Cortical morphological changes in chronic migraine in a Taiwanese cohort: Surface- and voxel-based analyses

Abstract

Background

Previous voxel- or surface-based morphometric analysis studies have revealed alterations in cortical structure in patients with chronic migraine, yet with inconsistent results. The discrepancies may be derived partly from the sample heterogeneity. Employing both methods in a clinically homogenous group may provide a clearer view.

Methods

Structural MRI data from 30 prevention-naïve patients with chronic migraine without medication overuse headache or a history of major depression and 30 healthy controls were analyzed. Vertex-wise (surface-based) or voxel-wise (voxel-based) linear models were applied, after controlling for age and gender, to investigate between-group differences. Averaged cortical thicknesses and volumes from regions showing group differences were correlated with parameters related to clinical profiles.

Results

Surface-based morphometry showed significantly thinner cortices in the bilateral insular cortex, caudal middle frontal gyrus, precentral gyrus, and parietal lobes in patients with chronic migraine relative to healthy controls. Additionally, the number of migraine days in the month preceding MRI examination was correlated negatively with right insular cortical thickness. Voxel-based morphometry (VBM) did not show any group differences or clinical correlations.

Conclusion

Patients with chronic migraine without medication overuse headache, major depression, or prior preventive treatment had reduced cortical thickness in regions within the pain-processing network. Compared to voxel-based morphometry, surface-based morphometry analysis may be more sensitive to subtle structural differences between healthy controls and patients with chronic migraine.

Keywords

Chronic migraine cortical thickness surface-based morphometry voxel-based morphometry

Introduction

Migraine is a leading cause of disability worldwide, with a prevalence of 10–15% (1). Epidemiological studies have shown that about 2% of the general population suffers from chronic migraine (CM) (2), with a 3% (2–4%) annual conversion rate from episodic migraine (EM) to CM (3).

The pathophysiology of migraine is far from being understood. Using structural neuroimaging methods; for example, surface-based morphometry (SBM) or voxel-based morphometry (VBM), researchers have found that, in EM, specific regions within the pain matrix exhibit alterations in cortical thickness or volume. These regions include the somatosensory cortex (4 –6), motor cortex (7,8), frontal/prefrontal cortices (7 –12), visual cortex (7,13), insula (6,11), and anterior/middle cingulate cortex (6,8,11), among others. However, the direction of change was inconsistent across these studies. For example, although some studies reported a thicker cortex in EM over somatosensory (4 –6), visual (7,13), or frontal regions (5,7) compared with healthy controls (HCs), other studies reported no differences (14 –16) or thinner cortex (6,7,17). Both SBM and VBM results have been conflicting. Differences in methodologies and clinical characteristics of the patients may partly explain the discrepancies.

Compared to EM, CM is characterized by an increased disease burden and could thus be associated with more widespread brain structural changes. Few studies have addressed this important issue with SBM (18) or VBM (19 –21), with the latter yielding inconsistent results. The only study (18) to investigate cortical thickness in CM did not find any significant differences relative to HCs; however, one third (10/30) of their study patients were comorbid with medication overuse headache (MOH) (22). As CM with and without MOH to some extent exhibit different morphological alterations (19), combining the two groups may obscure subtle changes in either group (23).

Although both SBM and VBM provide measures of brain structure, the two methods provide complementary rather than redundant information (5,6,16,24). As depicted in Figure 1, VBM measures gray matter (GM) volume, while SBM is capable of probing various aspects of cortical shape including cortical thickness, degree of gyral convolution (i.e. the gyrification index), and sulcal depth, among others. Thus, combining SBM and VBM may provide further insight into how brain structures are affected in CM.

Figure 1.

Basic workflow of VBM and SBM. After segmentation, VBM measures GM volume, while SBM can determine various aspects of cortical shape, including cortical thickness, degree of gyral convolution (the gyrification index), and sulcal depth, after applying different strategies in registration to the templates.

In the present study we aimed to evaluate brain structural changes using both SBM and VBM in a clinically homogeneous group of CM patients who were without MOH or major depression and naïve to preventive medications. We examined two hypotheses. Our first hypothesis was that these patients would show alterations in SBM or VBM metrics in pain-related brain regions. Our second hypothesis was that some of these alterations would correlate with clinical parameters, such as disease duration, CM severity, or headache frequency.

