Abstract

Our aim was identify brain areas involved in the premonitory phase of migraine using functional neuroimaging. To this end, we performed positron emission tomography scans with H215O to measure cerebral blood flow as a marker of neuronal activity. We conducted positron emission tomography scans at baseline, in the premonitory phase without pain and during migraine headache in eight patients. We used glyceryl trinitrate (nitroglycerin) to trigger premonitory symptoms and migraine headache in patients with episodic migraine without aura who habitually experienced premonitory symptoms during spontaneous attacks. The main outcome was comparing the first premonitory scans in all patients to baseline scans in all patients. We found activations in the posterolateral hypothalamus, midbrain tegmental area, periaqueductal grey, dorsal pons and various cortical areas including occipital, temporal and prefrontal cortex. Brain activations, in particular of the hypothalamus, seen in the premonitory phase of glyceryl trinitrate-triggered migraine attacks can explain many of the premonitory symptoms and may provide some insight into why migraine is commonly activated by a change in homeostasis.

Introduction

Migraine is among the most disabling neurological problems worldwide (Murray et al., 2012) and highly costly (Stewart et al., 2003; Olesen et al., 2012). Functional neuroimaging studies investigating the headache phase of spontaneous (Weiller et al., 1995; Afridi et al., 2005a) and nitroglycerin-induced (Bahra et al., 2001; Afridi et al., 2005b) migraine have consistently shown dorsal pontine activation, which persists after acute migraine treatment with sumatriptan. There have been no such studies in the premonitory phase that represents the earliest clinical manifestations of the attack and thus offers unique insights into migraine pathophysiology.

Many migraine patients experience premonitory symptoms that begin up to 3 days before headache and can be highly predictive of a migraine headache (Giffin et al., 2003). In a prospective electronic diary study, for 82% of selected patients, premonitory symptoms were followed by a migraine headache within 72 h more than 50% of the time (Giffin et al., 2003). Tiredness, neck stiffness and difficulty in concentration were the most common symptoms. Yawning, emotional changes and difficulties in reading and writing were the most predictive for a migraine headache. Electrophysiological studies also, have consistently shown lack of habituation of cortical responses to various stimuli that increases progressively in the period before headache, and normalizes with headache, consistent with the notion that the premonitory phase is likely to entail events key to migraine generation (Schoenen et al., 2003).

Remarkably as well as triggering migraine attacks phenotypically indistinct from other attacks, nitroglycerin can trigger the premonitory symptoms again with all the characteristics of a patient’s routine symptoms (Afridi et al., 2004). Triggering of attacks was first well characterized by Iversen et al. (1989), and has formed the basis of novel therapeutic approaches (Lassen et al., 1997), and pathophysiological understanding of the disorder (Olesen et al., 1995). Importantly, aura symptoms are not commonly provoked by nitroglycerin and patients with migraine with aura get migraine headache without aura on exposure to nitroglycerin (Christiansen et al., 1999; Afridi et al., 2004). Here, we set out to investigate the premonitory phase of nitroglycerin-induced migraine using PET scanning.

Patients and methods

We conducted telephone interviews, after advertisements in local media, to select patients who met the inclusion criteria: age 18–65 years, migraine without aura (Headache Classification Committee of The International Headache Society, 2004), <15 days of headache a month, premonitory symptoms before headache, no major medical conditions and not on preventive drugs for migraine or any other regular medications that could confound the study. We excluded patients with migraine aura to prevent confusion with premonitory symptoms, as both usually occur before headache. The study was approved by the UCSF Committee on Human Research and the Radiation Safety Committee. Written informed consent was obtained from all patients before study inclusion according to the Declaration of Helsinki.

Screening of patients

We initially screened 142 patients of which 25 satisfied the inclusion criteria and agreed to participate. The most common reasons for non-inclusion were: headache >15 days a month (n = 41), migraine aura (n = 12), confounding medications (n = 25), patients unwilling due to radiation exposure (n = 8) and miscellaneous (n = 31; Supplementary Table 1).

Infusion of nitroglycerin

The 25 patients fulfilling the inclusion criteria were invited for the first visit during which we infused intravenous nitroglycerin 0.5 µg/kg/min over 20 min to select patients who responded with premonitory symptoms before migraine headache. Patients were assessed before the nitroglycerin infusion and then at periodic intervals, initially every 5 min and then less frequently, every 10–15 min. Patients were asked about headache side, site, type and intensity, nausea, photophobia, phonophobia, movement sensitivity (not during scanning session), tiredness, neck stiffness, yawning, mood changes, micturition, thirst and cravings (Supplementary Table 2). Patients were assessed for all these parameters before the infusion and any significant change from baseline was considered positive.

Premonitory phase

Clinical observations

The premonitory phase was defined as the period following when the nitroglycerin-induced non-specific headache phase had completely ceased and patients started to experience symptoms warning them of an impending headache. These symptoms had to be present on at least two enquiries. Symptoms were recorded as present or absent without grading. Blood pressure, oxygen saturation and pulse rate were recorded initially every 5 min and then less frequently. Of these 25 patients, 16 had delayed migraine headache. Of these 16, 11 patients had premonitory symptoms before migraine headache. These 11 patients were invited for PET scans at least 7 days after the first nitroglycerin infusion. All patients had been pain-free for at least 72 h before the PET scans. The procedure for nitroglycerin infusion and recording of symptoms was repeated as on the first occasion.

