlated with the amount of white matter abnormalities.

Limitations

There is an important limitation as far as temporal resolution of BOLD-fMRI is concerned: a 2–4-s haemodynamic delay has been found between neuronal activity changes seen by electrophysiology and imaging (9). This has to be taken into account particularly for future studies in migraine searching for correlations between electrophysiology and imaging. Nevertheless, this delay should not affect the recording of repetitive events, such as cortical spreading depression.

Moreover, the exact relationship (linear or not) between the magnitude of the BOLD signal and the underlying neuronal activation remains obscure.

As far as more ‘technical’ details are concerned, BOLD-fMRI recordings can be easily artefacted by head movements, which limits the protocols that can be studied. Artefacts are also often seen in regions close to air, e.g. the sinuses and the orbitofrontal cortex.

Magnetic resonance spectroscopy

Definition

Magnetic resonance spectroscopy (MRS) of the brain is a non-invasive technique providing biochemical information about metabolites and, especially if used with a functional paradigm, neuronal function. Whereas other MRI techniques listed in this review (BOLD-fMRI, diffusion-weighted MRI and perfusion-weighted MRI) are based on ¹H ‘resonance’ (flip) in water molecules, in MRS ¹H atoms in other molecules, or other atoms such as ³¹P, are being resonated. The principle of MRS is that nuclei in different chemical structures have different resonance patterns. The molecular composition of a given brain region (called a ‘voxel’) is displayed as a spectrograph where the quantity of a substance is positively correlated with the area under the peak value. ¹H MRS allows the detection of the following compounds: N-acetylaspartate, lactate, glutamate, aspartate, GABA, choline, myoinositol and creatine, whereas ³¹P is able to measure intracellular pH, ATP metabolism and phospholipid metabolism.

Literature data

The available MR spectroscopy studies performed in migraine are listed in Table 2. The most common method used in migraine at present is ³¹P spectroscopy (10–19). More recently, ¹H spectroscopy studies (20–23) have also been performed, some of them with a functional paradigm. Ictal and, more often, interictal recordings are available, and, as far as ³¹P spectroscopy is concerned, brain and muscle studies before and after an exercise. Most (11–18, 20–23), but not all (10, 19), studies are compatible with an abnormal energy metabolism, possibly due to mitochondrial dysfunction, i.e. a reduced phosphorylation potential. Abnormal magnesium Mg2+ concentrations have been found in all migraine types by one study (18). In another trial (19), a significant decrease of Mg2+ was seen in hemiplegic migraine patients only, with a trend to abnormality in other migraine subgroups. Although the studies on brain metabolism were performed by different research groups, all studies on muscle metabolism in migraineurs came from the same group and have yet to be confirmed by others. MR spectroscopy abnormalities are found in all migraine subgroups, including MoA (15, 17, 18), and seem proportional to the severity of the migraine phenotype (17–19, 21).

Table 2

Main spectroscopy studies

First Author	Year	Material			Methods		Main findings in migraineurs
First Author	Year	Nb patients & CTRL	Mean age (SD/Range)	Timing of recordings	Spectroscopy type & target area	Measured components	Main findings in migraineurs
Welch	1988	? ‘Migraine’ ? CTRL	?	During attack	31P	pHi	No change in pHi during attack
Barbiroli	1990	8 MPA, MS & AMEA No CTRL	?	?	31P Brain and muscle	PCr, Pi, PCr/Pi ratio	Mitochondrial metabolic defects present in platelets and muscle are also expressed in the brain.
Bresolin	1991	1 MS with AMEA No CTRL	40	?	31P Brain and muscle	PCr, Pi	Energy metabolism defect in brain and muscle.
Barbiroli	1992	12 MA ? CTRL	?	Interictally	31P Brain and muscle	PCr, ADP, RR ATP	Low phosphocreatine and high ADP content in all patients Abnormal muscle mitochondrial function in 9 patients
Sacquegnia	1992	1 MPA ? CTRL	?	? Period Before and after exercise	31P Brain and muscle	PCr, ATP, Pi, ADP, pH	Impairment of energy oxidative metabolism both in brain and muscle in MPA
Montagna	1994	22 MO ? CTRL	?	Attack-free Before and after exercise	31P Brain and muscle	PCr, ADP, PP	Energy metabolism is abnormal in MO
Uncini	1995	1 FHM No CTRL	?	? Period Before and after exercise	31P Brain and muscle	PCr, ADP, PP, ATP	Defective energy metabolism of brain and muscle suggesting a multisystemic disorder of mitochondrial function in this FHM pedigree
Watanabe	1996	6 ‘Migraine’ 6 CTRL	?	Interictally	1H Visual cortex	Lac	Anaerobic glycolysis occurs in the brains of migraineurs interictally
Lodi	1997	18 MS or MPA 22 MA 23 MO 14 CH ? CTRL	?	Attack-free Before and after mixed glycolytic/aerobic exercise	31P Skeletal muscle	PCr, pHi, ATP, ADP	Abnormal mitochondrial function in complicated migraine at rest After exercise smaller pH decrease in migraineurs Mitochondrial impairment proportional to phenotype severity in migraine
Lodi	2001	7 MS 13 MPA 37 MA or BM 21 MO 13 CH No CTRL	38 (3) 25 (3) 23 (2) 32 (2) 39 (2)	Attack-free or out of cluster period	31P 1.5T scanner Occipital lobes	B-ATP, ATP hydrolysis, Mg2+, ADP, PCr, CrK	In all groups: reduction of Mg2+ and free energy In migraine: lack of energy release proportional to phenotype severity
Boska	2002	? MA ? MO ? FHM ? CTRL	?	Interictally	31P 3T scanner	PDE, Mg2+, PCr, PME, Pi, pH	No significant abnormalities of brain energy metabolism Disturbances of Mg2+ metabolism
Dichgans	2005	15 FHM 1 with CACNA1A mutation 17 CTRL	37.1 (16–63) 34.6 (22–53)	?	1H 1.5 T scanner Brain (superior cerebellar vermis, occipital and parietal cortex)	mI, Cho, Cr, NAA, Lac, Glu, BPF	Neuronal impairment in the superior cerebellar vermis: localized atrophy (BPF) and NAA levels Reduction in Glu may reflect impaired glutamatergic neurotransmission
Sarchielli	2005	22 MO 22 MA 10 CTRL	?	Interictally Baseline and during/after photic stimulation	1H Visual cortex	NAA, Lac	Less efficient mitochondrial functioning in MA patients.
Sandor	2005	5 MA 5 MA+ 11 CTRL	29.2 (6.6) 37.5 (11.4) 26.5 (14.0)	Interictally (72 h from last attack) Baseline & during 8-Hz checkerboard stimulation	1H 1.5 T scanner Visual and non-visual cortex	NAA, Lac, Cr, PCr, Cho	Abnormal metabolic strain in MA+ during stimulation, and predominant mitochondrial dysfunction in MA

Legend: MO = migraine without aura, MA = migraine with aura, MA+ = migraine with aura plus paraesthesia, paresis or dysphasia, MPA = migraine with prolonged aura, MS = migraine stroke, BM = basilar migraine, FHM = familial hemiplegic migraine, CH = cluster headache. AMEA = abnormal mitochondrial enzyme activity, CTRL = healthy controls. ADP = adenosine diphosphate, ATP = adenosine triphosphate, BPF = brain parenchyma fraction, Cho = choline, Cr = creatine, CrK = creatine kinase, Glu = glutamate, Lac = lactate, Mg2+ = magnesium, mI = myoinositol, NAA = N-acetyl-aspartate, PCr = phosphocreatine, PDE = phosphodiester, pHi = intracellular pH, Pi = inorganic phosphate, PME = phosphomonoester, PP = phosphorylation potential, RR ATP = relative rate of ATP.