Methods

Subjects

Patients with CM were recruited from the Department of Neurology at Taipei Veterans General Hospital. At their first visit, they completed a structured headache questionnaire consisting of headache profiles and assessment of disease severity (Migraine Disability Assessment [MIDAS]) and screening for anxiety and depression (Hospital Anxiety and Depression Scale [HADS]). Information on other common pain syndromes (e.g. widespread pain index as in fibromyalgia, arthralgia, toothache, and back pain), as well as the duration and severity of the pain, were collected. The inclusion criteria for CM patients were: Aged 18–65 years; fulfilling the ICHD-3 criteria for CM (22); and migraine onset before the age of 50 years. The exclusion criteria were: Headache type other than CM; comorbid MOH as defined by the ICHD-3 criteria (22); history of major depression; use of migraine prophylaxis before enrolment; CM onset after the age of 60 years; comorbidity with other pain syndrome for more than 3 months, being pregnant or nursing; having a history of renal or hepatic disorder, or other severe systemic disease; and having a condition incompatible with MRI, such as metallic or electric implants or claustrophobia. After the initial visit, eligible patients who met the ICHD-3 criteria for CM were asked to keep headache diaries during a 4-week prospective baseline period. Subjects with ≥15 monthly headache days (MHD) and ≥8 monthly migraine days (MMD) within this period then underwent brain MRI examination, with structural, functional, and spectroscopic measurements. Age and gender-matched HCs without a history of migraine were also enrolled. Infrequent tension-type headache (TTH) (<1/month) (22) was allowed in the HC group, in line with the inclusion criteria for HC in the field of migraine research (5,11,14,16). Apart from a diagnosis of CM in the patient group, all participants had normal physical and neurological examinations.

Prior to entering the study, all participants were informed about the study orally and in writing, and gave informed written consent to participate. The institutional review board of Taipei Veterans General Hospital approved the protocol (VGHIRB: 2012-04-046B).

MRI acquisition

The study was conducted on a 3-T magnetic resonance scanner (Trio, Siemens Medical Solutions, Erlangen, Germany) located at National Yang-Ming University. Images were acquired with a 32-channel head coil array. The participants’ heads were immobilized in the scanner by foam pads. The structural scan consisted of a high-resolution three-dimensional MPRAGE (Magnetization Prepared Rapid Acquisition Gradient Echo) sequence with the following parameters: Repetition time, 2530 ms; echo time, 3.03 ms; flip angle, 7 degrees; field of view, 224 × 256 × 192; and voxel size, 1 × 1 × 1 mm.

SBM

The SBM analysis was carried out with the Computational Anatomy Toolbox (CAT12, http://www.neuro.uni-jena.de/cat/), an extension for the Statistical Parametric Mapping 12 (SPM12; https://www.fil.ion.ucl.ac.uk/spm/software/spm12/) running in the MATLAB environment (R2017b; MathWorks, Natick, MA). The analysis was semi-automated by applying default parameters for all processing steps, as described by Niddam et al. (25). The high-resolution T1-images were initially segmented into different tissue types, including GM, white matter (WM), and cerebrospinal fluid (CSF). The pial surface and the WM surface (GM/WM boundary) were extracted, and an outer hull layer that tightly wraps the pial surface was reconstructed (Figure 1). Cortical thickness was estimated by calculating the distance between the pial and WM surfaces. Two additional geometric cortical measures (e.g. sulcal depth and degree of gyrification [gyrification index = $\frac{area o f l o cal pial surface}{area o f local hull surface}$ ]) were also derived (Figure 1). Surface reconstruction was then conducted in right and left hemispheres separately, using the projection-based thickness approach. This step provides topological correction for defects in the surface mesh as well as for spherical inflation and spherical registration (26 –28). Application of an adapted DARTEL (Diffeomorphic Anatomical Registration Through Exponentiated Lie Algebra) algorithm during spherical registration enabled inter-subject analysis by mapping the matrix onto a standardized spherical surface. Right and left hemispheres were then merged into a single mesh, resampled to a template space, and spatially smoothed. A 15-mm FWHM (full-width at half maximum) Gaussian kernel was used for smoothing the cortical thickness images, while a 20-mm FWHM Gaussian kernel was used for the gyrification index and sulcal depth images as recommended. The smoothed images were used for between-group analyses, as described below.

VBM

The VBM analysis was conducted in SPM 12, and followed the procedures described by Niddam et al. (29). In short, the high-resolution T1-images were segmented into different tissue types, including GM, WM, and CSF. An optimized GM template in standard space was created with the DARTEL toolbox. Registration parameters and deformation maps generated during this step were then used for spatial warping of each individual’s GM images. During this step, the GM volume was preserved to ensure that GM volumes would be compared rather than GM concentrations. Voxels with a GM concentration less than 0.2 were excluded from the analysis to avoid potential boundary effects between GM and WM, as well as to preserve homogeneity. Finally, a 4‐mm³ FWHM Gaussian kernel was used for spatial smoothing of the GM images. The smoothed images were used for between-group analyses.

Statistical analysis of structural images

Between-group differences in cortical thickness, gyrification index, and sulcal depth (by SBM) and GM volume (by VBM) were assessed with two-sample t-tests controlling for age and gender. Total intracranial volume (TIV) was also controlled for in the VBM models as recommended (29). Statistical maps were thresholded first at a voxel-level at p < 0.005 and then at a cluster-level at p < 0.05, corrected for the false discovery rate.

