Sage Journals: Discover world-class research

Abstract

Objective

To elucidate the hypothalamic involvement in episodic migraine and investigate the association between hypothalamic resting state functional connectivity changes and migraine patients’ clinical characteristics and disease progression over the years.

Methods

Ninety-one patients with episodic migraine and 73 controls underwent interictal resting state functional magnetic resonance imaging. Twenty-three patients and controls were re-examined after a median of 4.5 years. Hypothalamic resting state functional connectivity changes were investigated using a seed-based correlation approach.

Results

At baseline, a decreased functional interaction between the hypothalamus and the parahippocampus, cerebellum, temporal, lingual and orbitofrontal gyrus was found in migraine patients versus controls. Increased resting state functional connectivity between the hypothalamus and bilateral orbitofrontal gyrus was demonstrated in migraine patients at follow-up versus baseline. Migraine patients also experienced decreased right hypothalamic resting state functional connectivity with ipsilateral lingual gyrus. A higher migraine attack frequency was associated with decreased hypothalamic-lingual gyrus resting state functional connectivity at baseline, while greater headache impact at follow-up correlated with decreased hypothalamic-orbitofrontal gyrus resting state functional connectivity at baseline. At follow-up, a lower frequency of migraine attacks was associated with higher hypothalamic-orbitofrontal gyrus resting state functional connectivity.

Conclusions

During the interictal phase, the hypothalamus modulates the activity of pain and visual processing areas in episodic migraine patients. The hypothalamic-cortical interplay changes dynamically over time according to patients’ clinical features.

Keywords

Pathophysiology migraine hypothalamus biomarkers longitudinal fMRI

Introduction

Migraine is widely recognised as an intricate neurological disease that comprises the interaction of different brain networks (1). Neuroimaging studies have widened the understanding of migraine physiopathology, showing that the activation of specific brain regions can explain the broad clinical spectrum of migraine (2).

The hypothalamus is one of the key actors in the manifestation of migraine. Positron emission tomography and functional magnetic resonance imaging (fMRI) studies (3,4) demonstrated an early hypothalamic activation in the premonitory phase of migraine, suggesting that the hypothalamus could mediate symptoms like yawning, food craving, thirst, and fatigue. An increased hypothalamic activation has also been demonstrated in the pain phase of migraine (3 –5). During trigeminal nociceptive stimulation, the functional connection between the hypothalamus and brain regions implicated in the generation of migraine headache, such as the spinal trigeminal nucleus and dorsal pons, changes throughout the preictal and pain phase. These findings point to a hypothalamic involvement in the onset of the migraine attack and highlight dynamic functional changes of the hypothalamic connectivity during the migraine cycle (3).

Although many studies have investigated the role of the hypothalamus in the acute phases of the migraine attack in episodic migraine patients, only one small study has explored its involvement in the interictal phase. It has been shown that the functional connectivity (FC) between the hypothalamus and cortical regions implicated in pain processing, sympathetic and parasympathetic functions is altered in episodic migraine patients without aura studied during the interictal phase (6). Whether the interictal functional alterations of the hypothalamus could evolve over the years is unknown.

There is evidence supporting a role of the hypothalamus in migraine chronification (7). The hypothalamic interaction with pain processing brain areas is stronger in chronic migraine patients in comparison with patients with an episodic form of the disease (8). A recent study showed that a decreased resting state (RS) FC between the hypothalamus and medial prefrontal cortex, an area that is involved in pain perception, could modulate the pain severity perceived by chronic migraine patients during their attack (9). Even though the role of the hypothalamus in determining the severity of migraine is well established in patients with a chronic form of the disease (8,9), the association between hypothalamic functional alterations and clinical features of episodic migraine patients has not been investigated so far. In this study, we hypothesised that episodic migraine patients experience hypothalamic RS FC alterations that are influenced by patients’ clinical features (e.g. disease duration, migraine attack frequency, presence of aura). We also assumed that the functional interplay between the hypothalamus and other brain regions implicated in migraine pathophysiology might change over time, thus affecting the disease activity. To test our hypotheses, we explored cross-sectional hypothalamic RS FC alterations in a large cohort of episodic migraine patients and assessed the correlation between hypothalamic functional abnormalities and patients’ clinical features. We also followed the clinical evolution of patients over 4 years and explored longitudinal hypothalamic RS FC changes and their association with clinical data and disease progression, as measured by changes in migraine frequency. Finally, we investigated whether baseline hypothalamic RS FC alterations could influence migraine severity over time.