Limitations

As the number of the molecules studied with MRS is much lower than that of water molecules, wider areas and longer periods are needed to detect a signal. There is consequently a limitation of spatial and temporal resolution in comparison with conventional MRI. However, new techniques are being developed to reduce this constraint. Moreover, MRS is limited to mobile compounds and therefore cannot be used for the study of neurochemical receptors.

Diffusion-weighted MRI

Definition

Diffusion-weighted (DW) MRI is based on random ¹H movement in water molecules (24). The apparent diffusion coefficient (ADC) is the amount of water diffusion for a given pixel. Regions with low ADC will appear hyperintense. DWMRI is mostly used in the early diagnosis of acute cerebral ischaemia (24). It can detect a decrease in the ADC of the ischaemic brain, correlated with cytotoxic oedema and neuronal energy loss.

DWMRI can also be helpful in the detection of myelination (25).

Literature data

Most publications using DWMRI in migraine are case reports performed during spontaneous or persistent auras (26–28) and report normal findings (Table 3). Recently, Rocca et al. (8, 29), used diffusion-tensor imaging in migraine patients with and without MRI white matter lesions. They found that migraine patients had abnormalities of normal appearing white matter which were mild compared with other diseases associated with white matter abnormalities (29).

Table 3

Main Diffusion-weighted and Perfusion-weighted MRI studies

First Author	Year	Material			Methods		Main findings in migraineurs
First Author	Year	Nb patients & CTRL	Mean age (SD/Range)	Timing of recordings	MRI sequencies 4& target area	Measured components	Main findings in migraineurs
Cutrer	1998	4 MA No CTRL	?	Spontaneous aura	PW and DW Occipital cortex	Cerebral blood flow Cerebral blood volume Tissue mean transit time	No change in diffusion during aura.
Sanchez del Rio	1999	22 MA & MO No CTRL	?	During aura Headache after aura During MO	PW Occipital cortex	Cerebral blood flow Cerebral blood volume Tissue mean transit time	Decremental blood flow changes in occipital lobe most characteristic of MA No change in MO
Gutschalk	2002	1 FHM No CTRL	?	Prolonged hemiparesis and aphasia	PW and DW Whole brain ?	?	No change in diffusion during FHM hemiparesis.
Smith	2002	1 MPA (stopped prophylaxis) No CTRL	66	48h and 2 days after onset of aura	PW and DW Whole brain	Cerebral blood flow Cerebral blood volume Tissue mean transit time	No change in diffusion
Rocca	2003	28 MO 6 MA with / without brain lesions 17 CTRL	35.2 (18–55) 36 (21–55)	? 11 patients had treatment.	DTMRI Whole brain	Mean diffusivity and fractional anisotropy of normal appearing brain tissue (NABT)	Brain damage may extend beyond T2 weighted abnormalities in patients with migraine, but mild severity compared with other diseases with white matter abnormality.
Rocca	2003	15 MO with min 4 MRI lesions 15 CTRL	38 (16–60) 37.3 (26–59)	Interictally (range 2-12 weeks after last attack)	DTMRI BOLD fMRI 1.5T scanner DT: whole brain fMRI: primary sensorimotor cortex, supplementary motor area, secondary somatosensory cortex	Measures: volume, average mean diffusivity, average fractional anisotropy of all lesions seen on the dual-echo scans. Histograms of the normal-appearing white matter.	Functional cortical changes in patients with migraine and brain MRI abnormalities. Might be secondary to the extent of subcortical structural damage.
Relja	2005	1 MPA No CTRL	43	During persistent visual aura onset and 7 months after (treatment with lamotrigine)	PW (1.5 T scanner) and SPECT Grey matter of temporal, occipital, parietal and frontal lobes	Signal intensity vs. Time curves	Cortical hypoperfusion during persistent aura

Legend: MO = migraine without aura, MA = migraine with aura, MPA = migraine with prolonged aura, FHM = familial hemiplegic migraine, CTRL = healthy controls.

Limitations

DWMRI provides information about the effect of a recent ischaemic insult, but does not give information concerning the integrity of the vascular bed.

Perfusion-weighted MRI

Definition

Perfusion-weighted (PW) MRI is used to assess the blood supply of the brain (blood flow, blood volume and tissue mean transit time), i.e. the rate at which blood is delivered to the tissue. There are two main approaches: bolus tracking (more frequently employed, using an injected gadolinium-containing compound) and arterial spin labelling (without contrast product infusion).

Literature data—method

The literature on PWMRI is limited to case reports of MA, prolonged aura or familial hemiplegic migraine (FHM); partially contrasting results have been reported (Table 3). Cutrer et al. (26) and Relja et al. (30) have found cortical hypoperfusion during the aura, whereas in the study of Smith et al. (28) prolonged aura was associated with hyperperfusion and vasogenic leakage. Gutschalk et al. (27) performed PWMRI during a prolonged attack of FHM and found no change in perfusion during hemiparesis. Sanchez del Rio et al. found no change of perfusion in MoA (31).

Limitations

The bolus-tracking approach using intravenous gadolinium infusion cannot be repeated as often as less harmful techniques. Moreover, in migraine, PWMRI seems to be of interest only during or immediately before/after the attack. In the bolus-tracking method, the presence of a reduced arterial diameter may cause bolus dispersion, which has been shown to introduce significant errors in cerebral blood flow (CBF) estimations. It remains to be determined if this could explain part of the contradictory results seen in migraine studies.

Positron emission tomography

S Afridi & PJ Goadsby

Definition

Positron emission tomography (PET) is a tomographic nuclear imaging procedure, which uses positrons as radiolabels and positron-electron annihilation reaction-induced gamma rays to locate the radiolabel. It can provide a quantitative measurement of blood flow. The PET scanner consists of multiple detector rings to scan a number of transaxial planes simultaneously. In H₂ ¹⁵O PET, water is labelled with a positron emitter (¹⁵O) and injected intravenously into the patient, who is scanned by the tomographic system. The PET images are coregistered with a structural scan, usually an MRI, to enable accurate anatomical localization.

Literature data

Main PET studies mentioned below are summarized in Table 4.

Table 4

Main PET studies

First Author(s)	Year	Material			Method	Main findings in migraineurs
First Author(s)	Year	Nb patients & CTRL	Mean age (SD/Range)	Timing of recordings	PET tracers & measured components	Main findings in migraineurs
Woods	1994	1 ‘migraine’ No CTRL	?	During spontaneous attack	?	Bilateral spreading cerebral hypoperfusion during spontaneous migraine headache
Weiller	1995	9 MO No CTRL	29–57 years	During spontaneous attack and after pain relief with sumatriptan injection	rCBF	During attack increased CBF in cingulate, auditory and visual association cortices and in brain stem. After pain relief, persistance of an isolated brain stem activation
Andersson	1997	4 MO 6 MA	?	During headache, aura, after pain relief with sumatriptan and headache-free state	rCBF rCMRO₂ rOER	Significant reduction in rCBF and rCMRO₂ in primary visual cortex but no change in rOER during the headache phase Cerebral ischemia is not the primary cause of the attacks in these cases
Bednarczyk	1998	9 MO No CTRL	?	During spontaneous attack and interictally (48 hrs)	CBF CBV H₂ ¹⁵O C¹⁵O ¹⁵O₂	In MO, reduced CBF and CBV during the headache phase Cerebral oxygen metabolism and oxygen extraction not significantly affected
Bahra	2001	1 patient with cluster headache and migraine No CTRL	?	Typical MO after GTN	?	Activation in the dorsal rostral brainstem No activation in the region of the hypothalamus
Denuelle	2004	? ‘migraine’ ? CTRL	?	During spontaneous migraine attack	?	Brainstem and hypothalamic activation during spontaneous attack
Matharu	2004	8 patients with chronic migraine and improvement with bilateral suboccipital stimulation No CTRL	?	Stimulator switched OFF, ON and partially activated	rCBF	Significant changes in rCBF in the dorsal rostral pons, anterior cingulate cortex (ACC) and cuneus, correlated to pain scores, and in the ACC and left pulvinar, correlated to stimulation-induced paraesthesia scores The dorsal rostral pons may be a locus of neuromodulation by suboccipital stimulation
Afridi	2005a	5 ‘migraine’ No CTRL	?	During spontaneous attack and interictally	H₂ ¹⁵O	Significant activation in the dorsal pons, lateralized to the left Activation also seen in the right anterior cingulate, posterior cingulate, cerebellum, thalamus, insula, prefrontal cortex, and temporal lobes Area of deactivation in the migraine phase also located in the pons, lateralized to the right
Afridi	2005b	1 MA No CTRL	54	Migraine triggered by GTN	H₂ ¹⁵O	Activation in the primary visual area of the occipital cortex demonstrated during the aura
Afridi	2005c	16 MO 8 MA 8 CTRL	? (26–65)	Before, during and after GTN induced migraine	H₂ ¹⁵O	Lateralized dorsal pontine activation ipsilateral to headache side
Denuelle	2005a	7 MO No CTRL	?	During spontaneous attack, after pain relief with sumatriptan injection, and at pain-free state	H₂ ¹⁵O	Hypothalamic activation persisting after headache relief
Denuelle	2005b	7 MO No CTRL	?	During spontaneous attack, after pain relief with sumatriptan injection, and at pain-free state	H₂ ¹⁵O	Significant bilateral posterior cortical hypoperfusion persisting after headache relief
Fumal & Laureys	2006	16 CMOH 68 CTRL	42.5 (11) 45 (21)	Before and after analgesic withdrawal	¹⁸FDG	Medication overuse headache associated with persistent orbitofrontal hypofunction that could predispose subgroups of migraineurs to recurrent analgesic overuse