To label significant clusters with the corresponding anatomical structures, Montreal Neurological Institute (MNI) coordinates were first transformed into Talairach space using an MNI to Talairach Converter (https://bioimagesuiteweb.github.io/webapp/mni2tal.html), and then entered into the Talairach Client (http://www.talairach.org/client.html).

Given that previous conflicting results on brain structural changes in patients with migraine may have been due to different approaches, two additional analyses were performed. Firstly, threshold-free cluster enhancement (TFCE), a relatively liberal statistical method, was performed with a cluster-wise threshold at p < 0.05 and cluster size threshold > 50 voxels (30). Secondly, a region-of-interest (ROI) analysis was performed. Anatomical loci exhibiting between-group differences in one structural analysis, but not the other, were examined using an ROI analysis for the latter. Because a two-dimensional cluster derived from SBM cannot be transformed directly into a three-dimensional mask for VBM, and vice versa, an atlas-based ROI approach was used. For example, significant clusters in the SBM analysis were labeled using the IBASPM 116 atlas for VBM analysis. Two-sample t-tests were then applied to examine between-group (CM vs. HC) differences.

Statistical analysis of clinical parameters

Statistical analyses of clinical variables were conducted in SPSS software (version 21 for Windows; IBM Inc., Armonk, NY). Differences between groups were examined with independent t-tests for continuous data and with either the chi-squared test or Fisher’s exact test for nominal data, where appropriate. Statistical significance was defined as p < 0.05. For the correlation analysis between mean cortical thickness/volume, gyrification index, or sulcal depth from the clusters derived from SBM or VBM and clinical parameters, tests were performed for headache characteristics (i.e. migraine disease duration, CM disease duration, MIDAS score, and MHD/MMD at baseline). Statistical significance was defined as p < 0.01 (0.05/5), accounting for the multiple comparisons performed. All tests were two-tailed.

Results

Demographics and baseline clinical characteristics

Thirty patients with CM and 30 HCs were recruited for this study from June 2012 to December 2014. All participants were Taiwanese (East Asians). The demographic characteristics of the participants in each group are summarized in Table 1. The CM patient and HC groups were gender-matched (p = 0.766) and did not differ significantly in age (p = 0.723). Although infrequent TTH (n = 7) was allowed in the HC group, the majority were headache-free subjects (n = 23). CM patients suffered from a severe disease burden with a mean MIDAS score of 45.6 ± 50.7. None of the CM patients were overusing medications, had a history of or had taken medications for major depression. Their scores in the HAD scale were moderate with a mean HADS-A score of 7.3 ± 3.0, and HADS-D score of 5.4 ± 3.8. No structural lesions were found on the high-resolution T1-weighted MRI in either group. The mean volumes of GM, WM, CSF, brain (GM + WM), and TIV (GM + WM + CSF) did not differ significantly between the CM and HC groups (all p > 0.05).

Table 1.

Demographics of healthy controls and patients with chronic migraine.

	CM (n = 30)	HC (n = 30)	p
Age (years)	33.2 (9.8)	32.4 (8.3)	0.723
Gender, males/females	7/23	8/22	0.766
Migraine duration (years)	13.2 (8.8)	–	–
CM duration (years)	3.6 (4.4)	–	–
MIDAS	45.6 (50.7)	–	–
Headache frequency at M0 (/month)	24.0 (5.3)	–	–
Migraine frequency at M0 (/month)	16.5 (8.3)	–	–
Mean days of acute abortive medication at M0 (/month)	1.9 (2.8)	–	–
Mean acute abortive medication tablet taken at M0 (/month)	2.5 (4.4)	–	–
Types of acute abortive medication (nil/acetaminophen/NSAID/triptan)	12/5/8/5	–	–
HADS-A	7.3 (3.0)	–	–
HADS-D	5.4 (3.8)	–	–
HADS-T	12.7 (5.2)	–	–
GM volume (cm³)	677.1 (60.3)	690.3 (53.5)	0.371
WM volume (cm³)	497.6 (47.6)	499.9 (38.7)	0.839
CSF volume (cm³)	292.8 (27.4)	300.8 (34.7)	0.327
Brain volume (cm³)	1174.7 (104.6)	1190.3 (89.8)	0.539
Total intracranial volume (cm³)	1467.5 (127.7)	1491.0 (120.5)	0.466

Note: Data are shown as mean (SD).

CM: chronic migraine; CSF: cerebrospinal fluid; GM: gray matter; HADS-A/D/T: Hospital Anxiety and Depression Score, anxiety/depression/total score; HC: healthy control; M0: the month preceding MRI scan; MIDAS: migraine disability assessment scale; NSAID: non-steroidal anti-inflammatory drugs; WM: white matter.