Functional imaging data collection

We performed PET scans with the GE Discovery VCT PET/CT system in 3D mode with septa retracted. All patients were instructed to keep their eyes closed during the scans. The patients were positioned in the PET scanner, and their head immobilized with standard immobilization straps and a low dose CT scan performed for attenuation correction. CT scans for attenuation correction were repeated when patients exited the scanner for relaxation between conditions with a maximum three exits per subject. An antecubital vein cannula was used to administer the tracer, ∼370 MBq of H215O, which was repeated before each scan. The activity was infused into patients over 20 s at a rate of 10 ml/min. The interval between scans was ∼10 min allowing an interscan interval of five half-lives of H215O (t1/2 = 122 s). The PET data were acquired dynamically and summed for one 90-s frame beginning 5 s before the peak of the head curve.

Functional imaging conditions

Each patient had scans in three conditions: baseline, premonitory phase and migraine headache. We could not randomize the order of the scans since the premonitory and migraine headache phases were triggered by the nitroglycerin infusion sequentially. We planned to do three scans in each condition. The number of scans in the premonitory phase depended on how soon the migraine headache developed. Soon after the initiation of nitroglycerin, patients had a mild headache that lasted for 31 ± 12 min [mean ± (SD)]. Premonitory scans were performed after the nitroglycerin headache had completely subsided, premonitory symptoms were present and the migraine headache had not appeared. The time for the first premonitory scan after start of nitroglycerin infusion was 67 ± 34 min (range 40–142; see Fig. 1 for time line for scans and Table 1). Migraine scans were performed when the delayed headache was moderate or severe. The mean time for the first migraine headache scan after start of nitroglycerin infusion was 132 ± 69 min (range 82–270 min; Fig. 1 and Table 1). After the first scan during the premonitory and migraine headache phases, the subsequent scans were done at ∼10-min intervals.

Figure 1

Time line for scans after start of nitroglycerin infusion. After the first scan, the subsequent scans were done at ∼10-min intervals during the baseline, premonitory and migraine headache phases.

Figure 1

Time line for scans after start of nitroglycerin infusion. After the first scan, the subsequent scans were done at ∼10-min intervals during the baseline, premonitory and migraine headache phases.

Table 1

Number of scans in each phase and time after nitroglycerin, and details of premonitory symptoms and migraine headache

Patient ID Total number of premonitory scans Time for first premonitory scan after start of nitroglycerin infusion (min) Premonitory symptoms during scans Total number of migraine headache scans Maximum pain score during migraine headache (1–10 VAS) Time for first migraine scan after start of nitroglycerin infusion (min) 
82 Tiredness 143 
Neck stiffness 
45 Tiredness 270 
Neck stiffness 
Urination 
Thirst 
44 Nausea 54 
Photophobia 
Thirst 
62 Tiredness 159 
Nausea 
Photophobia 
142 Moody 10 165 
Tiredness 
Neck stiffness 
43 Urination 82 
Thirst 
76 Yawning 88 
Urination 
Thirst 
40 Tiredness 95 
Patient ID Total number of premonitory scans Time for first premonitory scan after start of nitroglycerin infusion (min) Premonitory symptoms during scans Total number of migraine headache scans Maximum pain score during migraine headache (1–10 VAS) Time for first migraine scan after start of nitroglycerin infusion (min) 
82 Tiredness 143 
Neck stiffness 
45 Tiredness 270 
Neck stiffness 
Urination 
Thirst 
44 Nausea 54 
Photophobia 
Thirst 
62 Tiredness 159 
Nausea 
Photophobia 
142 Moody 10 165 
Tiredness 
Neck stiffness 
43 Urination 82 
Thirst 
76 Yawning 88 
Urination 
Thirst 
40 Tiredness 95 

Using the above scans, the following contrasts were generated:

1. Early premonitory (main outcome). First premonitory scan of all patients (total 8) > all baseline scans of all patients (total 24), n = 8.

2. Late premonitory. Second premonitory scan of all patients (total 5) > all baseline scans of all patients (total 15), n = 5.

3. All premonitory scans of all patients (total 16) > all baseline scans of all patients (total 24), n = 8.

4. All migraine scans of all patients (total 26) > all baseline scans of all patients (total 24), n = 8.

After the first scan in each phase i.e. baseline, premonitory and migraine headache, subsequent scans were done at ∼10-min intervals.

VAS = visual–analogue scale.

Functional imaging scan distribution

Of 11 patients, eight could have at least one scan during the premonitory period when no pain was present. In the remaining three patients, either the nitroglycerin headache was followed by migraine headache without a pain-free interval or the premonitory stage was too short to allow the scans. Therefore, only these eight patients’ scans were used for the final analysis. All the eight patients had three scans each during baseline. In the premonitory phase, three patients had two scans each, three patients had one scan each; one patient had three and one patient had four scans. In the migraine headache phase, five patients had three scans each, one patient had two, one patient had four and one patient had five scans.

Data preprocessing and statistical analysis

Images were reconstructed by 3D iterative reconstruction into 47 image planes (separation 3.27 mm) and into a 128 × 128 image matrix (pixel size 2.1 × 2.1 mm2). SPM2 (www.fil.ion.ucl.ac.uk/spm) was used for data preprocessing and statistical analysis (Frackowiak and Friston, 1994) and paired t-tests were used to determine differences between conditions. Images were realigned with the first as reference and spatially normalized into MNI space. The normalized images were smoothed with a Gaussian filter of 8 mm full-width at half-maximum. Statistical parametric maps were derived with pre-specified contrasts as stated in the main text, comparing regional cerebral blood flow during states of interest. Correction for multiple comparisons was performed with false discovery rate (FDR) and P < 0.05 was considered significant. As all our patients had either right-sided or predominantly right-sided headache, and none of them had purely or predominantly left-sided headache, the scans were not flipped.

To capture the temporal sequence of activation, we applied the following voxel wise contrasts: (i) first premonitory scans (early premonitory) of all patients against baseline scans of all patients; (ii) second premonitory scans (late premonitory) of all patients against baseline scans of all patients; (iii) all premonitory scans of all patients against baseline scans of all patients; and (iv) all migraine scans of all patients against baseline scans of all patients. The main outcome was activation pattern in the early premonitory scans which represented the earliest premonitory symptoms after exposure to trigger (nitroglycerin).