Methods

Participants: Between October 2006 and May 2016, we prospectively studied 91 right-handed, episodic migraine patients (42 patients with migraine with aura (MWA) and 49 without aura (MWoA)) and 73 right-handed controls. All participants were asked to participate in a clinical and MRI follow-up evaluation. Both baseline and follow-up visits included a detailed clinical evaluation and an MRI session, including RS fMRI and structural MRI sequences. Results obtained from the structural MRI analysis performed on part of the patients included in the present study were previously published (10). To avoid measuring imaging changes associated to acute migraine symptoms, all patients had to be attack-free, including aura, in the 48 h before the MRI and during the exam. Exclusion criteria for patients and controls included the presence of vascular risk factors (e.g. vascular disease, heart disease, hypercholesteremia, hypertension, diabetes mellitus), abnormal neurological exam, systemic conditions, other psychiatric, or neurologic diseases. Controls were excluded from the study if they suffered from any headaches, with the exception of infrequent tension-type headache (<1 headache day/month). Patients who attended the Headache Outpatient Clinic of the IRCCS San Raffaele Scientific Institute (Milan, Italy) were enrolled in the study. Migraine was diagnosed applying the standard diagnostic criteria of the International Classification of Headache Disorders (11,12). Controls were recruited among hospital workers, university students and consented friends. The clinical assessment of all participants was performed by a single neurologist, before the MRI evaluation.

At baseline and follow-up, we obtained an accurate clinical history of patients, comprising their frequency of migraine attacks and disease duration. At follow-up, the median headache pain severity of the 3 months preceding the visit was evaluated using the Numerical Rating Scale (NRS) (13), and patients’ disability was quantified using the 6-item Headache Impact Test (HIT-6) (14) and Migraine Disability Assessment questionnaire (MIDAS) (15).

At baseline, 25 patients were taking preventive therapies for migraine, comprising topiramate, β-blockers, pizotifen, amitriptyline and flunarizine. During the follow-up, four patients never stopped taking preventive medications, eight patients stopped taking preventive treatments and one patient started a new preventive drug.

To explore whether hypothalamic functional changes might be related to disease progression and migraine phenotype, patients were divided into subgroups according to changes in migraine frequency over the follow-up (IoS = patients with decreased or stable attack frequency at follow-up, Wo = patients with increased attack frequency at follow-up) and presence/absence of aura. Patients’ improvement or worsening was evaluated, taking note of the number of migraine days reported by patients in their headache diaries at the baseline and follow-up visit.

Patient consents and protocol approvals: The Local Ethical Committee on human studies approved this study. Written informed consent was obtained from all participants before study entry.

Image acquisition: Using a 3.0 Tesla Intera scanner (Philips Medical Systems, Best, The Netherlands), the following brain sequences were obtained from all participants at baseline and follow-up: a) RS fMRI scans (T2*-weighted echo planar imaging sequence with 200 sets of 30 contiguous axial slices, slice thickness = 4 mm, matrix size = 96 × 96, reconstructed to 128 × 128, repetition time [TR]/echo time [TE] = 3000/35 ms, flip angle [FA] = 90°, field of view [FOV] = 240 mm²); b) T2-weighted turbo-spin echo (28 contiguous axial slices, 4 mm thick, TR/TE = 3000/120 ms, FA = 90°, matrix size = 512 × 512, FOV = 230 mm²); c) fluid attenuated inversion recovery (28 axial contiguous slices, 4 mm thick, TR/TE = 11000/120 ms, inversion time = 2800 ms, FA = 90°, matrix size = 256 × 256, FOV = 230 mm²); and d) 3D T1-weighted fast field echo (220 contiguous axial slices with voxel size = 0.89 × 0.89 × 0.8 mm, TR/TE = 25/4.6 ms; FA = 30°; matrix size = 256 × 256; FOV = 230 ×230 mm²). During the RS fMRI acquisition, participants were instructed to stay awake with their eyes closed. The same patient positioning procedure was used at the two study time points and baseline MRI localiser images were used as reference to achieve the same slice positioning on baseline and follow-up MRI exams.

MRI analysis: Statistical Parametric Mapping (SPM) 12 (https://www.fil.ion.ucl.ac.uk/spm/) and REST software (https://resting-fmri.sourceforge.net) were used to analyse baseline and follow-up RS fMRI images. The pre-processing of RS fMRI images comprised the following steps: a) rigid-body realignment of raw RS fMRI images to the mean of each session, to correct for minor head movements; b) rigid registration of realigned images to the 3D T1-weighted image; c) normalisation of RS fMRI data to the Montreal Neurological Institute (MNI) template using a standard affine transformation followed by nonlinear warping; d) linear detrending and band-pass filtering (0.01–0.08 Hz), performed to partially remove low frequency drifts and physiological high-frequency noise; e) removal of non-neuronal sources of synchrony between RS fMRI time series by regressing out the six motion parameters estimated by SPM12 and the average signals of the ventricular cerebro-spinal fluid and white matter; f) smoothing of normalised images using a 3D 6-mm Gaussian kernel.

After pre-processing, a seed-based correlation approach was applied to study voxel-wise baseline and follow-up RS FC between the left and right hypothalamus, separately, and the remaining voxels of the brain (16). Based on previous studies (4,7), we used a 6-mm sphere around the peak MNI coordinates of hypothalamic activation (X = ±6, Y = −6, Z = −10). Briefly, the correlation coefficients between the average time series extracted from the left and right hypothalamus and any other brain voxels were calculated. The Gaussianity of the obtained correlation coefficients was improved using a Fisher’s z transform (17).