SBM analysis

Comparisons of surface metrics between the CM and HC groups revealed significant between-group regional differences in cortical thickness (Figure 2), but not in sulcal depth or gyrification index. Compared to HCs, reduced cortical thickness was found in the CM group in clusters located in bilateral precentral gyri, caudal middle frontal gyri (MFG), insula, and parietal lobes (Tables 2 and 3). No clusters of increased cortical thickness were observed in the CM group relative to HCs. Correlation analyses of mean cortical thicknesses (extracted from significant clusters) with headache parameters in CM patients revealed a negative correlation between mean cortical thickness within the right insula and MMD (n = 30, r = −0.470, p = 0.009) (Figure 3).

Figure 2.

Cortical thickness abnormalities in patients with CM. Reduced cortical thickness in CM patients was observed in the bilateral precentral gyri, bilateral middle frontal gyri, bilateral insula, and bilateral parietal lobes (corrected p [FDR] < 0.05 at cluster-level). The color bar indicates the voxel-wise T-value.

Table 2.

Significant clusters exhibiting reduced cortical thickness in patients with CM compared with HCs.

Anatomical region	BA	p (FDR)	Cluster size (vertices)	Cluster size (mm²)	MNI coordinates (x, y, z)
LT precentral G.	4	<0.001	785	1688	−48	−11	49
LT middle frontal G.	6				−31	−8	45
RT middle frontal G.	6	0.047	214	460	30	−10	50
RT inferior parietal L.	40	0.002	401	862	52	−41	44
RT insula	13	0.002	374	804	38	−15	13
LT parietal L.	7	0.028	247	531	−28	−67	38
RT precentral G.	4	0.028	244	525	33	−20	45
LT insula	13	0.002	388	834	−30	22	6

Note: p-values corrected at cluster level; statistical maps initially thresholded at uncorrected voxel level of p < 0.005.

BA: Brodmann area; FDR: false discovery rate; G: gyrus; L: lobe/lobule; LT: left; MNI: Montreal Neurological Institute; RT: right.

Table 3.

Summary of findings from voxel- or surface-based morphometric studies.

Method	Parameters		Statistic approach	HC > CM	HC < CM
		${$	Voxel p < 0.005, FDR corrected	–	–
VBM	GM volume	${$	TFCE p < 0.05 (liberal approach)	–	–
			ROI based (liberal approach)	–	–
	Cortical thickness	$}$		v	–
SBM	Gyrification index	$}$	Voxel p < 0.005, FDR corrected	–	–
	Sulcal depth			–	–

CM: chronic migraine; FDR: false discovery rate; GM: gray matter; HC: healthy control; ROI: region-of-interest; SBM: surface-based morphometry; TFCE: threshold-free cluster enhancement; VBM: voxel-based morphometry.

Figure 3.

Association between mean cortical thickness and baseline migraine frequency. A negative correlation was observed between the mean cortical thickness of the right insula and monthly migraine frequency at baseline (n = 30, r = −0.470, p = 0.009). The cluster from which cortical thickness was obtained is encircled on the brain insert.

VBM analysis

Voxel-wise whole brain comparisons showed no significant differences in GM volume between the CM and HC groups (Table 3). No clusters passed the more liberal TFCE threshold. Finally, between-group comparisons of mean GM volumes in regions with corresponding significant clusters in the SBM analysis also failed to reveal any significant differences.

Analysis in women only

To further increase the specificity of the study cohort, an additional analysis restricted to women was performed (CM: n = 23; HC: n = 22). They were age-matched (p = 0.856), and the demographics of these subjects are provided in Supplemental Table 1. Considering the lower statistical power due to the reduction in case numbers, two different statistical criteria were applied (voxel-level at p < 0.005 followed by a cluster-level at p < 0.05, or voxel-level at p < 0.01 followed by a cluster size > 200). The results from the latter approach were similar to those derived from the whole study sample (Supplemental Figure 1, Supplemental Table 2).

Discussion

In the present study, we investigated cortical morphological changes using both SBM and VBM in patients with CM not comorbid with MOH or major depression, and not on any preventive medications. Furthermore, all subjects included in the study had a relatively homogenous demographic background (i.e. all Taiwanese). In contrast to our negative VBM results, SBM showed regional cortical thinning in CM patients, compared to HCs, in the bilateral insular cortex, caudal MFG, precentral gyrus, and parietal lobes. These regions are considered parts of the pain matrix and are involved in pain perception, evaluation, and modulation (31,32). In addition, we observed a correlation between the cortical thickness of the right insula and clinical severity.