Results

Eight patients (five females; mean age 30 years, range 19–47 years, all right-handed) were included in the final analysis (Table 2). During the scanning session, tiredness, neck stiffness and thirst were the three most common symptoms in the premonitory phase followed by frequent urination, photophobia, nausea, yawning and mood changes (Table 1). Incidentally, all patients had either right-sided or bilateral, but predominantly right-sided migraine headache although this was not an inclusion criterion.

Table 2

Basic characteristics of habitual migraine attacks

Patient ID Gender Headache days/month Age (years) Duration of disease (years) Acute medications Past preventives Other medications Family history 
39 23 Sumatriptan/aspirin/paracetamol None None Yes 
19 Ibuprofen, paracetamol None None Not sure 
47 33 Sumatriptan, ibuprofen None None Not sure 
21 None None None Yes 
35 23 Aspirin/paracetamol None None Yes 
30 25 Ibuprofen, paracetamol None None Yes 
10 26 Aspirin/paracetamol CoQ10 OCP Yes 
25 None None OCP Yes 
Patient ID Gender Headache days/month Age (years) Duration of disease (years) Acute medications Past preventives Other medications Family history 
39 23 Sumatriptan/aspirin/paracetamol None None Yes 
19 Ibuprofen, paracetamol None None Not sure 
47 33 Sumatriptan, ibuprofen None None Not sure 
21 None None None Yes 
35 23 Aspirin/paracetamol None None Yes 
30 25 Ibuprofen, paracetamol None None Yes 
10 26 Aspirin/paracetamol CoQ10 OCP Yes 
25 None None OCP Yes 

Mean ± SD for headache days/month; all fulfilling migraine headache = 4.6 ± 3.0; 95% confidence interval 2.0–7.1

OCP = oral contraceptive pill.

Brain areas activated

Early premonitory

Brain areas activated in the early premonitory phase compared to baseline are listed in Table 3 and shown in Fig. 2. Areas of note include the posterolateral hypothalamus, midbrain tegmental area and substantia nigra, periaqueductal grey, dorsal pons and various cortical areas including occipital, temporal and prefrontal cortex.

Figure 2

Brain activations in the early premonitory phase. Shown are areas of increased regional cerebral blood flow in the early premonitory phase greater than baseline in nitroglycerin-induced migraine. The results are superimposed on an anatomical reference derived from a representative T1-weighted MRI of one of the patients. Activations in the posterior hypothalamic region (A and B), the periaqueductal grey region (C and D), and in dorsal pons (E and F) are highlighted by circles. The colour bar indicates the colour coding of the z-scores. Images are displayed in radiological convention i.e. left side of image is right side of brain.

Figure 2

Brain activations in the early premonitory phase. Shown are areas of increased regional cerebral blood flow in the early premonitory phase greater than baseline in nitroglycerin-induced migraine. The results are superimposed on an anatomical reference derived from a representative T1-weighted MRI of one of the patients. Activations in the posterior hypothalamic region (A and B), the periaqueductal grey region (C and D), and in dorsal pons (E and F) are highlighted by circles. The colour bar indicates the colour coding of the z-scores. Images are displayed in radiological convention i.e. left side of image is right side of brain.

Table 3

MNI co-ordinates of areas with increased regional cerebral blood flow in the early premonitory phase > baseline

Brain regions MNI co-ordinates
 
Z-score P-value FDR corrected, significant < 0.05 
 x y z   
Right posterior hypothalamic region −6 −12 3.77 0.004 
Right ventral tegmental area/substantia nigra −9 −12 4.21 0.001 
Right periaqueductal grey −24 −9 4.65 0.000 
Right dorsal pons −36 −27 2.81 0.02 
Right putamen 24 −12 4.32 <0.001 
Left putamen −27 −3 3.02 0.01 
Right caudate nucleus 15 −6 3.08 0.01 
Right occipital cortex/BA 18 −102 6.08 <0.001 
Left occipital cortex/BA 18 −30 −102 6.60 <0.001 
Right prefrontal cortex/BA 10 36 60 12 4.47 <0.001 
Right frontal cortex/BA 6 27 21 63 4.55 <0.001 
Right temporal cortex/BA 20 54 −48 −15 4.98 <0.001 
Left parietal cortex/precuneus/BA 7 −6 −39 48 4.04 0.001 
Right cerebellum 48 −75 −36 5.85 <0.001 
Left cerebellum −36 −66 −33 4.53 <0.001 
Right anterior cingulate/BA 32 12 36 12 3.14 0.01 
Right posterior cingulate/BA 23 −36 33 4.62 <0.001 
Right pulvinar nucleus of thalamus −33 3.52 0.03 
Brain regions MNI co-ordinates
 
Z-score P-value FDR corrected, significant < 0.05 
 x y z   
Right posterior hypothalamic region −6 −12 3.77 0.004 
Right ventral tegmental area/substantia nigra −9 −12 4.21 0.001 
Right periaqueductal grey −24 −9 4.65 0.000 
Right dorsal pons −36 −27 2.81 0.02 
Right putamen 24 −12 4.32 <0.001 
Left putamen −27 −3 3.02 0.01 
Right caudate nucleus 15 −6 3.08 0.01 
Right occipital cortex/BA 18 −102 6.08 <0.001 
Left occipital cortex/BA 18 −30 −102 6.60 <0.001 
Right prefrontal cortex/BA 10 36 60 12 4.47 <0.001 
Right frontal cortex/BA 6 27 21 63 4.55 <0.001 
Right temporal cortex/BA 20 54 −48 −15 4.98 <0.001 
Left parietal cortex/precuneus/BA 7 −6 −39 48 4.04 0.001 
Right cerebellum 48 −75 −36 5.85 <0.001 
Left cerebellum −36 −66 −33 4.53 <0.001 
Right anterior cingulate/BA 32 12 36 12 3.14 0.01 
Right posterior cingulate/BA 23 −36 33 4.62 <0.001 
Right pulvinar nucleus of thalamus −33 3.52 0.03 

The co-ordinates represent the maxima within a cluster. Stated P-values are FDR corrected.