Legend: MO = migraine without aura, MA = migraine with aura, CMOH = chronic migraine with analgesic overuse, CTRL = healthy controls. CBF = cerebral blood flow (rCBF if regional); CBV = cerebral blood volume; CMRO2 = oxygen metabolism; OER = oxygen extraction.

Spontaneous migraine

In one of the earliest PET reports Woods et al. described a case report of a subject with migraine with no previous aura who unexpectedly developed a migraine during her participation in a visual activation paradigm whilst lying in a PET scanner (32). The subject described some visual blurring during one of the scans, but did not clearly describe any other features of typical aura. The migraine was associated with bilateral hypoperfusion starting in the occipital lobes and spreading anteriorly into the temporal and parietal lobes. The contiguous spread covered areas in the territories of the posterior and middle cerebral artery. In a later study, nine subjects were studied within 13 h of onset of MoA (33). They observed a 9.9% decrease in global CBF and a 5.2% decrease in cerebral blood volume persisting for at least 6 h. Oxygen metabolism and oxygen extraction remained unchanged. Most recently, a further study of spontaneous migraine has also reported bilateral posterior circulation hypoperfusion in MoA (34). These changes may represent hypoperfusion associated with cortical spreading depression (35), albeit without any clinical manifestations. Alternatively, they may represent a neurogenically mediated oligaemia, which may be seen in experimental animals with activation of the nucleus locus coeruleus (36) and is typically dominant in the posterior cortical structures (37).

The first PET study detailing regional activation during migraine was the study of Weiller and colleagues (38). Nine subjects with MoA presented with right-sided spontaneous headaches and were scanned within 6 h of onset of migraine, prior to having taken any medication. They were scanned during spontaneous migraine attacks, following sumatriptan and interictally. The study revealed brainstem activation during the migraine, which persisted after sumatriptan administration had relieved the pain. The resolution of the PET camera used was not high enough to identify specific nuclei, but the foci of maximum increase were around the dorsal midbrain and dorsolateral pons. Activation was also seen in the anterior cingulate, visual and auditory association cortices.

In a more recent study involving both MA and MoA, five patients were imaged in ictal and interictal states and the differences were analysed using statistical parametric mapping (39). Two patients had a typical migrainous aura prior to the onset of the headache. All the attacks studied fulfilled standard diagnostic criteria for migraine (40). Comparing the migraine scans with interictal scans there was significant activation in the dorsal pons, lateralized to the left, irrespective of the anatomical location of the pain. Activation was also seen in the right anterior cingulate, posterior cingulate, cerebellum, thalamus, insula, prefrontal cortex and temporal lobes.

Triggered migraine

A PET study by Andersson and colleagues (41) has investigated attacks provoked by red wine in 11 subjects with MA and MoA. They demonstrated reductions in regional CBF (23%) and in oxygen metabolism (22.5%) in an area corresponding to the primary visual cortex. They did not detect any significant increases in blood flow during aura or headache.

A case of a glyceryl trinitrate (GTN)-triggered migraine also revealed brainstem activation, on this occasion in the dorsolateral pons, which again persisted following abortion of the migraine (42).

The largest PET study to date involved 24 migraineurs (with and without aura) and eight healthy controls (43). The migraineurs were divided into three groups according to the site of their headache: right/left/bilateral. In each group a migraine was induced using a GTN infusion. The subjects were scanned at various points (preinfusion, during GTN, during migraine, post migraine). Significant brainstem activation was seen in the dorsal pons during the migraine state vs. the pain-free state when comparing migraineurs with controls. When each group was analysed separately to investigate laterality it was found that the dorsal pontine activation was ipsilateral in the right-sided and left-sided groups and bilateral in the bilateral headache group with a left-sided preponderance.

Chronic migraine

A PET study has examined eight patients with chronic migraine who had had bilateral suboccipital stimulators implanted. The authors reported activation in the dorsolateral pons that persisted when the stimulator was on and the patients had pain relief (44). The area was the same as demonstrated in the episodic migraine cases described above, suggesting that both chronic and episodic migraine share important common biological features in terms of areas of brain activation. Remarkably, the main difference between the pain state and the pain-free state was a change in the pattern of thalamic activation.

Limitations

Although the temporal and spatial resolution provided by PET is relatively good, it is lower than for fMRI. The spatial resolution of PET is dependent upon the position and type of detectors. The temporal resolution is dependent on the time taken for the tracer to reach the brain and on the half-life of the tracer (124 s for ¹⁵O). This enables a relatively good comparison of prolonged brain states such as migraine vs. pain free, but is not fast enough to follow, for example, progression of migraine aura.

Another limitation of PET is that during the scanning process ionizing radiation is administered, limiting the number of PET scans per subject.

Voxel-based morphometry

A May

Definition

Voxel-based morphometry (VBM) is a whole-brain technique capable of investigating subtle, regionally specific changes in grey matter by averaging across subjects. It gives information about density and volume. VBM has been cross-validated with region-of-interest measurements and functional data in a number of studies (45–47). This method is based on high-resolution structural T1-weighted 3D MR images, registered in a common stereotactic space, and is designed to find significant regional differences throughout the brain by applying voxel-wise statistics in the context of Gaussian random fields (48).

Literature data

Unlike functional data using PET and fMRI, where a substantial amount of data is available, there are to date only four VBM papers on headache (49) (Table 5), two on chronic back pain (50, 51) and one on phantom pain (52). Of the headache studies, one studied patients suffering from cluster headache (47), one MA and MoA patients (53) and the third chronic tension-type headache (TTH) and medication overuse headache (54). All three studies compared patients with healthy controls (cohort studies). In cluster headache a significant unilateral increase of grey matter in the posterior inferior hypothalamus was found. In TTH, but not in patients with medication overuse headache, a significant negative alteration in grey matter was found in the pain matrix (cingulate cortex, anterior insulae, brainstem) (54). However, no significant morphometric changes were found in grey or white matter between patients suffering from episodic migraine and healthy controls. In contrast, a recent study of VBM in migraineurs with white matter T2 hyperintensities has shown decreased grey matter density in frontal and temporal areas, but increased densities in periaqueductal grey matter and dorsolateral pons. Reduced, but not increased, grey matter density was strongly related to age, disease duration and T2-visible lesion load (55).