SBM might be more sensitive than VBM

Although VBM and SBM adopt similar preprocessing steps (e.g. GM/WM segmentation), the two methods use different strategies in aligning each individual’s image to the template (Figure 1). VBM applies voxel- (volume-) based registration to compensate for individual variability, which is relatively prone to registration artifacts induced by global volume change (15). The highly variable folding pattern across individuals may further reduce the accuracy of alignment in VBM (33). SBM, in contrast, adopts an alternative approach for registration by matching the gyral and sulcal geometry to an inflated spherical atlas (5,15,24). Such an approach greatly reduces the potential misalignment caused by complex folding pattern, as well as global volume change. Moreover, SBM is capable of measuring different important cortical features, such as cortical thickness, sulcal depth, gyrification index, and surface area among others, while VBM only provides a measure of GM volume, which may depend on some of the surface derived characteristics (5).

In this study, between-group differences were found with SBM (cortical thickness), but not VBM (GM volume). A similar result was reported recently in a study applying both methods in patients with EM (9). Both studies thus lend support to the above-mentioned notion that SBM might be more accurate or sensitive than VBM in detecting morphological changes within the brain. Of note, in our SBM analysis, only cortical thickness, but not gyrification index or sulcal depth, showed between-group differences. The absence of changes in the latter features may suggest that no overt atrophic changes occur in patients with migraine, as alteration in gyrification index or sulcal depth are frequently seen in neurodevelopmental (34) or neurodegenerative diseases (35).

The observation that cortical thinning was not associated with a corresponding reduction in GM volume also deserves further discussion. The superior spatial registration used in SBM may in part explain this difference. In addition, although GM volume is assumed to be influenced by both surface area and thickness, there are indications that GM volume depends more on the former than the latter (24). It can, however, not be ruled out that an enlarged surface area developed in patients with CM to compensate for the reduction in cortical thickness, and resulted in a preserved GM volume. Nevertheless, further studies are needed to clarify this point.

Comparison to previous morphological studies in CM and EM

Our results in CM patients are partly in accordance with previous studies of CM, as well as those of EM (see summary in Table 4). For example, in another recent VBM study of CM, no between-group differences were observed (21). Nevertheless, two other VBM studies of CM demonstrated regional differences in the precentral gyrus, prefrontal cortex, temporal lobe, and amygdala (19,20), and one SBM study failed to show between-group differences (18). The causes of the discrepancies may partly be attributed to differences in comorbidities, disease burden, MRI parameters, and analytic approaches (36). For instance, CM with MOH is known to differ from CM without MOH with respect to cortical alterations (19). Of note, 33–71% of the CM patients were comorbid with MOH in the aforementioned previous studies (18,20). Regarding disease burden, the mean MIDAS score in our patient cohort (45.6 ± 50.7) was lower compared to the others (ranging from 46.2–67.3) (18,20,37,38), which could be explained again by the absence of MOH (37) or comorbid depression (38) as both were associated with higher MIDAS scores. On the other hand, the difference in the inclusion criteria for HC (i.e. the allowance of infrequent TTH (5,11,14,16) vs. completely headache-free (18 –21)) may also partly contribute to the inconsistent results.

Table 4.

Summary of brain morphological studies in patients with migraine.

Note: Most commonly involved regions, including those found in the current study, are marked in rectangles.

ANOVA: analysis of variance; A/MCC: anterior/middle cingulate cortex; CM: chronic migraine; CS: central sulcus; EM: episodic migraine; FDR: false discovery rate; FWE: family-wise error; HF: high-frequency migraine; k: number of voxel; LF: low-frequency migraine; MA: migraine with aura; MO: migraine without aura; MOH: medication-overuse headache; PFC: prefrontal cortex; Post/PreCG: post-/pre-central gyrus; s: SBM-based analysis; SBM: surface-based morphometry; SMC: sensorimotor cortex; S/M/IFG: superior/middle/inferior frontal gyrus; S/M/ITG: superior/middle/inferior temporal gyrus; TFCE: threshold free cluster enhancement; TP: temporal pole; v: VBM-based analysis; VBM: voxel-based morphometry; – : no between-group differences; #: meta-analysis, †: region-of-interest (ROI)-based analysis.

Interestingly, the CM results reported here are comparable to those derived from SBM studies, as well as from meta-analyses involving VBM studies, in EM (39 –41) (see summary in Table 4). This finding may imply that SBM is capable of detecting subtle structural differences that are only apparent with VBM using a much larger study cohort.

Comparison to previous morphological studies in other chronic pain syndromes

It should be emphasized that the regions demonstrating between-group differences in Table 4 are not specific to EM/CM. Previous studies on non-migraine headache disorders and other chronic pain syndromes also showed alterations in these regions; that is, the pain matrix (42,43). The observation that nearly all studies investigating brain alterations in patients with various types of chronic pain disorders yield comparable results suggests that these changes are highly correlated with the presence of chronic pain and may either result from repeated input or an inherited susceptibility to develop chronic pain (42). As insular, premotor cortex, and caudal MFG were affected in this study, it could be that the affective-motivation and motor components (43) were more selectively involved in our patient population. Nevertheless, further large-scale studies or meta-analyses using SBM may help delineate the extent of involvement of the pain matrix in patients with CM.