Statistical analysis: Given the challenges of retaining a big sample of patients and controls throughout the entire follow-up period in a 4-year study, based on previous literature (18), we estimated that a sample size of at least 20 patients would have been acceptable to assess fMRI abnormalities in headache disorders. The distribution of the data was assessed using the Q-Q plots, Kolmogorov–Smirnov and Shapiro–Wilk tests. Since the data were not normally distributed, between-group differences in demographic and clinical variables were assessed using the Mann–Whitney test for continuous variables and Fisher’s exact test for categorical variables (version 26.0; SPSS software, IBM, Armonk, NY). Using SPM12, the following linear models were performed on hypothalamic RS FC maps: a) average baseline hypothalamic RS FC within the group of migraine patients and controls, separately (age- and sex-adjusted one sample t-tests); b) between-group comparisons (patients vs. controls, MWA vs. MWoA) of hypothalamic RS FC at baseline (age- and sex-adjusted two-sample t-tests and conjunction analyses); c) changes over time of hypothalamic RS FC in the entire group of patients with migraine, each subgroup of patients and controls, separately (age-adjusted paired sample t-tests); and d) time-by-group interactions tested comparing the RS FC delta at follow-up vs. baseline between migraine patients and controls and between the different subgroups of patients (age- and sex-adjusted two-sample t-tests). To exclude a possible effect of migraine preventive treatments, we performed a between-group comparison of hypothalamic RS FC at baseline considering the use of preventive drugs (age- and sex-adjusted two-sample t-tests and conjunction analyses) and explored longitudinal hypothalamic RS FC changes in the subgroup of patients who were not taking any preventive drugs at baseline and follow-up. The use of preventive therapies was included as an additional covariate in the analysis investigating the longitudinal effect of disease progression. To prevent misinterpretations associated with anticorrelated connections, only brain regions showing positive RS FC with the hypothalamus in both patients and controls were included in the between-group comparison analyses (19). In migraine patients, multiple linear regression models, adjusted for age and sex, were performed using SPM12 to investigate the association between abnormal hypothalamic RS FC and clinical characteristics at baseline and follow-up (frequency of migraine attacks, disease duration, NRS, MIDAS and HIT-6 scores). A statistical threshold of p < 0.05, FWE corrected, and p < 0.001, uncorrected was used to test the results. Exploratory analyses in the subgroups of migraine patients were also tested at a p < 0.05, uncorrected for multiple comparisons.

Data availability: Data supporting the results of this study are available from the corresponding author, upon reasonable request.

Results

Clinical and demographic findings: The main clinical and demographic characteristics of participants are summarised in Table 1. Twenty-three migraine patients (11 MWA and 12 MWoA) and 23 controls agreed to participate to the clinical and MRI follow-up evaluation after a median of 4.5 years (interquartile range = 2–5 years; patients: mean follow-up years: 4.5, range: 3–6; controls: mean follow-up years: 3.9, range: 2–6). Sixty-eight patients and 50 controls withdrew from the follow-up evaluation because they moved to another city, due to familial or work commitments, pregnancy or interest loss. During the follow-up, 11 patients (48%) reported a reduction of migraine attack frequency, eight patients (35%) had an increased number of migraine attacks and in the remaining four patients (17%) the migraine attack frequency did not change. We found no association between patients’ disease progression during the follow-up and the use of migraine preventive drugs at baseline (p = 0.2) and follow-up (p = 1.0).

Table 1.

Main demographic and clinical characteristics of migraine patients and controls enrolled in the study.

Baseline
	Controls	Migraine patients	MWA patients	MWoA patients	p-values patients vs. controls			p-values MWA vs. MWoA patients
Women/men	47/26	65/26	29/13	36/13	0.4			0.6
Age (years)	27 (25–41)	34 (27–42)	33 (26–39)	37 (28–47)	0.02*			0.04*
Attack frequency per month	–	3.5 (1–6)	2.3 (1–5)	4 (2–8)	–			0.009*
Disease duration (years)	–	15 (10–21)	14 (4–21)	16 (10–22)	–			0.3
Use of preventive medications	–	25	9	16	–			0.3
Follow-up
	Controls	Migraine patients	MWA patients	MWoA patients	IoS patients	Wo patients	p-values patients vs. controls		p-values MWA vs. MWoA patients	p values IoS vs. Wo patients
Women/men	9/14	15/8	9/2	6/6	10/5	5/3	0.1		0.06	1
Age (years)	36 (26–45)	45 (36–48)	37 (31–48)	45 (41–50)	46 (37–50)	44 (33–48)	0.02*		0.06	0.6
Attack frequency per month	–	3 (1–5)	1.5 (0.1–4)	4 (2–7)	1.5 (0.4–4)	4.5 (2 –9)	–		0.03*	0.01*
Duration of follow-up (years)	4.5 (2–5)	4.7 (4-5)	4.8 (4–5)	4.5 (4–5)	4.7 (4–5)	4.5 (4–5)	0.3		0.4	0.8
Use of preventive medications	–	5	1	4	3	2	–		0.4	1.0
MIDAS score	–	11 (6–30)	15 (8–28)	9 (2–46)	11 (4–30)	17 (6–41)	–		0.5	0.9
HIT-6 score	–	60 (55–67)	59 (55–67)	60 (58–68)	58 (55–68)	64 (59–67)	–		0.5	0.3
NRS score	–	7 (6–8)	7 (5–8)	7 (6–8)	6.5 (5–8)	8 (7 –9)	–		0.9	0.5

HIT-6: 6-item Headache Impact Test; IoS: improved or stable migraine; MIDAS: Migraine Disability Assessment; MWA: migraine with aura; MWoA: migraine without aura; NRS: Numerical Rating Scale; Wo: worsening migraine.