Table 5

Main VBM studies

First Author	Year	Material		Main findings in migraineurs
First Author	Year	Nb patients & CTRL	Mean age (SD/Range)	Main findings in migraineurs
May	1999	29 Cluster headache not migraine	47 (25–74)	Structural abnormalities in hypothalamic region, correlation with PET findings (cluster headache not migraine)
Matharu	2003	17 MO 11 MA 28 CTRL	34 (24–57) 31 (23–52) Matched CTRL	No difference of brain global and regional macroscopic structure between MA, MO and CTRL
Rocca	2006	16 migraine with T2 abnormalities 15 CTRL	?	Structural GM abnormalities can be detected in migraine patients with brain T2-visible lesions using VBM and a high-field MRI scanner.
Schmidt-Wilcke	2006	20 MOH, 20 cTTH, 40 CTRL	MOH 34 MOH 40 CTRL 38	Structural GM abnormalities in the pain matrix can be detected in patients with tension type headache but not in patients with MOH or episodic migraine.

Legend: MO = migraine without aura, MA = migraine with aura, MOH = medication overuse headache, CTRL = healthy controls.

Advantage of the method

Given that cluster headache and migraine have subtly different brain structures, the question arises whether other primary headache types may also be distinguishable on a morphological level. The advantage of VBM studies is that the method is non-invasive (T1-weighted MRI scans) and that it can be replicated in the same individual at different time points, e.g. during an attack or after pharmacotherapy.

Limitations of the method

VBM has still to prove that it generates reproducible results. VBM is for research purposes only and requires groups of at least 20 subjects per group to produce stable results. VBM is sensitive for changes in grey matter, and recent cross-sectional VBM studies have demonstrated learning-dependent changes in adult human brain and suggested anatomical correlates for navigation, arithmetic, linguistic, procedural and musical learning abilities (56–60). However, VBM is not sensitive for white matter changes, where another morphometric method, diffusion tensor imaging, would be the method of choice (61). Because of the limited knowledge regarding the cellular substrate detectable with VBM, we can only speculate about the nature of the structural changes. VBM would detect changes of grey matter concentration per voxel as well as changes of the classification of individual voxels, e.g. from white to grey matter (62) and is probably a combination of both. In general, an increase of grey matter could be due to a simple increase in cell size, neural or glial cell genesis, spining and even changes in blood flow or interstitial fluid. Finally, the need to average in Talairach's coordinates, i.e. the standard reference for brain locations in the scientific literature, may lead to include ‘mixed’ voxels containing grey and white matter as well as cerebrospinal fluid in certain areas. Independently of these limitations, there is a need for better standardization between different studies and centres (63).

Recommendations for improved methodology

To date, no standardized method allows comparisons between laboratories. If the software package Statistical Parametric Mapping (SPM) is used, the optimized protocol by Good et al. (62) is relevant.

The following information should be included in the Methods section of published studies:

•

Data acquisition: scanner type and voxel volume.

•

Data preprocessing: which SPM. Whether or not the optimized protocol by Good et al. (62) has been used. Choice of template. Information regarding modulation of volume changes. Smoothing factor.

•

Data analysis: if small volume correction is used: information about the choice of sphere volume.

•

Information about grey and white matter changes in both directions (increases or decreases).

Which optimal methodology can be proposed in headache studies?

In clinical practice, to obtain reliable and reproducible VBM data, the following method is recommended:

For cohort studies: n > 15 per group, better >20 per group

Age and gender matching (variables influencing cortical and subcortical anatomy)

Homogeneous groups, optimum: genetically determined cohorts

Always the same scanner with identical imaging parameters

Data preprocessing: optimized protocol (62, 64)

Smoothing factor of 10 mm full width at half maximum

Longitudinal analysis (same patient over time) better than cohort analysis (patients vs. controls)

Additional regression analysis using external parameters (such as electrophysiological data, etc.) to better fit the model of the a priori hypothesis.

Studies needed

•

Migraine has a strong genetic component and, to render groups as homogeneous as possible, the ideal inclusion criteria could be based on genotype (cohort study) or response to treatment (longitudinal study including controls)

•

Standardization of protocols is needed

•

A way to compare directly groups investigated on different scanners is needed (65)

•

Histological data of the structures identified with VBM is needed

•

External parameters (neurophysiological data see evaluation and proposal for optimization of neurophysiological tests in migraine) need to be compared to changes in brain structure.

Migraine-provoking test

L Bendtsen, H Kaube & G Sandrini

Genuine migraine is difficult to study because of its attack-wise appearance. A large number of migraine-provoking agents have been described, including nitroglycerin (NTG), histamine, prostacyclin, reserpine, m-clorophenylpiperazine, adenosine, tyramine-containing foods, chocolate, sustained tooth clenching (for review see (66)), calcitonin gene-related peptide (CGRP) (67) and sildenafil (68). However, NTG is the only substance for which there is sufficient information on reproducibility and reliability.

Definition

GTN is widely used for its antianginal effect, and it has been known for >150 years that one of its most common side-effects is headache. It is believed that GTN is metabolized to nitric oxide and/or S-nitrosothiols, and that these metabolites cause the headache induced by GTN (66). GTN has been used as a migraine-provoking substance for more than five decades, and a specific human experimental model of migraine using GTN has been developed and tested in a series of studies by Iversen and colleagues within the last two decades (66, 69). More recently, other groups have adopted this model (43, 70–73).

Literature data

The main literature is listed in Table 6.