Clinical implication from correlation analysis

With respect to the observations of cortical thinning, it is important to clarify whether such changes are causal to or consequent of chronic pain. The observation that reduced cortical thickness within the right insular cortex correlated with baseline migraine frequency suggests that the alteration within this region may result from the impact of repetitive high-frequency attacks. This supposition is supported by a recent study showing decreased GM volume in the insular cortex in high-frequency EM patients (8–14 days/month) with respect to both HC and low-frequency EM patients (1–2 days/month) (6). The fact that effective treatment of some forms of chronic pain results in a reversal of localized GM reductions (e.g. within the insular cortex, anterior cingulate cortex, and dorsolateral prefrontal cortex) further strengthens this viewpoint (44).

It is worth noting that most of the regions showing between-group differences did not correlate with clinical parameters. A recent SBM study in a large cohort of patients with EM also showed similar findings (17). This could mean that the observed alterations may be inherent to the migraine brain, and might potentially be a predisposing factor. This interpretation is supported by a recent SBM study in twins with migraine with aura, demonstrating morphological changes as being a trait rather than a consequence of the disease (45).

Limitations

Our study has limitations. First, the number of subjects was relatively small and hence may not fully reflect group differences. Further large-scale studies may be needed to demonstrate cortical thickness alterations not detected here. Second, because we recruited a relatively homogenous group of patients with CM, our results may not be generalizable to the broader CM population, of whom 50% have comorbid MOH and 60% of whom have major depression. Third, owing to the limitation of the software (CAT12), surface area was not calculated. The speculation that patients with CM had enlarged surface area to compensate for cortical thinning requires further studies. Fourth, since this is a cross-sectional study, it is difficult to determine whether clusters showing between-group differences, yet without clinical correlation, are predisposing factors for the development of CM. Further studies using a longitudinal follow-up design are needed. Finally, given the inherent limitation, SBM is not capable of analyzing subcortical regions (46,47), including the hypothalamus and the brainstem, both of which have been recognized recently to play crucial roles in migraine pathogenesis (48 –51). The dynamic alterations of functional connectivity between the hypothalamus and some specific nuclei within the brainstem (e.g. the dorsal pons, and the spinal trigeminal nucleus) have been shown to be responsible for the cyclic nature of migraine attacks (49,50). Also, the hyper-responsiveness to painful trigeminal stimulation in the anterior part of the hypothalamus is involved in migraine chronification (51). All these findings were demonstrated by a newly developed high-resolution functional MRI paradigm focusing on brainstem structure (52). In the coming era of ultra-high field MRI, combined analysis of both high-resolution structural and functional images may provide more information regarding the regions involved in the pathogenesis of CM.

Conclusions

In summary, our study showed that in patients with CM, specific regions located within pain-processing networks had reduced cortical thickness compared to HCs. SBM appears to be more sensitive than VBM for studying subtle structural differences in cortical regions. Further longitudinal studies aimed at examining cortical thickness within these regions after successful treatment are warranted.

Clinical implications

Patients with chronic migraine without MOH, major depression or prior prophylactic treatment had reduced cortical thickness within regions related to pain process, which could be a consequence of or a predisposing factor to chronic pain.

Surface-based morphometry analysis appears to be more sensitive than using voxel-based morphometry for studying subtle structural differences in cortical regions.

Supplemental Material

sj-pdf-1-cep-10.1177_0333102420920005 - Supplemental material for Cortical morphological changes in chronic migraine in a Taiwanese cohort: Surface- and voxel-based analyses

Supplemental material, sj-pdf-1-cep-10.1177_0333102420920005 for Cortical morphological changes in chronic migraine in a Taiwanese cohort: Surface- and voxel-based analyses by Kuan-Lin Lai, David M Niddam, Jong-Ling Fuh, Wei-Ta Chen, Jaw-Ching Wu and Shuu-Jiun Wang in Cephalalgia

Supplemental Material

sj-pdf-2-cep-10.1177_0333102420920005 - Supplemental material for Cortical morphological changes in chronic migraine in a Taiwanese cohort: Surface- and voxel-based analyses

Supplemental material, sj-pdf-2-cep-10.1177_0333102420920005 for Cortical morphological changes in chronic migraine in a Taiwanese cohort: Surface- and voxel-based analyses by Kuan-Lin Lai, David M Niddam, Jong-Ling Fuh, Wei-Ta Chen, Jaw-Ching Wu and Shuu-Jiun Wang in Cephalalgia

Supplemental Material

sj-pdf-3-cep-10.1177_0333102420920005 - Supplemental material for Cortical morphological changes in chronic migraine in a Taiwanese cohort: Surface- and voxel-based analyses