BA = Brodmann area; MNI = Montreal Neurological Institute

Late premonitory

Brain areas activated in the late premonitory phase compared to baseline are listed in Table 4 (Fig. 3). Please note persistence of dorsal pontine activation but not of the hypothalamic and midbrain activations. Additionally, activations were seen in right pulvinar and left thalamic areas.

Figure 3

Brain activations in late premonitory, all premonitory (early and late) and migraine headache phases in nitroglycerin-induced migraine. Areas of increased regional cerebral blood flow in late premonitory, all premonitory (early and late) and migraine headache greater than baseline are shown. The results are superimposed on an anatomical reference derived from a T1- weighted MRI of one of the patients. The colour bar indicates colour coding of the z-scores. Results are thresholded at P < 0.005 uncorrected for display purposes.

Figure 3

Brain activations in late premonitory, all premonitory (early and late) and migraine headache phases in nitroglycerin-induced migraine. Areas of increased regional cerebral blood flow in late premonitory, all premonitory (early and late) and migraine headache greater than baseline are shown. The results are superimposed on an anatomical reference derived from a T1- weighted MRI of one of the patients. The colour bar indicates colour coding of the z-scores. Results are thresholded at P < 0.005 uncorrected for display purposes.

Table 4

MNI co-ordinates of areas with increased regional cerebral blood flow in the late premonitory phase > baseline

Brain regions MNI co-ordinates
 
Z-score P-value, FDR corrected, significant < 0.05 
 x y z   
Right pulvinar −24 3.55 0.01 
Right dorsal pons −36 −30 4.21 0.01 
Left thalamus −6 −12 3.40 0.03 
Left occipital cortex/BA 18 −18 −78 3.92 0.01 
Right anterior cingulate cortex/BA 32 42 3.94 0.01 
Brain regions MNI co-ordinates
 
Z-score P-value, FDR corrected, significant < 0.05 
 x y z   
Right pulvinar −24 3.55 0.01 
Right dorsal pons −36 −30 4.21 0.01 
Left thalamus −6 −12 3.40 0.03 
Left occipital cortex/BA 18 −18 −78 3.92 0.01 
Right anterior cingulate cortex/BA 32 42 3.94 0.01 

The co-ordinates represent the maxima within a cluster.

BA = Brodmann area; MNI = Montreal Neurological Institute

All premonitory

Brain areas activated in the premonitory phase (early and late together) compared to baseline are listed in Table 5 (Fig. 3).

Table 5

MNI co-ordinates of areas with increased regional cerebral blood flow for all premonitory phase scans > baseline

Brain regions MNI co-ordinates
 
Z-score P-value, FDR corrected, significant < 0.05 
 x y z   
Right dorsal pons −33 −27 3.52 0.02 
Right ventral tegmental area/substantia nigra −18 −21 3.25 0.03 
Right occipital cortex/BA 17 −102 5.05 0.001 
Left occipital cortex/BA 18 −12 −102 12 3.55 0.02 
Right temporal cortex/BA 20 63 −30 −24 4.39 0.004 
Right cuneus/BA 19 −84 42 3.96 0.01 
Right precuneus/BA 7 −57 36 4.25 0.006 
Right anterior cingulate cortex/BA 32 42 3.45 0.02 
Right middle frontal gyrus/BA10 39 54 4.14 0.007 
Left middle frontal gyrus/BA10 −39 54 −3 3.84 0.01 
Left superior parietal lobule/BA 7 39 −60 54 4.52 0.003 
Left cerebellum −48 −72 −39 4.01 0.009 
Right cerebellum 33 −78 −48 3.80 0.14 
Brain regions MNI co-ordinates
 
Z-score P-value, FDR corrected, significant < 0.05 
 x y z   
Right dorsal pons −33 −27 3.52 0.02 
Right ventral tegmental area/substantia nigra −18 −21 3.25 0.03 
Right occipital cortex/BA 17 −102 5.05 0.001 
Left occipital cortex/BA 18 −12 −102 12 3.55 0.02 
Right temporal cortex/BA 20 63 −30 −24 4.39 0.004 
Right cuneus/BA 19 −84 42 3.96 0.01 
Right precuneus/BA 7 −57 36 4.25 0.006 
Right anterior cingulate cortex/BA 32 42 3.45 0.02 
Right middle frontal gyrus/BA10 39 54 4.14 0.007 
Left middle frontal gyrus/BA10 −39 54 −3 3.84 0.01 
Left superior parietal lobule/BA 7 39 −60 54 4.52 0.003 
Left cerebellum −48 −72 −39 4.01 0.009 
Right cerebellum 33 −78 −48 3.80 0.14 

The co-ordinates represent the maxima within a cluster.

BA = Brodmann area; MNI = Montreal Neurological Institute

Migraine headache

Brain areas activated in the migraine headache phase compared to baseline are listed in Table 6 (Fig. 3). It is notable that while dorsal pontine activation continues the postero-lateral hypothalamus and periaqueductal grey are not significantly activated.