Table 6

Main migraine-provoking tests studies

First Author	Year	Material		Methods			Main findings
First Author	Year	Nb patients & CTRL	Mean age (SD/Range)	NTG/GTN (way, dose & time)	Total time of follow-up after NTG	Recordings	Main findings
Sicuteri	1987	24 ‘migraine’ parents 50 children with/ without familial history of migraine 36 CTRL parents	45.32 (12.54) 24.74 (14.18) 25.83 (13.60) 48.09 (13.92)	SL 0.6 mg	12 hours	None	Delayed headache after NTG in 28.6% of CTRL with familial history of migraine but not in CTRL without familial history 66.7% of parents with migraine had delayed headache Delayed NTG-induced headache seems similar to migraine headache
Zanette	1991	?MO ? CTRL	?	SL	?	TCD with measure of systolic, diastolic and time-mean blood flow velocity and pulsatility index of main cerebral arteries before and after NTG administration.	In MO, marked response to NTG with decrease in systolic and time-mean blood velocity and pulsatility index In CTRL, only time-mean velocity decrease.
Thomsen	1993	17 MO 9 ETTH 17 CTRL	?	IV 0.015, 0.03 & 0.25 µg/kg/min 15 min each	?	TCD with measure of blood flow velocity of middle cerebral artery before, at the end of each infusion and 30 and 60 min after the last infusion	Decrease of blood velocity with increasing GTN dose, more pronounced in migraine for 2 higher doses NO sensitivity could be important in mechanism of migraine pain
Iversen	1996	No migraineurs 10 CTRL	?	IV 0.12 µg/kg/min 20 min	?	TCD with measure of blood flow velocity of middle cerebral artery, ultrasound measure of diameter of radial and temporal arteries. Measure before and after GTN, and after treatment of headache with either 6 mg subcutaneous sumatriptan or placebo	Sumatriptan reduced intensity of GTN-induced headache in CTRL Blood velocity in middle cerebral artery unaffected by sumatriptan but decreased diameter of radial and temporal arteries
Thomaides	1996	19 MO 19 CTRL	35 (6)	SL 1 mg	?	Electroencephalographic recordings before NTG, during headache phase and 30 min after 6 mg subcutaneous sumatriptan	In MO, increase of theta and delta activities during headache and decrease after sumatriptan
Christiansen	1999	12 MA 14 CTRL	?	IV 0.5 µg/kg/min 20 min	? At least 240 min	None	Aura symptoms not elicited in any subject In 6 patients with MA, typical MO fulfilling IHS criteria after 240 min
Fullerton	1999	No migraineurs 10 CTRL	?	IV 0.5 µg/kg/min 30 min	?	TCD with measure of blood flow velocity of middle cerebral artery Measure before and after GTN, and after treatment of headache with either 2 or 6 mg subcutaneous sumatriptan	9 of the 10 CTRL developed headache Compared with baseline: decrease of blood flow velocity with GTN, increase with sumatriptan GTN would be a suitable model to study effect of migraine drugs
Christiansen	2000	29 MO 14 CTRL	?	IV 0.5 µg/kg/min	?	None	No influence of attack frequency on GTN triggered headache In migraineurs, more severe attack and more often fulfillment of the IHS criteria
Ashina	2000	16 CTTH 16 CTRL	?	IV 0.5 µg/kg/min 20 min	12 hours 1 week	None	NO-induced biphasic response with immediate and delayed headache is common to chronic tension-type headache and migraine The delayed headache has the characteristics of the primary headache disorder
Bank	2001	No migraineurs 1 CTRL	46	SL	2 hours	None	Typical migraine with aura triggered by NTG in a subject without personal or familial history of migraine
Bednarczyk	2002	No migraineurs 12 CTRL	27.5 (8.7)	IV 0.125, 0.25 & 0.5 µg/kg/min 15-min steps	? At least 2 hours	TCD and CBF H₂ ¹⁵O PET measurements at baseline, after each GTN step and 15, 30 and 60 min after sumatriptan 6 mg sc injection	GTN increased global CBF while flow velocities decreased. Regions of increased flow included cingular cortex, regions of decreased flow included occipital cortex. Regional changes are similar to spontaneous migraine, but global CBF and velocities effects are different. No effect of sumatriptan.
Thomaides	2003	45 MO 15 CTRL	38.8 (2.3) 39.1 (3.4)	SL 1 mg	? At least 5 days	TCD with measure of blood flow velocity and pulsatility index of middle cerebral artery Several sessions:before NTG, at maximum pain, every 30 min for 2 hours after oral treatment with either zolmitriptan or sumatriptan	Significant decrease of blood flow velocity during headache in zolmitriptan, sumatriptan and untreated migraine group but not in the CTRL group. One hour after triptan intake, disappearance of these abnormalities. No effect on pulsatility index.
Kowacs	2003	No migraineurs 8 CTRL	? (22–33)	IV 0.5 µg/kg/min 30 min	? 45 min to 105 min	Measure of nociception specific blink reflex R2 area under the curve before and at 15 and 45 min after the start of infusion of GTN or placebo	No significant effect of GTN on blink reflex's R2 response during early headache phase 6 subjects had mild immediate headache but none had delayed headache.
Afridi	2004	23 MO 21 MA 12 CTRL	46 (25–68) 44 (27–65) 38 (21–58)	IV 0.5 µg/kg/min		None	32 patients had a MO attack and 1 had a typical MA attack 12 patients reported typical premonitory symptoms A repeat attack was triggered in all subjects but one
de Tommaso	2004	9 MO No CTRL	35.6 (9.5)	Oral 0.6 mg	? At least 72 h	CO2 laser evoked potentials (hand and forehead stimulation) 2 sessions: mean 45 min after onset of moderate to severe headache and at least 72h after its resolution	Attack triggered in all subjects Reduction of pain threshold and increase of LEP amplitude during headache phase NTG would be able to induce central sensitization phenomenon
Tvedskov	2004a	12 MO Treatment with valproate 1000mg/day or placebo No CTRL	? (18–65)	IV 0.25 µg/kg/min 20 min	12–24h	Ultrasound of superficial temporal and radial arteries, TCD of both middle cerebral arteries.	After valproate: significant reduction of migraine induced by GTN, reduction of mean headache peak intensity, no effect on diameter of temporal and radial arteries before and after GTN, reduced velocity in middle cerebral arteries after GTN.
Tvedskov	2004b	19 MO Treatment with 160 mg propranolol or placebo 16 CTRL	34.3 (22–61) 35.8 (22–59)	IV 0.25 µg/kg/min 20 min	12–24h	Ultrasound of superficial temporal and radial arteries, TCD of both middle cerebral arteries.	All MO and CTRL developed headache after GTN, 8 MO and 2 CTRL fulfilled IHS 1.1 criteria Propranolol treatment induced radial artery constriction in CTRL only No effect of propranolol on GTN induced headache and cerebral arteries GTN would act at a deeper level of the pathophysiological cascade
Afridi	2005a	16 MO 8 MA 8 CTRL	? (26–65) ? (21–55)	IV 0.5µg/kg/min 20 min	Up to 4 hours	H₂ ¹⁵O PET at baseline, during non-specific headache phase, during typical migraine headache and after pain relief with sumatriptan 6 mg SC	Non-specific headache in all MIG and 6 CTRL, evolution to migraine in all MIG Activation of the dorsolateral pons in clinically typical migraine triggered by GTN, persistance after pain relief with sumatriptan
Afridi	2005b	1 MA No CTRL	54	IV 0.5µg/kg/min 20 min	? 165 min	H₂ ¹⁵O PET at baseline, during typical visual aura and headache phase	Activation of primary visual area in occipital cortex during aura First case of MA triggered reproductibly with GTN
Di Clemente	2005	No migraineurs 10 CTRL	?	SL 1.2 mg	5 hours	Measure of nociception specific blink reflex R2 area under the curve before,1 hour and 4 hour after NTG	Reduced pain threshold at 1 and 4 hours after NTG, and reduced R2 latency and increased R2 AUC at 4 hours after NTG. Suggests NTG is able to produce delayed central trigeminal sensitization in CTRL

Legend: MO = migraine without aura, MA = migraine with aura, TTH = tension-type headache (Episodic or Chronic), CTRL = healthy controls. IV = intravenous, SL = sublingual. TCD = transcranial doppler.

Iversen et al. have reported acceptable day-to-day variation in headache intensity experienced by healthy volunteers following intravenous GTN. Out of 50 paired infusions, only five differed by more than one score when scoring maximal headache intensity on an 0–10 scale (74). The standard deviation of the day-to-day difference of headache scores was 0.7, which allows detection of a difference of one headache score, with type 1 and 2 errors of 5%, in a crossover study of 12 subjects (66). Headache characteristics were also reproducible. In 10 healthy subjects headache quality did not differ and in one subject only the localization differed slightly between days (74). Thus, NTG induced headache with acceptable reproducibility in healthy subjects.

Olesen et al. compared GTN-induced headache responses in 17 sufferers with migraine, 17 headache-free subjects and nine patients with episodic TTH (75). Migraineurs developed an immediate headache within the first hour after infusion that most often did not fulfil the criteria for migraine, but 11 of the 17 migraineurs developed a delayed typical migraine within the next 24 h. Only one of the controls developed a delayed migraine. Thomsen et al. (76) infused GTN in 10 patients suffering from MoA. Eight patients developed immediate headache, only one of which fulfilled migraine criteria, whereas eight patients developed a regular migraine attack within 12 h. Peak migraine headache occurred at a mean of 5.5 h after the infusion. Pain characteristics and accompanying symptoms were very similar to the spontaneous attacks recorded by the patients. Christiansen et al. induced migraine attacks, but no aura, in six of 12 MA patients (77). Juhasz et al. have reported that sublingual NTG induced attacks of MoA in 10 of 15 patients with MoA and in two of eight controls (70). Afridi and colleagues (43) challenged 23 patients without and 21 patients with aura with GTN. Migraine attacks were triggered in 33 (75%) of the patients (83% in MoA and 67% in MA). Only one of these attacks was MA. Twenty-nine of the 33 patients who developed migraine were re-challenged at least a week later, and all but one developed a repeat attack. In one patient a visual aura was triggered on both occasions. Thus, NTG induced MoA with high reproducibility both in MoA and in MA.