Supplemental material, sj-pdf-3-cep-10.1177_0333102420920005 for Cortical morphological changes in chronic migraine in a Taiwanese cohort: Surface- and voxel-based analyses by Kuan-Lin Lai, David M Niddam, Jong-Ling Fuh, Wei-Ta Chen, Jaw-Ching Wu and Shuu-Jiun Wang in Cephalalgia

Footnotes

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 disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Brain Research Center, National Yang-Ming University from The Featured Areas Research Center Program within the framework of the Higher Education Sprout Project by the Ministry of Education (MOE) in Taiwan; Ministry of Science and Technology of Taiwan [MOST 108-2321-B-010-014 -MY2 and 108-2321-B-010 001 and 108-2314-B-010-023-MY3 and 107-2314-B-010-018-MY3]; Ministry of Health and Welfare, Taiwan [MOHW 103-TDU-B-211-113-003, MOHW 104-TDU-B-211-113-003, MOHW 105-TDU-B-211-113-003, MOHW 106-TDU-B-211-113001, MOHW 107- TDU-B-211-123001, MOHW 108- TDU-B-211-133001]. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

ORCID iDs

Kuan-Lin Lai

Shuu-Jiun Wang

References

Dodick

DW.

Migraine. Lancet 2018; 391: 1315–1330.

Natoli

Manack

Dean

, et al. Global prevalence of chronic migraine: A systematic review. Cephalalgia 2010; 30: 599–609.

Lipton

RB.

Tracing transformation: Chronic migraine classification, progression, and epidemiology. Neurology 2009; 72: S3–S7.

DaSilva

Granziera

Snyder

, et al. Thickening in the somatosensory cortex of patients with migraine. Neurology 2007; 69: 1990–1995.

Kim

Suh

, et al. Thickening of the somatosensory cortex in migraine without aura. Cephalalgia 2014; 34: 1125–1133.

Maleki

Becerra

Brawn

, et al. Concurrent functional and structural cortical alterations in migraine. Cephalalgia 2012; 32: 607–620.

Messina

Rocca

Colombo

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

Rocca

Ceccarelli

Falini

, et al. Brain gray matter changes in migraine patients with T2-visible lesions: A 3-T MRI study. Stroke 2006; 37: 1765–1770.

Hougaard

Amin

Hoffmann

, et al. Structural gray matter abnormalities in migraine relate to headache lateralization, but not aura. Cephalalgia 2015; 35: 3–9.

10.

Valfre

Rainero

Bergui

, et al. Voxel-based morphometry reveals gray matter abnormalities in migraine. Headache 2008; 48: 109–117.

11.

Kim

Suh

Seol

, et al. Regional grey matter changes in patients with migraine: A voxel-based morphometry study. Cephalalgia 2008; 28: 598–604.

12.

Maleki

Barmettler

Moulton

, et al. Female migraineurs show lack of insular thinning with age. Pain 2015; 156: 1232–1239.

13.

Granziera

DaSilva

Snyder

, et al. Anatomical alterations of the visual motion processing network in migraine with and without aura. PLoS Med 2006; 3: e402.

14.

Hougaard

Amin

Arngrim

, et al. Sensory migraine aura is not associated with structural grey matter abnormalities. Neuroimage Clin 2016; 11: 322–327.

15.

Datta

Detre

Aguirre

, et al. Absence of changes in cortical thickness in patients with migraine. Cephalalgia 2011; 31: 1452–1458.

16.

Chong

Dodick

Schlaggar

, et al. Atypical age-related cortical thinning in episodic migraine. Cephalalgia 2014; 34: 1115–1124.

17.

Magon

May

Stankewitz

, et al. Cortical abnormalities in episodic migraine: A multi-center 3T MRI study. Cephalalgia 2019; 39: 665–673.

18.

Woldeamanuel

DeSouza

Sanjanwala

, et al. Clinical features contributing to cortical thickness changes in chronic migraine – a pilot study. Headache 2019; 59: 180–191.

19.

Lai

Chou

Fuh

, et al. Gray matter changes related to medication overuse in patients with chronic migraine. Cephalalgia 2016; 36: 1324–1333.

20.

Neeb

Bastian

Villringer

, et al. Structural gray matter alterations in chronic migraine: Implications for a progressive disease? Headache 2017; 57: 400–416.

21.

Coppola

Petolicchio

Di Renzo

, et al. Cerebral gray matter volume in patients with chronic migraine: Correlations with clinical features. J Headache Pain 2017; 18: 115.

22.

Headache Classification Committee of the International Headache Society (IHS). The International Classification of Headache Disorders, 3rd edition. Cephalalgia 2018; 38: 1–211.

23.

Schwedt

Chong

, et al. Accurate classification of chronic migraine via brain magnetic resonance imaging. Headache 2015; 55: 762–777.

24.

Winkler

Kochunov

Blangero

, et al. Cortical thickness or grey matter volume? The importance of selecting the phenotype for imaging genetics studies. Neuroimage 2010; 53: 1135–1146.