Note: Measures are reported as medians and interquartile ranges (25th–75th percentiles). Sex and the use of preventive medications are reported as frequencies.

*Mann–Whitney test, p < 0.05.

We did not find any sex differences in migraine patients versus controls (baseline: p = 0.4, follow-up: p = 0.1). At the two time points, patients with migraine were older than controls (baseline: p = 0.02, follow-up: p = 0.02). At baseline, patients with MWoA were older than patients with MWA (p = 0.04). Compared to MWA, patients with MWoA had a higher attack frequency at baseline (p = 0.0009) and follow-up (p = 0.03). At follow-up, except from migraine attack frequency, we did not find any significant demographic and clinical differences among migraine patients with improved or stable migraine and those with worsening migraine (Table 1).

Baseline demographic and clinical characteristics did not differ between the entire group of participants and the subgroup of patients and controls who underwent longitudinal assessment (Supplementary Table 1).

Baseline hypothalamic RS FC: Supplementary Table 2 summarises the brain areas with positive hypothalamic RS FC in patients with migraine and controls at baseline. Both patients and controls showed positive left and right hypothalamic RS FC with cerebellar, occipital, frontal and temporal regions.

Decreased RS FC between the right and left hypothalamus and left parahippocampus and bilateral orbitofrontal gyrus (OFG) was found in migraine patients versus controls (Figure 1 and Table 2). Compared to controls, migraine patients experienced also decreased RS FC between the left hypothalamus and right cerebellar crus II, as well as decreased RS FC between the right hypothalamus and the right lingual gyrus, cerebellar lobule VI and inferior temporal gyrus (Figure 1 and Table 2). Similar findings were obtained when the subgroup of patients who were on preventive drugs for migraine and those who were not taking any treatments were separately compared to controls (Supplementary Table 3). We found no difference of hypothalamic RS FC in patients who were on preventive drugs for migraine compared to those who were not taking medications.

Figure 1.

Hypothalamic resting state functional connectivity alterations in migraine patients at baseline. Areas of decreased hypothalamic resting state (RS) functional connectivity (FC) in migraine patients compared to controls (two sample t-test, p < 0.05, clusterwise FWE-corrected for multiple comparisons), color-coded in blue based on their t values. (a) RS FC alterations of the left hypothalamus; (b) RS FC alterations of the right hypothalamus.

Table 2.

Regions showing significant decrease of hypothalamic resting state functional connectivity in patients with migraine compared to controls at baseline, as well as among subgroups of patients.

Migraine patients vs. controls
Left hypothalamus
Connected regions	Brodmann area	t values*	Cluster extent (no of voxels)	MNI coordinates (X, Y, Z)
R cerebellum (crus II)	–	5.31	141	50, −62, −46
L orbitofrontal gyrus	11	4.62	167	−16, 10, −16
L parahippocampal gyrus	28	4.10	167	−12, −4, −24
R orbitofrontal gyrus	11	4.62	185	16, 26, −18

Right hypothalamus
Connected regions	Brodmann area	t values*	Cluster extent (no of voxels)	MNI coordinates(X, Y, Z)
L parahippocampal gyrus	28	4.43	480	−12, −4, −24
L orbitofrontal gyrus	11	3.85		−18, 10, −16
R orbitofrontal gyrus	11	4.13		18, 18, −20
R inferior temporal gyrus	37	4.32	228	60, −50, −22
R lingual gyrus	18	4.27	244	20, −74, 2
R cerebellum (lobule VI)	–	3.79	244	6, −64, −10

MWoA vs. MWA

Left hypothalamus
Connected regions	Brodmann area	t values*	Cluster extent (no of voxels)	MNI coordinates (X, Y, Z)
R temporal pole	48	3.76	131	66, 4, −2

MWoA vs. controls

Left hypothalamus

Connected regions	Brodmann area	t values**	Cluster extent (no of voxels)	MNI coordinates (X, Y, Z)
R temporal pole	48	4.32	77	64, 6, −4

*Age- and sex-adjusted linear model (p < 0.05, clusterwise FWE-corrected for multiple comparisons).

**Age- and sex-adjusted linear model (p < 0.001, uncorrected).

L: left; MWA: migraine with aura; MWoA: migraine without aura; R: right.

Compared with MWA patients and controls, MWoA patients showed decreased RS FC between the left hypothalamus and the right temporal pole (Table 2).

Similar findings were found when analysing the subgroup of controls and patients that were studied longitudinally (Supplementary Table 4).