Sances et al. have studied the reliability of NTG-induced migraine in a large study including 197 migraineurs, 175 without and 22 with aura (71). The study was methodologically well-performed. Thus, all diagnoses were made on the basis of a diagnostic headache diary kept during a 12-week run-in period, and the NTG test was performed by an observer blinded to the diagnosis. Nitroglycerin was given sublingually. The authors calculated sensitivity (the probability that a test will be positive in a subject affected by a given disease), specificity (the probability that the test will be negative in a subject not affected by a given disease) and accuracy (proportion of true-positive and true-negative tests out of total number of tests performed). A total of 138 of 175 patients with MoA (82.1%) developed MoA within the 8-h observation period, and the same was true for two of 53 healthy controls. In MoA the test sensitivity was 82.1%, specificity 96.2% and accuracy 85.5%. The reliability of the NTG test was less satisfactory in patients with MA, where only three of 22 patients developed MA (sensitivity 13.6%, specificity 96.2% and accuracy 72.0%). Thus, reliability of the NTG test was satisfactory in MoA but not in MA.

Advantages of the method and perspectives

There are a number of practical advantages of the NTG migraine-provoking test, together with characteristics of the induced attack very similar to the genuine attack that make it an interesting tool for the exploration of migraine pathophysiology. First, it is reliable, reproducible and critically evaluated by several independent groups. Second, it is able to trigger premonitory symptoms (yawning, tiredness, irritability, neck stiffness, frequency of urination, hunger and low mood) similar to those of the genuine attack (78). Third, the triggering aspect of the test bears resemblance to genuine attacks which are also often triggered, e.g. by stress, hormonal factors or alcohol. Fourth, it responds to acute (79) and prophylactic medications (80) known to have an effect in genuine migraine. Fifth, the efficacy of nitric oxide inhibitors in acute migraine (81) demonstrates that nitric oxide Key molecule in migraine. Thus, the NTG test provides excellent opportunities for prospective studies of the migraine attack, given that the limitations mentioned above are considered when interpreting the findings.

Premonitory symptoms can be studied prospectively (43) and central processes can be studied during the attack, e.g. central pain processing by means of the nociception-specific blink reflex (82), auditory evoked potentials (83), visual evoked potentials (84) and laser evoked potentials (72). Imaging studies may provide information about temporal and spatial activation of various pain-sensitive centres in the brain (43, 73), and biochemical studies may unravel the important changes in, for example, nitric oxide (85) and CGRP (86) concentrations during the attack. The mechanism of action of antimigraine drugs could be explored, e.g. by recording of plasma CGRP concentrations during sumatriptan treatment of NTG-induced migraine attacks (87). Genetic factors could be studied, e.g. by comparing the response to the NTG test between patients with and without a strong family history of migraine. The test can be used to evaluate the efficacy of acute (79) and prophylactic migraine medications (80). Rare forms of migraine, e.g. FHM or basilar type migraine, could be investigated in multicentre studies, as the test methodology can be strictly standardized. Furthermore, since the test can be applied also in animals, it provides opportunities to bridge the gap between animal and human experimental migraine models. Thus, there are wide perspectives for using the test in migraine as well as in other primary headaches in the future.

Limitations of the method

The major limitation of the NTG migraine-provoking test is that, although the induced migraine has many similarities with a spontaneous attack, it is not a genuine attack. This implies that it is unknown whether the changes observed, e.g. in imaging or biochemical studies, would also be true for spontaneous attacks. It should also be considered whether changes observed in the laboratory are caused by the attack itself or by laboratory-induced stress.

The long period needed to observe the delayed headache puts a burden on the patient and makes studies time-consuming and expensive. Moreover, the likely outcome of the provocation, the migraine attack, is not exactly what the normal migraine patient is imagining. Together, these factors impede the inclusion of a large number of subjects in the studies.

Recommendations for improved methodology

It is of crucial importance that headaches are classified meticulously before as well as during the NTG test, because NTG has a differential effect in MoA and in MA, as described above, and because NTG induces cluster headache in cluster patients (71, 88), and TTH in patients with chronic TTH (89). The NTG test should be performed by a trained observer who is blinded to the diagnosis.

The limited number of available studies indicates that the sensitivity of the NTG test for provoking an attack of MoA in MoA patients is approximately 80% whether provocation is done with intravenous infusion of GTN 0.5 µg/kg per min over 20 min (43, 76) or with 0.9 mg NTG sublingually (71). Using lower doses significantly decreases the sensitivity of the test (80). Higher doses of GTN cannot be recommended, since dose–response studies have demonstrated a ceiling effect at 0.5 µg/kg per min (74). Given the comparable sensitivity of intravenous GTN and sublingual NTG, it is unlikely that increasing the sublingual NTG dose will be beneficial, but this has not been examined. End-points should be predefined and may include maximal headache intensity, area under the headache curve or sum of pain scores and the type of headache provoked. At least a 2-day interval between GTN challenges is recommended in healthy subjects and longer, e.g. 1 week, in migraineurs (66).

Acknowledgements

The leading authors D.M. and J.S. are grateful to Mrs P. Gerard for help with literature search and referencing. This work is supported by the EU STREP EUROHEAD (LSHM-CT-2004-5044837) (Workpackage 9).

References

1 Sandor

Roon

Ferrari

van Dijk

Schoenen

. Repeatability of the intensity dependence of corticalauditory evoked potentials in the assessment of cortical information processing. Cephalalgia 1999; 19:873–9.

Roon

Sandor

Schoonman

Lamers

Schoenen

Ferrari

Auditory evoked potentials in the assessment of central nervous system effects of antimigraine drugs. Cephalalgia 1999; 19:880–5.

3 Ogawa

Lee

Kay

Tank

. Brain magnetic resonance imaging with contrast dependent on blood oxygenation. Proc Natl Acad Sci USA 1990; 87:9868–72.

4 Cao

Welch

Aurora

Vikingstad

. Functional MRI-BOLD of visually triggered headache in patients with migraine. Arch Neurol 1999; 56:548–54.

5 Cao

Aurora

Nagesh

Patel

Welch

. Functional MRI-BOLD of brainstem structures during visually triggered migraine. Neurology 2002; 59:72–8.

Hadjikhani

Sanchez Del Rio

Schwartz

Bakker

Fischl

Mechanisms of migraine aura revealed by functional MRI in human visual cortex. Proc Natl Acad Sci USA 2001; 98:4687–92.

7 Vincent

Pedra

Mourao-Miranda

Bramati

Henrique

Moll

. Enhanced interictal responsiveness of the migraineous visual cortex to incongruent bar stimulation: a functional MRI visual activation study. Cephalalgia 2003; 23:860–8.

Rocca

Colombo

Pagani

Falini

Codella

Scotti

Evidence for cortical functional changes in patients with migraine and white matter abnormalities on conventional and diffusion tensor magnetic resonance imaging. Stroke 2003; 34:665–70.

9 Logothetis

Pauls

Augath

Trinath

Oeltermann

. Neurophysiological investigation of the basis of the fMRI signal. Nature 2001; 412 (6843):150–7.

10.

10 Welch

Levine

D'Andrea

Helpern

. Brain pH in migraine: an in vivo phosphorus-31 magnetic resonance spectroscopy study. Cephalalgia 1988; 8:273–7.

11.

Barbiroli

Montagna

Cortelli

Martinelli

Sacquegna

Zaniol

Complicated migraine studied by phosphorus magnetic resonance spectroscopy. Cephalalgia 1990; 10:263–72.

12.

Bresolin

Martinelli

Barbiroli

Zaniol

Ausenda

Montagna

Muscle mitochondrial DNA deletion and 31P-NMR spectroscopy alterations in a migraine patient. Neurol Sci 1991; 104:182–9.

13.

Barbiroli

Montagna

Cortelli

Funicello

Iotti

Monari

Abnormal brain and muscle energy metabolism shown by 31P magnetic resonance spectroscopy in patients affected by migraine with aura. Neurology 1992; 42:1209–14.

14.

Sacquegna

Lodi

De Carolis

Tinuper

Cortelli

Zaniol

Brain energy metabolism studied by 31P-MR spectroscopy in a case of migraine with prolonged aura. Acta Neurol Scand 1992; 86:376–80.