Table 6

MNI co-ordinates of areas with increased regional cerebral blood flow in the migraine headache phase > baseline

Brain regions MNI co-ordinates
 
Z-score P-value FDR corrected, significant < 0.05 
 x y z   
Right dorsal pons −36 −24 3.41 0.04* 
Right claustrum 33 4.04 0.03 
Right insula/BA 13 48 −24 21 4.09 0.02 
Right Precentral gyrus/BA 6 48 44 3.75 0.04 
Right prefrontal cortex/BA 10 39 42 15 3.83 0.03 
Left post central gyrus/BA 3 −51 −15 51 3.91 0.03 
Right parietal lobe/BA 7 12 −66 60 3.81 0.04 
Right cerebellum 27 −57 −51 3.82 0.04 
Left cerebellum −30 −51 −42 4.58 0.01 
Brain regions MNI co-ordinates
 
Z-score P-value FDR corrected, significant < 0.05 
 x y z   
Right dorsal pons −36 −24 3.41 0.04* 
Right claustrum 33 4.04 0.03 
Right insula/BA 13 48 −24 21 4.09 0.02 
Right Precentral gyrus/BA 6 48 44 3.75 0.04 
Right prefrontal cortex/BA 10 39 42 15 3.83 0.03 
Left post central gyrus/BA 3 −51 −15 51 3.91 0.03 
Right parietal lobe/BA 7 12 −66 60 3.81 0.04 
Right cerebellum 27 −57 −51 3.82 0.04 
Left cerebellum −30 −51 −42 4.58 0.01 

The co-ordinates represent the maxima within a cluster.

*After small volume correction with a sphere of radius 12 mm on co-ordinates x = 3, y = −36, z = −24.

Uncorrected P-values < 0.001 for all co-ordinates.

BA = Brodmann area; MNI = Montreal Neurological Institute

Table 7 shows the comparison of the activations in early and premonitory phases and migraine headache. Table 1 summarizes the total number of scans for each patient in each phase and how the scans were used to generate the contrasts.

Table 7

Comparison of brain activations in early premonitory, late premonitory and migraine headache phases

Brain area Early premonitory Late premonitory Migraine headache 
Hypothalamus Yes No No 
Ventral midbrain Yes No No 
Periaqueductal grey Yes No No 
Dorsal pons Yes Yes Yes 
Putamen Yes No No 
Pulvinar Yes Yes No 
Caudate Yes No No 
Thalamus No Yes No 
Occipital cortex Yes Yes No 
Cingulate cortex Yes Yes No 
Insula No No Yes 
Temporal cortex Yes No No 
Frontal cortex Yes No Yes 
Parietal cortex/precuneus Yes No Yes 
Cerebellum Yes No Yes 
Brain area Early premonitory Late premonitory Migraine headache 
Hypothalamus Yes No No 
Ventral midbrain Yes No No 
Periaqueductal grey Yes No No 
Dorsal pons Yes Yes Yes 
Putamen Yes No No 
Pulvinar Yes Yes No 
Caudate Yes No No 
Thalamus No Yes No 
Occipital cortex Yes Yes No 
Cingulate cortex Yes Yes No 
Insula No No Yes 
Temporal cortex Yes No No 
Frontal cortex Yes No Yes 
Parietal cortex/precuneus Yes No Yes 
Cerebellum Yes No Yes 

For anatomical labelling of cortical and diencephalic areas, we initially converted the MNI coordinates into Talairach coordinates (non-linear) using http://www.brainmap.org/ale/. Then, we identified the nearest grey matter area for a given set of co-ordinates using the Talairach daemon software (http://www.talairach.org/client). For brainstem areas, we used the human atlas by DeArmond et al. (1976) to identify the areas.

There was no significant difference between mean systolic and diastolic blood pressure measurements in baseline, premonitory and migraine headache phases (paired t-tests, P > 0.05).

Discussion

We performed H215O PET cerebral blood flow scans as a surrogate for neuronal activity, referred to hereafter as activations (Frackowiak and Friston, 1994), in nitroglycerin-induced migraine. In the early premonitory phase as compared to baseline, we found activations in the right posterior and lateral regions of the hypothalamus (MNI co-ordinates = 6, −6, −12) and the adjacent right midbrain ventral tegmentum (Fig. 1).These activations were more anterior to those observed in acute cluster headache (May et al., 1998) and distant from those previously reported in the headache phase of the attack (Denuelle et al., 2007). The data offer insight into areas of the brain pivotal for the early phases of migraine suggesting structures that can be explored at the bench and neurotransmitter systems that may have therapeutic potential at the bedside.

Hypothalamic and ventral tegmental involvement would explain yawning, possibly related to dopaminergic mechanisms (Argiolas and Melis, 1998); frequent urination and thirst, possibly relate to reduced vasopressin (Krowicki and Kapusta, 2011) and mood changes through hypothalamic connections with the limbic system (Sokolowski and Corbin, 2012). The dopamine D2 receptor antagonist domperidone is reported to be effective in preventing migraine headache when taken early in the attack, highlighting the importance of dopamine in the early stages (Waelkens, 1982, 1984). The posterior/lateral hypothalamic region is the main location for hypocretin/orexin messenger RNA, which produces two related neuropeptides called hypocretin/orexin A and B (de Lecea et al., 1998). Projections of the orexinergic system have been implicated in a variety of functions such as feeding, sleep–wake cycle, neuroendocrine and autonomic functions (Ferguson and Samson, 2003; Sakurai, 2005), all of which are relevant in migraine, most strikingly in the premonitory phase, although the role of orexin in trigeminal nociception can be either pro- or anti-nociceptive depending on the receptor activated (Bartsch et al., 2004). Only one previous neuroimaging study has shown hypothalamic activation in migraine; this was anterior to what is now reported, and crucially seen during migraine headache (Denuelle et al., 2007). These investigators scanned their patients within 4 h of onset of spontaneous migraine headache, which is earlier than in other previously reported studies in spontaneous migraine, with in 6 h (Weiller et al., 1995) or within 24 h (Afridi et al., 2005). Hence, hypothalamic activation may occur in phases: right at the start of the attack during the premonitory phase, and later on during the headache. No previous neuroimaging study has investigated migraineurs just as the earliest signs of an attack are manifest—the premonitory phase.