Longitudinal hypothalamic RS FC changes: No significant longitudinal hypothalamic RS FC changes were found in controls. At follow-up versus baseline, the left and right hypothalamus showed increased RS FC with bilateral OFG in migraine patients (Figure 2 and Table 3). Migraine patients also experienced decreased right hypothalamic RS FC with the ipsilateral lingual gyrus (Figure 2 and Table 3). We found no significant time-by-group interaction in migraine patients compared to controls.

Figure 2.

Longitudinal hypothalamic resting state functional connectivity changes in patients with migraine. Areas of increased hypothalamic RS FC are represented in red and areas of decreased hypothalamic RS FC are shown in blue, color-coded for their t values (paired sample t-test, p < 0.001, uncorrected for display). (a) Left hypothalamic RS FC changes; (b) right hypothalamic RS FC changes.

Table 3.

Regions showing significant hypothalamic resting state functional connectivity changes in the whole group of migraine patients over time, as well as in subgroups of patients.

Whole migraine patient group
Increased hypothalamic RS FC at follow-up vs. baseline
Left hypothalamus
Connected regions	Brodmann area	t values	Cluster extent (no of voxels)	MNI coordinates (X, Y, Z)
L orbitofrontal gyrus	11	5.46*	108	−22, 28, −8
R orbitofrontal gyrus	11	4.56**	35	24, 36, −10

Right hypothalamus

Connected regions	Brodmann area	t values**	Cluster extent (no of voxels)	MNI coordinates (X, Y, Z)
L orbitofrontal gyrus	11	4.17	33	−22, 28, −12
R orbitofrontal gyrus	11	3.51	23	24, 32, −10
Decreased hypothalamic RS FC at follow-up vs. baseline
Right hypothalamus
Connected regions	Brodmann area	t values**	Cluster extent (no of voxels)	MNI coordinates (X, Y, Z)
R lingual gyrus	17	4.84	15	4, −72, 4

Patients with improved or stable migraine at follow-up^§
Increased hypothalamic RS FC at follow-up vs. baseline
Left hypothalamus
Connected regions	Brodmann area	t values**	Cluster extent (no of voxels)	MNI coordinates (X, Y, Z)
L orbitofrontal gyrus	11	4.75	7	−24, 24, −6
Patients with worsening migraine at follow-up
Decreased hypothalamic RS FC at follow-up vs. baseline

Left hypothalamus

Connected regions	Brodmann area	t values*	Cluster extent (no of voxels)	MNI coordinates (X, Y, Z)
L orbitofrontal gyrus	11	16.65	77	−12, 16, −24

Right hypothalamus

Connected regions	Brodmann area	t values*	Cluster extent (no of voxels)	MNI coordinates (X, Y, Z)
L orbitofrontal gyrus	11	12.67	61	−14, 16, −24

*Age- and sex-adjusted linear model (p < 0.05, clusterwise FWE-corrected for multiple comparisons).

**Age- and sex-adjusted linear model (p < 0.001, uncorrected).

^§The use of preventive therapies was included as an additional covariate.

L: left; R: right.

Similar results were detected in the exploratory analysis investigating longitudinal hypothalamic RS FC changes in the subgroup of patients who were not taking any preventive treatments at the two time points (Supplementary Table 5).

Effect of disease progression and aura over time: The exploratory subgroup analysis showed an increased left hypothalamic RS FC with the ipsilateral OFG in patients with an improved or stable migraine at follow-up, while patients with worsening migraine experienced decreased RS FC between bilateral hypothalamus and left OFG (Figure 3 and Table 3). We found no significant longitudinal changes in the subgroups of patients with and without aura. No significant time-by-group interactions were observed in patients with improved or stable migraine versus patients with worsening migraine, as well as in MWA versus MWoA patients.

Figure 3.

Longitudinal hypothalamic resting state functional connectivity changes and disease progression in patients with migraine. Longitudinal hypothalamic resting state functional connectivity changes in patients with an improved or stable migraine (a) and patients with worsening migraine (b) (paired sample t-test, p < 0.001, uncorrected for display). Areas of increased hypothalamic RS FC are represented in red and areas of decreased hypothalamic RS FC are shown in blue, color-coded for their t values.

Correlation analysis: At baseline, in migraine patients the lower the right hypothalamic RS FC with the right lingual gyrus, the higher the migraine attack frequency (r = −0.39, p < 0.05, voxelwise FWE-corrected for multiple comparisons). Moreover, the decreased hypothalamic RS FC with the right OFG observed in migraine patients at baseline correlated with greater headache impact at follow-up (left hypothalamus: r = −0.82, p < 0.001, uncorrected; right hypothalamus: r = −0.87, voxelwise FWE-corrected for multiple comparisons). At follow-up, the increased RS FC between the right hypothalamus and the ipsilateral OFG was associated with lower migraine attack frequency (r = −0.73, p < 0.001 uncorrected).

Discussion

The main finding of this study is that episodic migraine patients experience hypothalamic RS FC alterations during the interictal phase according to their clinical features. Using a longitudinal fMRI study design, we showed that the hypothalamic-cortical interplay changes dynamically over time in episodic migraine patients. Changes in the functional coupling between the hypothalamus and pain and visual processing brain regions can influence the course of migraine.

