White matter microstructural alterations in migraine: A diffusion-weighted MRI study

White matter microstructural alterations in migraine: A diffusion-weightedMRI study

Nikoletta Szabó a, Zsigmond Tamás Kincses a,!, Árpád Párdutz a, János Tajti a, Délia Szok a,Bernadett Tuka a, András Király a, Magor Babos b, Erika Vörös c, Giuseppe Bomboi d, Francesco Orzi d,László Vécsei aaDepartment of Neurology, Albert Szent-Györgyi Clinical Center, University of Szeged, Szeged, Hungaryb Euromedic Diagnostics Hungary Ltd., Szeged, HungarycDepartment of Radiology, Albert Szent-Györgyi Clinical Center, University of Szeged, Szeged, HungarydDepartment of Neurological Sciences, II Faculty of Medicine, Sapienza University of Rome, Rome, Italy

Sponsorships or competing interests that may be relevant to content are disclosed at the end of this article.

a r t i c l e i n f o

Article history:Received 9 March 2011Received in revised form 19 October 2011Accepted 29 November 2011

Keywords:MigraineDiffusion tensor imagingMRITract-Based Spatial Statistics

a b s t r a c t

Migraine is a common and disabling neurological disease. The pathomechanism that underlies the disor-der is not entirely understood, and reliable biomarkers are missing. In the current analysis we looked formicrostructural alterations of the brain white matter in migraine patients by means of diffusion-weighted magnetic resonance imaging. The measurements were carried out with a novel approach basedon fine-tuned nonlinear registration and nonparametric permutation test in an alignment-invariant tractrepresentation (Tract-Based Spatial Statistics). We found reduced fractional anisotropy in the right frontalwhite matter cluster of migraine patients. In the same region we also found increased mean diffusivityand increased radial diffusivity. The probabilistic tractography showed connection of this cluster to otherparts of the pain network (orbitofrontal cortex, insula, thalamus, dorsal midbrain). We speculate thatthese findings reflect maladaptive plastic changes or white matter disintegration.

! 2011 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved.

1. Introduction

Migraine is a chronic, predominantly genetically determineddisorder, characterised by recurrent attacks of throbbing headache,aggravated by physical activity, usually associated with centralsymptoms like perceptual hypersensitivity and nausea [48].Although not life-threatening, the disorder confers considerabledisability, robs quality of life, and bears an enormous socioeco-nomic burden [49,50]. The complex pathomechanism that under-lies the disease is not entirely understood, and establishedindicators of the disease are missing. Functional imaging altera-tions have been frequently reported in migraine [14,75,76], butthe studies on brain morphological changes are sparse and not con-clusive. Studies generally identified abnormalities both in brain-stem [42,57,77] and hemispheric structures [19,20,29].

Structural magnetic resonance imaging (MRI) studies, in spite ofthe different methodological approaches, consistently revealed

loss of gray matter in regions associated with pain processing,including the frontal cortex, temporal lobe, insula, and brainstem[40,57,64,65,74]. Whether such changes are consequences of thedisease is still debated [53].

Among the structural abnormalities, white matter microstruc-ture changes, as defined by diffusion-weighted MRI, are receivingmore and more attention. Diffusion-weighted MRI is sensitive toself-diffusion of water molecules, which in the brain is largelyrestricted by the membranes of the cellular elements. By fitting adiffusion tensor model, it is possible to estimate diffusion parame-ters that reflect the microscopic organisation of the measuredvolume [9].

Thus, alterations of white matter microstructure have been re-ported in migraine by a number of authors [20,29,47]. For instance,lower white matter fractional anisotropy (FA) and increased meandiffusivity (MD) were found in migraine patients by means of a his-togram analysis [58], while Li and co-workers, on the basis of a re-gion-of-interest (ROI) analysis approach showed reduced FA in thegenu, splenium, and body of the corpus callosum [47]. Similarly, inanother ROI-based analysis, Rocca and co-workers found reducedFA and higher MD in the right optic radiation of patients with aura[59]. DaSilva and colleagues, using a voxel-based morphometry-style analysis, described lower FA in the thalamocortical tract of

0304-3959/$36.00 ! 2011 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved.doi:10.1016/j.pain.2011.11.029

! Corresponding author. Address: Neuroimaging Research Group, Department ofNeurology, Albert Szent-Györgyi Clinical Center, University of Szeged, Semmelweisu. 6, Szeged 6725, Hungary.