The activations we found extended into the right midbrain, in the region of substantia nigra and the ventral tegmental area that gives rise to a dopaminergic projection to the hypothalamus through the medial forebrain bundle (You et al., 2001). Rewarding stimulation of the lateral hypothalamus leads to release of dopamine in the ventral tegmental area and nucleus accumbens, suggesting these structures could be involved in craving, which is a characteristic premonitory symptom (You et al., 2001; Giffin et al., 2003; Kelman, 2004). We also found activations in the region of the periaqueductal grey and the red nucleus. The periaqueductal grey is an important region in pain control in general and migraine in particular. In a previous study, stereotactic electrodes placed in this area for chronic pain produced migraine-like headaches in patients without previous history of migraines (Raskin et al., 1987). Dorsal pontine activation was seen in the region of the locus coeruleus. The locus coeruleus is a noradrenergic nucleus controlling the extracerebral and intracerebral vascular tone and has a modulatory effect on cortical excitability, thus possibly contributing to sensory dysmodulation leading to symptoms, such as photophobia and phonophobia (Akerman et al., 2011). The hypothalamus (Charbit et al., 2009), periaqueductal grey (Knight et al., 2002) and locus coeruleus (Lance et al., 1983) have been shown to influence physiology relevant to migraine. It can be speculated that activations in these areas in the premonitory phase could represent dysfunction that can potentially remove the top–down inhibitory effect on the trigeminocervical complex leading on to the initial neck stiffness/discomfort and eventually to the attack (Akerman et al., 2011).

We found activations in predominantly right-sided frontal, and temporal cortex, and bilateral occipital cortex. The latter likely contributes to photophobia, which three of eight of our patients had. The cortical activation became less prominent as the headache appeared, in line with electrophysiological data showing restitution of habituation with headache (Schoenen et al., 2003). Our study does not indicate whether the cortical or subcortical structures are involved first, although a previous study showed activation of the substantia nigra and red nucleus before occipital cortical activation in visually triggered migraine (Cao et al., 2002). During the late premonitory phase, activations in the hypothalamus and midbrain regions were not seen and additionally activations were seen in the right pulvinar nucleus of thalamus and the left thalamus, contralateral to headache. Similarly, activation in the hypothalamic region was not seen when all premonitory scans were compared against baseline in our study, suggesting this activation is a very early phenomenon. Finally, in the migraine headache phase, activations were seen in the dorsal rostral pons and structures generally involved in pain processing as has been previously reported (Table 3) (Weiller et al., 1995; Bahra et al., 2001; Afridi et al., 2005a, b).

It is noteworthy that the hypothalamic and brainstem activations in the premonitory phase were exclusively right-sided and patients developed purely or predominantly right-sided headache during the migraine headache phase. This suggests that ipsilateral involvement of these structures before the onset of headache may be involved in headache generation. In another study primarily looking at laterality of brainstem changes during nitroglycerin-triggered migraine headache it was shown that these changes occur ipsilateral to headache (Afridi et al., 2005). The data are consistent with the anatomy of the first-order trigeminovascular neurons and the ipsilateral influence of these structures, particularly the periaqueductal grey and serotonergic raphe nuclei, on the trigeminocervical complex. However, it should be said that laterality was not an aim of our study and we did not choose patients with right-sided headache only during habitual attacks. The occurrence of pure or predominantly right-sided headache during the scanning session was by chance. Furthermore, one needs to study left-sided headache also to fully establish the ipsilateral involvement of these structures in the premonitory phase.

The activations we report are on the basis of increases in regional cerebral blood flow. We believe these changes are reflective of neuronal activity, as has been assumed in previous PET studies in migraine. We believe that there is no convincing literature supporting the view that the coupling between neuronal activity and haemodynamic response is abnormal in migraineurs globally. Rather previous literature indicates it may be normal in migraineurs, at least interictally (Schytz et al., 2010). However, neurovascular coupling has not been systematically studied in the premonitory or migraine headache phases when triggered by nitroglycerin. All we can conclude is that the increased regional cerebral blood flow suggests a change in neuronal activity without making any assumptions about this is stimulatory or inhibitory to the connected neurons or the degree/intensity of the change.

Potential study limitations

Our study did not include a contemporaneous control group, so we cannot absolutely rule out that the observed activation changes during the premonitory and migraine headache phase relate to nitroglycerin administration rather than to migraine. However, we have previously studied the effect of nitroglycerin in PET using the same approach (Afridi et al., 2005b). In addition to featureless headache soon after nitroglycerin and activations in the anterior cingulate cortex areas, we saw no changes in the regions here reported for the premonitory phase. It could be argued that exposing further controls to this method is unnecessary. Furthermore, the half-life of intravenous nitroglycerin is short, <5 min, as it is rapidly broken down into two inactive metabolites (Abrams, 1985). Considering the mean time of 67 min to the first premonitory scan after the start of the nitroglycerin infusion, a vasodilator effect seems unlikely. A second potential limitation is the study was relatively small. The sample size is typical for PET in migraine (Weiller et al., 1995; Afridi et al., 2005a, b), and given the biological plausibility, and the nature of the analysis, a positive result with smaller numbers is regarded by methodological experts as a very strong outcome (Friston, 2012). Lastly, we screened 142 patients of which eight patients were included in the final analysis. This selection goes to generalizability. Our data certainly do not address patients with chronic migraine or those with confounding medical conditions and treatments. By definition we cannot address patients who do not have premonitory symptoms. The large exclusion rate reflects our design to have a pure sample, and the nature of social media and clinic patients who are generally more disabled and thus tend to reply to such advertisements. It seems likely our data reflect the more routine migraineur.