15.

Montagna

Cortelli

Monari

Pierangeli

Parchi

Lodi

31P-magnetic resonance spectroscopy in migraine without aura. Neurology 1994; 44:666–9.

16.

Uncini

Lodi

Di Muzio

Silvestri

Servidei

Lugaresi

Abnormal brain and muscle energy metabolism shown by 31P-MRS in familial hemiplegic migraine. J Neurol Sci 1995; 129:214–22.

17.

Lodi

Kemp

Montagna

Pierangeli

Cortelli

Iotti

Quantitative analysis of skeletal muscle bioenergetics and proton efflux in migraine and cluster headache. J Neurol Sci 1997; 146:73–80.

18.

Lodi

Iotti

Cortelli

Pierangeli

Cevoli

Clementi

Deficient energy metabolism is associated with low free magnesium in the brains of patients with migraine and cluster headache. Brain Res Bull 2001; 54:437–41.

19.

19 Boska

Welch

Barker

Nelson

Schultz

. Contrasts in cortical magnesium, phospholipid and energy metabolism between migraine syndromes. Neurology 2002; 58:1227–33.

20.

20 Watanabe

Kuwabara

Ohkubo

Tsuji

Yuasa

. Elevation of cerebral lactate detected by localized 1H-magnetic resonance spectroscopy in migraine during the interictal period. Neurology 1996; 47:1093–5.

21.

Sarchielli

Tarducci

Presciutti

Gobbi

Pelliccioli

Stipa

Functional 1H-MRS findings in migraine patients with and without aura assessed interictally. Neuroimage 2005; 24:1025–31.

22.

Sandor

Dydak

Schoenen

Kollias

Hess

Boesiger

MR-spectroscopic imaging during visual stimulation in subgroups of migraine with aura. Cephalalgia 2005; 25:507–18.

23.

23 Dichgans

Herzog

Freilinger

Wilke

Auer

. 1H-MRS alterations in the cerebellum of patients with familial hemiplegic migraine type 1. Neurology 2005; 64:608–13.

24.

24 Fisher

Prichard

Warach

. New magnetic resonance techniques for acute ischemic stroke. JAMA 1995; 274:908–11.

25.

Sakuma

Nomura

Takeda

Tagami

Nakagawa

Tamagawa

Adult and neonatal human brain: diffusional anisotropy and myelination with diffusion-weighted MR imaging. Radiology 1991; 180:229–33.

26.

Cutrer

Sorensen

Weisskoff

Ostergaard

Sanchez Del Rio

Lee

Perfusion-weighted imaging defects during spontaneous migrainous aura. Ann Neurol 1998; 43:25–31.

27.

Gutschalk

Kollmar

Mohr

Henze

Ille

Schwaninger

Multimodal functional imaging of prolonged neurological deficits in a patient suffering from familial hemiplegic migraine. Neurosci Lett 2002; 332:115–8.

28.

28 Smith

Cros

Sheen

. Hyperfusion with vasogenic leakage by fMRI in migraine with prolonged aura.Neurology 2002; 58:1309–10.

29.

29 Rocca

Colombo

Inglese

Codella

Comi

Filippi

. A diffusion tensor magnetic resonance imaging study of brain tissue from patients with migraine.J Neurol Neurosurg Psychiatry 2003; 74:501–3.

30.

30 Relja

Granato

Ukmar

Ferretti

Antonello

Zorzon

. Persistent aura without infraction; description of the first case studied with both brain SPECT and perfusion MRI. Cephalalgia 2005; 25:56–9.

31.

Sanchez Del Rio

Bakker

Agosti

Mitsikostas

Ostergaard

Perfusion weighted imaging during migraine: spontaneous visual aura and headache. Cephalalgia 1999; 19:701–7.

32.

32 Woods

RPW

Iacoboni

Mazziota

. Bilateral spreading cerebral hypoperfusion during spontaneous migraine headache. N Engl J Med 1994; 331:1689–92.

33.

33 Bednarczyk

Remler

Weikart

Nelson

Reed

. Global cerebral blood flow, blood volume, and oxygen metabolism in patients with migraine headache. Neurology 1998; 50:1736–40.

34.

34 Denuelle

Fabre

Payoux

Chollet

Geraud

. Brainstem and hypthalamic activation in spontaneous migraine attacks. Cephalalgia 2004; 24:775–814.

35.

35 Lauritzen

. Pathophysiology of the migraine aura. The spreading depression theory. Brain 1994; 117:199–210.

36.

36 Goadsby

Lambert

Lance

. Differential effects on the internal and external carotid circulation of the monkey evoked by locus coerulus stimulation. Brain Res 1982; 249:247–54.

37.

37 Goadsby

Duckworth

. Low frequency stimulation of the locus coeruleus reduces regional cerebral blood flow in the spinalized cat. Brain Res 1989; 476:71–7.

38.

Weiller

May

Limmroth

Juptner

Kaube

Schayck

Brain stem activation in spontaneous human migraine attacks. Nat Med 1995; 1:658–60.

39.

Afridi

Giffin

Kaube

Friston

Ward

Frackowiak

A positron emission tomographic study in spontaneous migraine. Arch Neurol 2005; 62:1270–5.

40.

40 Headache Classification Subcommittee of the International Headache Society. The International Classification of Headache Disorders, 2nd edition. Cephalalgia 2004; 24 (Suppl. 1).

41.

41 Andersson

Muhr

Lilja

Lundberg

Langstrom

. Regional cerebral blood flow and oxygen metabolism during migraine with and without aura. Cephalalgia 1997; 17:570–9.

42.

42 Bahra

Matharu

Buchel

Frackowiak

Goadsby

. Brainstem activation specific to migraine headache. Lancet 2001; 357 (9261):1016–7.

43.

Afridi

Matharu

Lee

Kaube

Friston

Frackowiak

A PET study exploring the laterality of brainstem activation in migraine using glyceryl trinitrate. Brain 2005; 128:932–9.

44.

44 Matharu

Bartsch

Ward

Frackowiak

Weiner

Goadsby

. Central neuromodulation in chronic migraine patients with suboccipital stimulators. Brain 2004; 127:220–30.

45.

Vargha-Khadem

Watkins

Price

Ashburner

Alcock

Connelly

Neural basis of an inherited speech and language disorder. Proc Natl Acad Sci USA 1998; 95:12695–700.

46.

46 Woermann

Free

Koepp

Sisodiya

Duncan

. Abnormal cerebral structure in juvenile myoclonic epilepsy demonstrated with voxel-based analysis of MRI. Brain 1999; 122:2101–8.

47.

May

Ashburner

Buchel

McGonigle

Friston

Frackowiak

Correlation between structural and functional changes in brain in an idiopathic headache syndrome. Nat Med 1999; 5:836–8.

48.

48 Ashburner

Friston

. Voxel-based morphometry—the methods. Neuroimage 2000; 11:805–21.

49.

49 May

. The role of imaging in the pathophysiology and diagnosis in headache. Curr Opin Neurol 2005; 18:293–7.

50.

Apkarian

Sosa

Sonty

Levy

Harden

Parrish

Chronic back pain is associated with decreased prefrontal and thalamic gray matter density.J Neurosci 2004; 24:10410–5.

51.

Schmidt-Wilcke

Leinisch

Ganssbauer

Draganski

Bogdahn

Altmeppen

Affective components and intensity of pain correlate with structural differences in gray matter in chronic back pain patients. Pain 2006; 125:89–97.

52.

Draganski

Moser

Lummel

Ganssbauer

Bogdahn

Haas

Decrease of thalamic gray matter following limb amputation. Neuroimage 2006; 31:951–7.

53.

53 Matharu

Good

May

Bahra

Goadsby

. No change in the structure of the brain in migraine: a voxel-based morphometric study. Eur J Neurol 2003; 10:53–7.

54.