Recent evidence highlighted an important hypothalamic involvement in migraine physiopathology, showing an altered activity of the hypothalamus during all phases of the migraine attack. The hypothalamus is involved in numerous functions including pain modulation, sleep, cognition, autonomic and homeostatic regulation (1). An early activation of the hypothalamus can mediate migraine premonitory symptoms and facilitate the onset of the migraine headache (3,4). It has been shown that the hypothalamus can contribute to migraine chronification and influence the severity of migraine attacks in chronic migraine patients (7,9). Only one study (6) has investigated interictal hypothalamic RS FC alterations in 12 patients with episodic migraine, showing an altered functional coupling between the hypothalamus and brain structures involved in pain and autonomic functions. Consistent with the previous cross-sectional study (6), we found an altered hypothalamic RS FC with brain regions implicated in nociception and migraine pathophysiology, including the cerebellum, parahippocampal, lingual and OFG, in a large sample of interictal episodic migraine patients. The parahippocampus, cerebellar crus II and lobule VI are associated with sympathetic and parasympathetic regulation (6). The altered connection we have found between the hypothalamus and these brain areas at baseline may contribute to explain autonomic nervous system dysfunctions accompanying the migraine attack (20). Our cross-sectional analysis showed only a decreased hypothalamic RS FC, while both decreased and increased hypothalamic RS FC have been previously described in migraine patients (6). This variability among studies could relate to different study designs, statistical approaches and sample size of patients.

A valuable strategy to elucidate the hypothalamic involvement in the course of migraine is to study patients longitudinally. Previous fMRI studies have investigated migraine-phase dependent hypothalamic changes (3,4). Here, we explored longitudinal changes of the hypothalamic connectivity, showing that migraine patients studied interictally developed increased hypothalamic RS FC with frontal nociceptive regions as well as decreased RS FC between the hypothalamus and visual areas after 4 years.

Numerous imaging studies reported a specific hypothalamic activation during the pain phase of primary headache disorders, such as cluster headache and migraine, and showed that the hypothalamus can modulate the activity of brain areas usually implicated in pain transmission and perception (2,5). The hypothalamus is highly connected with the trigeminal cervical complex, as well as brainstem, thalamic and cortical structures involved in nociception (1). Nociceptive inputs originating from the trigeminal cervical complex can converge into the hypothalamus, which mediates autonomic, neuroendocrine, cognitive and behavioural responses to pain through reciprocal connections with cortical and subcortical areas (21). Similar to previous studies (6,9,22) supporting reciprocal connections between the hypothalamus and frontal areas, we found a weak RS FC between the hypothalamus and OFG at baseline, which was strengthened after 4 years. The OFG is part of the descending pain-inhibitory pathway and is involved in the modulation of pain threshold and emotional aspects of pain (23). This frontal area is innervated by hypothalamic orexinergic neurons (9,24,25). The neuropeptides orexins A and B are exclusively produced by the hypothalamus and are involved in autonomic functions, pain modulation, arousal, feeding and sleep regulation (1). The orexinergic neurons seem to modulate patients’ susceptibility to the onset of a migraine attack and the duration of migraine pain (21). In line with previous data in chronic migraine (9), the decreased hypothalamic RS FC with the OFG we have detected in our cross-sectional analysis may lead to an impairment of the antinociceptive system in patients with episodic migraine. Notably, an adaptive coping response that lowers the migraine attack frequency over the years occurred in migraine patients. We showed that the interaction between the hypothalamus and OFG was reinforced after 4 years, and the increased RS FC was associated to a lower migraine attack frequency at follow-up. In support of this hypothesis, we found that the hypothalamic-orbitofrontal connection was strengthened over time in those patients who reported fewer migraine attacks or remained stable, whereas it decreased in patients with worsening migraine.

It also interesting to note that, at baseline, a lower hypothalamic functional interaction with the OFG was related to greater headache impact at follow-up, suggesting that weakened hypothalamic-frontal connections may be a prognostic marker for migraine disability. This finding is in agreement with the notion that functional and structural abnormalities of the OFG may predispose to a more severe form of migraine (22,26).

Visual pathways have been widely studied in MWA and MWoA patients (27,28). It has been suggested that neural mechanisms associated to visual manifestations commonly described by patients with migraine during and between attacks, like light hypersensitivity, can involve a complex brain network including retinal afferents, trigeminal nuclei, hypothalamus, thalamus, and visual processing cortical areas (1,29). Recent studies demonstrated direct connections of the hypothalamus with visual cortical areas, with possible implications in visuospatial processing, regulating the circadian rhythm and encoding the salience of multisensory information (9,30). Both the hypothalamus and visual cortex show an abnormal activation across all phases of the migraine attack (3,4,27). Previous research (9,22) has found abnormal hypothalamic functional and structural connectivity with occipital brain regions in patients with episodic and chronic migraine, highlighting the role of visual areas in migraine pathophysiology and its chronification via direct and indirect connections between the visual, thalamic, hypothalamic and trigeminovascular pathways. Here, we showed a reduced RS FC between the right hypothalamus and ipsilateral lingual gyrus that further decreased over time. Functional and structural alterations of the lingual gyrus, an extrastriate visual area involved in higher order visual functions, such as visual memory and perception of colour, have been widely described in migraine patients (27,31). Interestingly, in our cohort of patients a lower functional interaction between the hypothalamus and lingual gyrus was related to higher migraine attack frequency at baseline, suggesting that an aberrant interaction between these two regions during the interictal phase might facilitate the onset of a migraine attack. These findings further corroborate the interplay between the hypothalamus, visual and trigeminovascular systems.