In conclusion, we provide neuroimaging evidence for the involvement of the hypothalamus and midbrain areas, periaqueductal grey and dorsal pons, in the earliest stages of migraine before the appearance of headache. For the first time the earliest part of the migraine attack has been imaged and there are clear changes without any pain. Hypothalamic involvement can explain many of the premonitory symptoms and may provide some insight into why migraine is commonly triggered by a change in homeostasis. The data offer targets for neurobiological studies at the bench and plausible neurotransmitter systems, such as dopaminergic and orexinergic targets, for bedside development of novel therapeutics.

Acknowledgements

We are thankful to Carole Schreck (chief PET technician); Seante Baker, BS (Study coordinator); Janet Corroo, RN; Denise Chou MD, Henry VanBrocklin PhD, and Mei Mei Church, NP. We are most grateful to all our patients for their participation in the study.

Funding

This work was supported by an unrestricted grant from GlaxoSmithKline and by Miles For Migraine (http://milesformigraine.org/). Till Sprenger was supported by a grant from the Deutsche Forschungsgemeinschaft (SP1215/1-1).

Supplementary material

Supplementary material is available at Brain online.

References

Abrams
J
Pharmacology of nitroglycerin and long-acting nitrates
Am J Cardiol
 , 
1985
, vol. 
56
 (pg. 
12A
-
18A
)
Afridi
S
Giffin
NJ
Kaube
H
Friston
KJ
Ward
NS
Frackowiak
RSJ
, et al.  . 
A PET study in spontaneous migraine
Arch Neurol
 , 
2005a
, vol. 
62
 (pg. 
1270
-
5
)
Afridi
S
Kaube
H
Goadsby
PJ
Glyceryl trinitrate triggers premonitory symptoms in migraineurs
Pain
 , 
2004
, vol. 
110
 (pg. 
675
-
80
)
Afridi
S
Matharu
MS
Lee
L
Kaube
H
Friston
KJ
Frackowiak
RSJ
, et al.  . 
A PET study exploring the laterality of brainstem activation in migraine using glyceryl trinitrate
Brain
 , 
2005b
, vol. 
128
 (pg. 
932
-
9
)
Akerman
S
Holland
P
Goadsby
PJ
Diencephalic and brainstem mechanisms in migraine
Nat Rev Neurosci
 , 
2011
, vol. 
12
 (pg. 
570
-
84
)
Argiolas
A
Melis
MR
The neuropharmacology of yawning
Eur J Pharmacol
 , 
1998
, vol. 
343
 (pg. 
1
-
16
)
Bahra
A
Matharu
MS
Buchel
C
Frackowiak
RSJ
Goadsby
PJ
Brainstem activation specific to migraine headache
Lancet
 , 
2001
, vol. 
357
 (pg. 
1016
-
7
)
Bartsch
T
Levy
MJ
Knight
YE
Goadsby
PJ
Differential modulation of nociceptive dural input to [hypocretin] Orexin A and B receptor activation in the posterior hypothalamic area
Pain
 , 
2004
, vol. 
109
 (pg. 
367
-
78
)
Cao
Y
Aurora
SK
Nagesh
V
Patel
SC
Welch
KM
Functional MRI-BOLD of brainstem structures during visually triggered migraine
Neurology
 , 
2002
, vol. 
59
 (pg. 
72
-
8
)
Charbit
AR
Akerman
S
Holland
PR
Goadsby
PJ
Neurons of the dopaminergic/CGRP A11 cell group modulate neuronal firing in the trigeminocervical complex: an electrophysiological and immunohistochemical study
J Neurosci
 , 
2009
, vol. 
29
 (pg. 
12532
-
41
)
Christiansen
I
Thomsen
LL
Daugaard
D
Ulrich
V
Olesen
J
Glyceryl trinitrate induces attacks of migraine without aura in sufferers of migraine with aura
Cephalalgia
 , 
1999
, vol. 
19
 (pg. 
660
-
7
)
de Lecea
L
Kilduff
TS
Peyron
C
Gao
X
Foye
PE
Danielson
PE
, et al.  . 
The hypocretins: hypothalamus-specific peptides with neuroexcitatory activity
Proc Natl Acad Sci USA
 , 
1998
, vol. 
95
 (pg. 
322
-
7
)
DeArmond
SJ
Fusco
MM
Dewey
MM
Photographic atlas—structure of the human brain
 , 
1976
2nd edn
New York
Oxford University Press
Denuelle
M
Fabre
N
Payoux
P
Chollet
F
Geraud
G
Hypothalamic activation in spontaneous migraine attacks
Headache
 , 
2007
, vol. 
47
 (pg. 
1418
-
26
)
Ferguson
AV
Samson
WK
The orexin/hypocretin system: a critical regulator of neuroendocrine and autonomic function
Front Neuroendocrinol
 , 
2003
, vol. 
24
 (pg. 
141
-
50
)
Frackowiak
RSJ
Friston
KJ
Functional neuroanatomy of the human brain: positron emission tomography—a new neuroanatomical technique
J Anat
 , 
1994
, vol. 
184
 (pg. 
211
-
25
)
Friston
K
Ten ironic rules for non-statistical reviewers
Neuroimage
 , 
2012
, vol. 
61
 (pg. 
1300
-
10
)
Giffin
NJ
Ruggiero
L
Lipton
RB
Silberstein
S
Tvedskov
JF
Olesen
J
, et al.  . 
Premonitory symptoms in migraine: an electronic diary study
Neurology
 , 
2003
, vol. 
60
 (pg. 
935
-
40
)
Headache Classification Committee of The International Headache Society
The international classification of headache disorders (second edition)
Cephalalgia
 , 
2004
, vol. 
24
 