Schmidt-Wilcke

Leinisch

Straube

Kampfe

Draganski

Diener

Gray matter decrease in patients with chronic tension type headache. Neurology 2005; 65:1483–6.

55.

Rocca

Ceccarelli

Falini

Colombo

Tortorella

Bernasconi

Brain gray matter changes in migraine patients with T2-visible lesions: a 3-T MRI study. Stroke 2006; 37:1765–70.

56.

Maguire

Gadian

Johnsrude

Good

Ashburner

Frackowiak

Navigation-related structural change in the hippocampi of taxi drivers. Proc Natl Acad Sci USA 2000; 97:4398–403.

57.

57 Sluming

Barrick

Howard

Cezayirli

Mayes

Roberts

. Voxel-based morphometry reveals increased gray matter density in Broca's area in male symphony orchestra musicians. Neuroimage 2002; 17:1613–22.

58.

58 Golestani

Paus

Zatorre

. Anatomical correlates of learning novel speech sounds. Neuron 2002; 35:997–1010.

59.

59 Gaser

Schlaug

. Brain structures differ between musicians and non-musicians. J Neurosci 2003; 23:9240–5.

60.

60 Draganski

Gaser

Busch

Schuierer

Bogdahn

May

. Neuroplasticity: changes in grey matter induced by training. Nature 2004; 427 (6972):311–2.

61.

61 Buchel

Raedler

Sommer

Sach

Weiller

Koch

. White matter asymmetry in the human brain: a diffusion tensor MRI study. Cereb Cortex 2004; 14:945–51.

62.

62 Mortimer

Good

Marsters

. The VEP in acephalgic migraine. Headache 1990; 30:285–8.

63.

63 May

Gaser

. Magnetic resonance-based morphometry: a window into structural plasticity of the brain. Curr Opin Neurol 2006; 19:407–11.

64.

64 Friston

Holmes

Ashburner

Poline

. SPM99; http://www.fil.icn.ucl.ac.uk/spm, 1999.

65.

Etgen

Draganski

Ilg

Schroeder

Geisler

Hajak

Bilateral thalamic gray matter changes in patients with restless legs syndrome. Neuroimage 2004; 24:1242–7.

66.

66 Iversen

. Glyceryl trinitrate-induced headache. An experimental model of headache in humans. Own Press, Copenhagen; 2004:1–75.

67.

67 Lassen

Haderslev

Jacobsen

Iversen

Sperling

Olesen

. CGRP may play a causative role in migraine. Cephalalgia 2002; 22:54–61.

68.

68 Kruuse

Thomsen

Birk

Olesen

. Migraine can be induced by sildenafil without changes in middle cerebral artery diameter. Brain 2003; 126:241–7.

69.

69 Thomsen

. Investigations into the role of nitric oxide and the large intracranial arteries in migraine headache. Cephalalgia 1997; 17:873–95.

70.

Juhasz

Zsombok

Modos

Olajos

Jakab

Nemeth

NO-induced migraine attack: strong increase in plasma calcitonin gene-related peptide (CGRP) concentration and negative correlation with platelet serotonin release. Pain 2003; 106:461–70.

71.

71 Sances

Tassorelli

Pucci

Ghiotto

Sandrini

Nappi

. Reliability of the nitroglycerin provocative test in the diagnosis of neurovascular headaches. Cephalalgia 2004; 24:110–9.

72.

de Tommaso

Guido

Libro

Losito

Difruscolo

Puca

Topographic and dipolar analysis of laser-evoked potentials during migraine attack. Headache 2004; 44:947–60.

73.

Bednarczyk

Wack

Kassab

Burch

Trinidad

Haka

Brain blood flow in the nitroglycerin (GTN) model of migraine: measurement using positron emission tomography and transcranial Doppler. Cephalalgia 2002; 22:749–57.

74.

74 Iversen

Olesen

Tfelt-Hansen

. Intravenous nitroglycerin as an experimental model of vascular headache. Pain 1989; 38:17–24.

75.

75 Olesen

Iversen

Thomsen

. Nitric oxide supersensitivity: a possible molecular mechanism of migraine pain. Eur J Neurol 1994; 1:73–80.

76.

76 Thomsen

Kruuse

Iversen

Olesen

. A nitric oxide donor (nitroglycerin) triggers genuine migraine attacks. Eur J Neurol 1994; 1:73–80.

77.

77 Christiansen

Thomsen

Daugaard

Ulrich

Olesen

. Glyceryl trinitrate induces attacks of migraine without aura in sufferers of migraine with aura. Cephalalgia 1999; 19:660–7; discussion 626.

78.

78 Afridi

Kaube

Goadsby

. Glyceryl trinitrate triggers premonitory symptoms in migraineurs. Pain 2004; 110:675–80.

79.

79 Iversen

Olesen

. Headache induced by a nitric oxide donor (nitroglycerin) responds to sumatriptan, a human model for development of migraine drugs. Cephalalgia 1996; 16:412–8.

80.

80 Tvedskov

Thomsen

Iversen

Gibson

Williams

Olesen

. The prophylactic effect of valproate on glyceryltrinitrate induced migraine. Cephalalgia 2004; 24:576–85.

81.

81 Lassen

Ashina

Christiansen

Ulrich

Olesen

. Nitric oxide synthase inhibition in migraine. Lancet 1997; 349 (9049):401–2.

82.

82 Kaube

Katsarava

Przywara

Drepper

Ellrich

Diener

. Acute migraine headache: possible sensitization of neurons in the spinal trigeminal nucleus? Neurology 2002; 58:1234–8.

83.

83 Ambrosini

Rossi

De Pasqua

Pierelli

Schoenen

. Lack of habituation causes high intensity dependence of auditory evoked cortical potentials in migraine. Brain 2003; 126:2009–15.

84.

Bohotin

Fumal

Vandenheede

Gérard

Bohotin

Maertens de Noordhout

Effects of repetitive transcranial magnetic stimulation on visual evoked potentials in migraine. Brain 2002; 125:912–22.

85.

85 Ashina

Simonsen

Bendtsen

Jensen

Olesen

. Glyceryl trinitrate may trigger endogenous nitric oxide production in patients with chronic tension-type headache. Cephalalgia 2004; 24:967–72.

86.

86 Ashina

Bendtsen

Jensen

Schifter

Olesen

. Calcitonin gene-related peptide levels during nitric oxide-induced headache in patients with chronic tension-type headache. Eur J Neurol 2001; 8:173–8.

87.

87 Juhasz

Zsombok

Jakab

Nemeth

Szolcsanyi

Bagdy

. Sumatriptan causes parallel decrease in plasma calcitonin gene-related peptide (CGRP) concentration and migraine headache during nitroglycerin induced migraine attack. Cephalalgia 2005; 25:179–83.

88.

88 Ekbom

. Nitroglycerin as a provocative agent in cluster headache. Arch Neurol 1968; 19:487–93.

89.

89 Ashina

Bendtsen

Jensen

Olesen

. Nitric oxide-induced headache in patients with chronic tension-type headache. Brain 2000; 123:1830–7.

Evaluation and Proposal for Optimization of Neurophysiological Tests In Migraine: Part 2—Neuroimaging and The Nitroglycerin Test

Abstract

Keywords

Functional magnetic resonance

Generalities on optimal method, personal experience

BOLD-fMRI

Definition

Literature data

Limitations

Magnetic resonance spectroscopy

Definition

Literature data

Limitations

Diffusion-weighted MRI

Definition

Literature data

Limitations

Perfusion-weighted MRI

Definition

Literature data—method

Limitations

Positron emission tomography

Definition

Literature data

Spontaneous migraine

Triggered migraine

Chronic migraine

Limitations

Voxel-based morphometry

Definition

Literature data

Advantage of the method

Limitations of the method

Recommendations for improved methodology

Which optimal methodology can be proposed in headache studies?

Studies needed

Migraine-provoking test

Definition

Literature data

Advantages of the method and perspectives

Limitations of the method

Recommendations for improved methodology

Acknowledgements

References