It is also worth noting that we found functional alterations between the hypothalamus and visual areas regardless of the presence of migraine with aura. The only difference we have observed between patients with and without aura was a lower hypothalamic connectivity with the right temporal pole, a nociceptive processing brain area involved in migraine pathophysiology (32), in patients without aura. Cortical spreading depression (CSD) is the most widely accepted neurophysiological mechanism underlying migraine aura. It is characterised by a wave of neuronal depolarisation followed by prolonged inhibition that originates from striate and extrastriate visual areas (33). Recent experimental studies showed that the CSD can also activate the thalamus and trigeminovascular system, suggesting an involvement of the CSD also in migraine without aura (34). Which is the role of CSD in the onset of migraine pain and whether “silent CSD attacks” originating from the visual cortex can activate the trigeminovascular system or vice versa is still unclear (27,28). Consistent with previous studies showing functional and structural alterations in visual processing areas regardless of the presence of migraine aura, our results support the notion that migraine patients with and without aura may share common pathogenic mechanisms (2,35).

It is noteworthy that cross-sectional and longitudinal hypothalamic RS FC changes we have observed here were not influenced by the use of migraine preventive medications at baseline and follow-up. Based on these findings we could suppose that changes in the hypothalamic connectivity we have observed here and their clinical correlates reflect the natural history of the disease. Future studies including drug-naïve patients are necessary to support this conclusion.

A few limitations of the study should be noted. The sample size of the migraine subgroups was quite small and the lack of a strict control of false positives using an uncorrected statistical threshold in the comparison between subgroups should be considered. Moreover, information regarding the time elapsed between the MRI and the following migraine attack was missing, as well as data concerning the headache pain severity and patients’ disability at baseline. Although all statistical models were controlled for age, the group of patients was older than that of controls. In addition, we did not have reliable pieces of information concerning female hormonal changes over the years that could have influenced the hypothalamic function. At last, the duration of follow-up was shorter compared to the migraine disease duration. Further larger studies with a longer follow-up are needed to confirm our results. In addition, future studies should investigate longitudinal hypothalamic connectivity changes in patients at an early onset of the disease, like paediatric patients, as well as in chronic migraine patients.

Overall, our results showed that the hypothalamus modulates the activity of key brain areas implicated in migraine physiopathology, thus affecting the course of the disease. As such, we highlighted a possible target for novel pharmacological and neuromodulation approaches.

Article highlights

The functional coupling between the hypothalamus and brain regions implicated in pain and visual processing changes dynamically over time in episodic migraine patients.

Changes in the hypothalamic-cortical interplay can influence the course of migraine.

Weakened hypothalamic-orbitofrontal gyrus connections may be a prognostic marker for migraine disability.

Supplemental Material

sj-pdf-1-cep-10.1177_03331024211046618 - Supplemental material for Clinical correlates of hypothalamic functional changes in migraine patients

Supplemental material, sj-pdf-1-cep-10.1177_03331024211046618 for Clinical correlates of hypothalamic functional changes in migraine patients by Roberta Messina, Maria A Rocca, Paola Valsasina, Paolo Misci and Massimo Filippi in Cephalalgia

Footnotes

Author contributions

RM: design of the work, patient enrolment, data analysis, statistical analysis and drafting/revising the work. MAR: drafting/revising the work and data interpretation. PV: analysis of the data. PM: MRI analysis. MF: design of the work, drafting/revising the work and data interpretation. He also acted as study supervisor. All the authors approved the current version of the manuscript.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Roberta Messina

References

Goadsby

Holland

Martins-Oliveira

, et al. Pathophysiology of migraine: A disorder of sensory processing. Physiol Rev 2017; 97: 553–622.

Messina

Filippi

Goadsby

PJ.

Recent advances in headache neuroimaging. Curr Opin Neurol 2018; 31: 379–385.

Schulte

May

The migraine generator revisited: Continuous scanning of the migraine cycle over 30 days and three spontaneous attacks. Brain 2016; 139: 1987–1993.

Maniyar

Sprenger

Monteith

, et al. Brain activations in the premonitory phase of nitroglycerin-triggered migraine attacks. Brain 2014; 137: 232–241.

Denuelle

Fabre

Payoux

, et al. Hypothalamic activation in spontaneous migraine attacks. Headache 2007; 47: 1418–1426.

Moulton

Becerra

Johnson

, et al. Altered hypothalamic functional connectivity with autonomic circuits and the locus coeruleus in migraine. PLoS One 2014; 9: e95508.