Suppl 1
(pg. 
1
-
160
)
Iversen
HK
Olesen
J
Tfelt-Hansen
P
Intravenous nitroglycerin as an experimental headache model
Basic characteristics. Pain
 , 
1989
, vol. 
38
 (pg. 
17
-
24
)
Kelman
L
The premonitory symptoms (prodrome): a tertiary care study of 893 migraineurs
Headache
 , 
2004
, vol. 
44
 (pg. 
865
-
72
)
Knight
YE
Bartsch
T
Kaube
H
Goadsby
PJ
P/Q-type calcium channel blockade in the PAG facilitates trigeminal nociception: a functional genetic link for migraine?
J Neurosci
 , 
2002
, vol. 
22
 (pg. 
1
-
6
)
Krowicki
ZK
Kapusta
DR
Microinjection of glycine into the hypothalamic paraventricular nucleus produces diuresis, natriuresis, and inhibition of central sympathetic outflow
J Pharmacol Exp Ther
 , 
2011
, vol. 
337
 (pg. 
247
-
55
)
Lance
JW
Lambert
GA
Goadsby
PJ
Duckworth
JW
Brainstem influences on cephalic circulation: experimental data from cat and monkey of relevance to the mechanism of migraine
Headache
 , 
1983
, vol. 
23
 (pg. 
258
-
65
)
Lassen
LH
Ashina
M
Christiansen
I
Ulrich
V
Olesen
J
Nitric oxide synthesis inhibition in migraine
Lancet
 , 
1997
, vol. 
349
 (pg. 
401
-
2
)
May
A
Bahra
A
Buchel
C
Frackowiak
RS
Goadsby
PJ
Hypothalamic activation in cluster headache attacks
Lancet
 , 
1998
, vol. 
352
 (pg. 
275
-
8
)
Murray
CJ
Vos
T
Lozano
R
Naghavi
M
Flaxman
AD
Michaud
C
, et al.  . 
Disability-adjusted life years (DALYs) for 291 diseases and injuries in 21 regions, 1990-2010: a systematic analysis for the Global Burden of Disease Study 2010
Lancet
 , 
2012
, vol. 
380
 (pg. 
2197
-
223
)
Olesen
J
Gustavsson
A
Svensson
M
Wittchen
HU
Jonsson
B
The economic cost of brain disorders in Europe
Eur J Neurol
 , 
2012
, vol. 
19
 (pg. 
155
-
62
)
Olesen
J
Thomsen
LL
Lassen
LH
Olesen
IJ
The nitric oxide hypothesis of migraine and other vascular headaches
Cephalalgia
 , 
1995
, vol. 
15
 (pg. 
94
-
100
)
Raskin
NH
Hosobuchi
Y
Lamb
S
Headache may arise from perturbation of brain
Headache
 , 
1987
, vol. 
27
 (pg. 
416
-
20
)
Sakurai
T
Reverse pharmacology of orexin: from an orphan GPCR to integrative physiology
Regul Pept
 , 
2005
, vol. 
126
 (pg. 
3
-
10
)
Schoenen
J
Ambrosini
A
Sandor
PS
Maertens de Noordhout
A
Evoked potentials and transcranial magnetic stimulation in migraine: published data and viewpoint on their pathophysiologic significance
Clin Neurophysiol
 , 
2003
, vol. 
114
 (pg. 
955
-
72
)
Schytz
HW
Ciftci
K
Akin
A
Ashina
M
Bolay
H
Intact neurovascular coupling during executive function in migraine without aura: interictal near-infrared spectroscopy study
Cephalalgia
 , 
2010
, vol. 
30
 (pg. 
457
-
66
)
Sokolowski
K
Corbin
JG
Wired for behaviors: from development to function of innate limbic system circuitry
Front Mol Neurosci
 , 
2012
, vol. 
5
 pg. 
55
 
Stewart
WF
Ricci
JA
Chee
E
Morganstein
D
Lipton
R
Lost productive time and cost due to common pain conditions in the US workforce
JAMA
 , 
2003
, vol. 
290
 (pg. 
2443
-
54
)
Waelkens
J
Domperidone in the prevention of complete classical migraine
BMJ
 , 
1982
, vol. 
284
 pg. 
944
 
Waelkens
J
Dopamine blockade with domperidone: bridge between prophylactic and abortive treatment of migraine? A dose-finding study
Cephalalgia
 , 
1984
, vol. 
4
 (pg. 
85
-
90
)
Weiller
C
May
A
Limmroth
V
Juptner
M
Kaube
H
Schayck
RV
, et al.  . 
Brain stem activation in spontaneous human migraine attacks
Nat Med
 , 
1995
, vol. 
1
 (pg. 
658
-
60
)
You
ZB
Chen
YQ
Wise
RA
Dopamine and glutamate release in the nucleus accumbens and ventral tegmental area of rat following lateral hypothalamic self-stimulation
Neuroscience
 , 
2001
, vol. 
107
 (pg. 
629
-
39
)

Author notes

*These authors contributed equally to this work.
Present address: Department of Neurology, The Royal London Hospital (Barts and the London NHS Trust), Whitechapel Road, Whitechapel, London E1 1BB, UK and Basildon and Thurrock University Hospitals NHS Foundation Trust, Nethermayne, Essex SS16 5NL, UK
#Present address: Department of Neurology and Division of Neuroradiology, University Hospital Basel, Switzerland, Petersgraben 4, 4031 Basel, Switzerland
§Present address: Department of Neurology, University of Miami, Miami FL, USA