25.

Niddam

Lee

, et al. Altered cortical morphology in patients with chronic shoulder pain. Neurosci Lett 2019; 712: e134515.

26.

Dahnke

Yotter

Gaser

Cortical thickness and central surface estimation. Neuroimage 2013; 65: 336–348.

27.

Yotter

Dahnke

Thompson

, et al. Topological correction of brain surface meshes using spherical harmonics. Hum Brain Mapp 2011; 32: 1109–1124.

28.

Yotter

Thompson

Gaser

Algorithms to improve the reparameterization of spherical mappings of brain surface meshes. J Neuroimaging 2011; 21: e134–e147.

29.

Niddam

Lee

, et al. Brain structural changes in patients with chronic myofascial pain. Eur J Pain 2017; 21: 148–158.

30.

Smith

Nichols

TE.

Threshold-free cluster enhancement: Addressing problems of smoothing, threshold dependence and localisation in cluster inference. Neuroimage 2009; 44: 83–98.

31.

Borsook

Veggeberg

Erpelding

, et al. The insula: A “hub of activity” in migraine. Neuroscientist 2016; 22: 632–652.

32.

Torta

Legrain

Mouraux

, et al. Attention to pain! A neurocognitive perspective on attentional modulation of pain in neuroimaging studies. Cortex 2017; 89: 120–134.

33.

Anticevic

Dierker

Gillespie

, et al. Comparing surface-based and volume-based analyses of functional neuroimaging data in patients with schizophrenia. Neuroimage 2008; 41: 835–848.

34.

Van Essen

Dierker

Snyder

, et al. Symmetry of cortical folding abnormalities in Williams syndrome revealed by surface-based analyses. J Neurosci 2006; 26: 5470–5483.

35.

Liu

Lipnicki

Zhu

, et al. Cortical gyrification and sulcal spans in early stage Alzheimer’s disease. PLoS One 2012; 7: e31083.

36.

Han

Jovicich

Salat

, et al. Reliability of MRI-derived measurements of human cerebral cortical thickness: The effects of field strength, scanner upgrade and manufacturer. Neuroimage 2006; 32: 180–194.

37.

Freitag

Lyss

Nissan

GR.

Migraine disability, healthcare utilization, and expenditures following treatment in a tertiary headache center. Proc (Bayl Univ Med Cent) 2013; 26: 363–367.

38.

Yalinay Dikmen

Onur Aysevener

Kosak

, et al. Relationship between MIDAS, depression, anxiety and alexithymia in migraine patients. Acta Neurol Belg. Epub ahead of print 16 November 2017. DOI: 10.1007/s13760-017-0856-x.

39.

Dai

Zhong

Xiao

, et al. Gray matter correlates of migraine and gender effect: A meta-analysis of voxel-based morphometry studies. Neuroscience 2015; 299: 88–96.

40.

Guo

Chen

, et al. A meta-analysis of voxel-based morphometric studies on migraine. Int J Clin Exp Med 2015; 8: 4311–4319.

41.

Jia

Grey matter alterations in migraine: A systematic review and meta-analysis. Neuroimage Clin 2017; 14: 130–140.

42.

May

Structural brain imaging: A window into chronic pain. Neuroscientist 2011; 17: 209–220.

43.

Davis

Moayedi

Central mechanisms of pain revealed through functional and structural MRI. J Neuroimmune Pharmacol 2013; 8: 518–534.

44.

Rodriguez-Raecke

Niemeier

Ihle

, et al. Brain gray matter decrease in chronic pain is the consequence and not the cause of pain. J Neurosci 2009; 29: 13746–13750.

45.

Gaist

Hougaard

Garde

, et al. Migraine with visual aura associated with thicker visual cortex. Brain 2018; 141: 776–785.

46.

Desikan

Segonne

Fischl

, et al. An automated labeling system for subdividing the human cerebral cortex on MRI scans into gyral based regions of interest. Neuroimage 2006; 31: 968–980.

47.

Destrieux

Fischl

Dale

, et al. Automatic parcellation of human cortical gyri and sulci using standard anatomical nomenclature. Neuroimage 2010; 53: 1–15.

48.

Vila-Pueyo

Hoffmann

Romero-Reyes

, et al. Brain structure and function related to headache: Brainstem structure and function in headache. Cephalalgia 2019; 39: 1635–1660.

49.

Schulte

May

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

50.

Schulte

Mehnert

May

Longitudinal neuroimaging over 30 days: Temporal characteristics of migraine. Ann Neurol 2020; 87: 646–651.

51.

Schulte

Allers

May

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

52.

Schulte

Sprenger

May

Physiological brainstem mechanisms of trigeminal nociception: An fMRI study at 3T. Neuroimage 2016; 124: 518–525.

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

0.12 MB

0.32 MB

0.10 MB