Schulte

Allers

May

Hypothalamus as a mediator of chronic migraine: Evidence from high-resolution fMRI. Neurology 2017; 88: 2011–2016.

Lee

Park

Cho

, et al. Increased connectivity of pain matrix in chronic migraine: A resting-state functional MRI study. J Headache Pain 2019; 20: 29.

Coppola

Di Renzo

Petolicchio

, et al. Increased neural connectivity between the hypothalamus and cortical resting-state functional networks in chronic migraine. J Neurol 2020; 267: 185–191.

10.

Messina

Rocca

Colombo

, et al. Gray matter volume modifications in migraine: A cross-sectional and longitudinal study. Neurology 2018; 91: e280–e292.

11.

Headache Classification Committee of the International Headache Society. The International Classification of Headache Disorders, 3rd edition (beta version). Cephalalgia 2013; 33: 629–808.

12.

Headache Classification Subcommittee of the International Headache Society. The International Classification of Headache Disorders, 2nd edition. Cephalalgia 2004; 24: 9–160.

13.

Lines

Vandormael

Malbecq

A comparison of visual analog scale and categorical ratings of headache pain in a randomized controlled clinical trial with migraine patients. Pain 2001; 93: 185–190.

14.

Kosinski

Bayliss

Bjorner

, et al. A six-item short-form survey for measuring headache impact: The HIT-6. Qual Life Res 2003; 12: 963–974.

15.

Stewart

Lipton

Dowson

, et al. Development and testing of the Migraine Disability Assessment (MIDAS) Questionnaire to assess headache-related disability. Neurology 2001; 56: S20–S28.

16.

Biswal

Yetkin

Haughton

, et al. Functional connectivity in the motor cortex of resting human brain using echo-planar MRI. Mag Res Med 1995; 34: 537–541.

17.

Lowe

Mock

Sorenson

JA.

Functional connectivity in single and multislice echoplanar imaging using resting-state fluctuations. Neuroimage 1998; 7: 119–132.

18.

May

Pearls and pitfalls: Neuroimaging in headache. Cephalalgia 2013; 33: 554–565.

19.

Hidalgo de la Cruz

Valsasina

Mesaros

, et al. Clinical predictivity of thalamic sub-regional connectivity in clinically isolated syndrome: A 7-year study. Mol Psychiatry 2020; 26: 2163–2174.

20.

Gazerani

Cairns

BE.

Dysautonomia in the pathogenesis of migraine. Expert Rev Neurother 2018; 18: 153–165.

21.

May

Burstein

Hypothalamic regulation of headache and migraine. Cephalalgia 2019; 39: 1710–1719.

22.

Chong

Aguilar

Schwedt

TJ.

Altered hypothalamic region covariance in migraine and cluster headache: A structural MRI study. Headache 2020; 60: 553–563.

23.

Petrovic

Kalso

Petersson

, et al. Placebo and opioid analgesia – imaging a shared neuronal network. Science 2002; 295: 1737–1740.

24.

Karimi

Hamidi

Fatahi

, et al. Orexin 1 receptors in the anterior cingulate and orbitofrontal cortex regulate cost and benefit decision-making. Prog Neuro-Psychoph Biol Psych 2019; 89: 227–235.

25.

Saper

CB.

Hypothalamic connections with the cerebral cortex.

Prog Brain Res 2000; 126: 39–48.

26.

Schwedt

Chong

CD.

Medication overuse headache: Pathophysiological insights from structural and functional brain MRI research. Headache 2017; 57: 1173–1178.

27.

Puledda

Ffytche

O’Daly

, et al. Imaging the visual network in the migraine spectrum. Front Neurol 2019; 10: 1325.

28.

Burke

Joutsa

Cohen

, et al. Mapping migraine to a common brain network. Brain 2020; 143: 541–553.

29.

Schulte

Peng

KP.

Current understanding of premonitory networks in migraine: A window to attack generation. Cephalalgia 2019; 39: 1720–1727.

30.

Kamali

Ghazi Sherbaf

Rahmani

, et al. A direct visuosensory cortical connectivity of the human limbic system. Dissecting the trajectory of the parieto-occipito-hypothalamic tract in the human brain using diffusion weighted tractography. Neurosci Lett 2020; 728: 134955.

31.

Messina

Rocca

Colombo

, et al. Cortical abnormalities in patients with migraine: A surface-based analysis. Radiology 2013; 268: 170–180.

32.

Schwedt

Berisha

Chong

CD.

Temporal lobe cortical thickness correlations differentiate the migraine brain from the healthy brain. PLoS One 2015; 10: e0116687.

33.

Charles

Baca

SM.

Cortical spreading depression and migraine. Nat Rev Neurol 9: 637–644.

34.

Noseda

Burstein

Migraine pathophysiology: Anatomy of the trigeminovascular pathway and associated neurological symptoms, CSD, sensitization and modulation of pain. Pain 2013; 154: S44–S53.

35.

Zielman

Wijnen

Webb

, et al. Cortical glutamate in migraine. Brain 2017; 140: 1859–1871.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.35 MB