E-mail addresses: [email protected], [email protected](Z.T. Kincses).

URL: http://www.nepsy.szote.u-szeged.hu/~kincsesz (Z.T. Kincses).

PAIN"153 (2012) 651–656

www.e l sev ie r . com/ loca t e / pa in

migraineurs, and similar alterations were found in the trigemino-thalamic tract and in the periaqueductal gray of patients withand without aura, respectively [20]. By using a similar approach,Granziera and colleagues found reduced FA in the visual motion-processing network [29].

Despite the undisputed merits, these studies have limitations.Histogram- and ROI-based analyses have low spatial localisingpower. Whole-brain voxel-based methods provide us with aremarkable spatial resolution, but optimal analysis is often com-promised by the standard registration algorithms and by unsatis-factory alignment of images of individual subjects. A furthersource of concern refers to the appropriateness of the statistical ap-proach used in previous studies, as the parametric statistical anal-ysis presumes normal data distribution, which is often notapplicable to these studies.

In the current analysis, we have investigated brain white matterdiffusion tensor imaging (DTI) microstructural alterations in mi-graine patients as compared to controls. The novelty of our studylies in an improved analytical approach, which was achieved byperforming a carefully tuned nonlinear registration, followed byprojection onto an alignment-invariant tract representation (the‘‘mean FA skeleton’’) as implemented in the Oxford Centre forFunctional MRI of the Brain (FMRIB) Software Library’s (FSL’s)Tract-Based Spatial Statistic (TBSS) toolbox [68] (http://www.fmri-b.ox.ac.uk/fsl/tbss/index.html). Statistical inference was carriedout with a nonparametric permutation test [55].

2. Methods

2.1. Participants

Twenty-one female migraine patients, without any historyof other neurological disorder, were recruited from outpatientsof the Headache Outpatient Clinic of the Department ofNeurology, University of Szeged. The diagnosis was based onthe criteria of the International Headache Society [48]. In orderto rule out possible confounding factors, all the patients werescreened for depression by the Hamilton Depression RatingScale [35]; 4 patients scoring higher than 16 points wereexcluded from the study. Demographics of the study groupare depicted in Table 1.

Seventeen age-matched, healthy female individuals, with nohistory of long-term headache, migraine, or other neurological orpsychiatric diseases, were included as controls.

The study was approved by the local ethics committee (author-ity number: 87/2009), and written consent was provided by all thesubjects.

2.2. Image acquisition

MRI scanning was carried out on a 1.5T GE Signa Excite HDxtMR scanner (GE Healthcare, Chalfont St Giles, UK). Three-dimen-sional spoiled gradient echo images (FSPGR: echo time [TE]:4.1 ms, repetition time [TR]: 10.276 ms, matrix: 256 ! 256, field

of view [FOV]: 25 ! 25 cm, Flip angle: 15#, in-plane resolution:1 ! 1 mm, slice thickness: 1 mm) and 60 direction diffusion-weighted images with 6 non-diffusion-weighted reference volume(TE: 93.8 ms, TR: 16,000 ms, matrix: 96 ! 96, FOV: 23 ! 23 cm, Flipangle: 90#, in-plane resolution: 2.4 ! 2.4 mm slice thickness:2.4 mm, b: 1000 s/m2, number of excitations [NEX]: 2, array spatialsensitivity encoding technique [ASSET]) were acquired for allsubjects.

2.3. Image analysis

Diffusion data were corrected for eddy currents and movementartefacts by 12 df affine linear registration to the first non-diffu-sion-weighted reference image [38]. Diffusion gradient directionswere reoriented according to the result of eddy current correction[45]. Diffusion tensors at each voxel were fitted by an algorithmincluded in FMRIB’s Diffusion Toolbox (FDT) of FSL (v. 4.0,www.fmrib.ox.ac.uk/fsl; [69]). FA, MD, and diffusivity parallel(k1) and perpendicular ((k2 + k3)/2) to the principal diffusiondirection were computed for the whole brain. In order to reducethe possible errors arising from misalignment of the images, weused the TBSS method [68]: All subjects’ FA data were aligned intoa common space derived from 58 high-resolution FA images ofhealthy subjects, using the FMRIB’s Nonlinear Registration Tool,FNIRT [3], which uses a b-spline representation of the registrationwarp field [62]. A mean FA image was created and then threshol-ded at FA = 0.2, deriving a mean FA skeleton that represents thecentres of all tracts common to the group. Each subject’s alignedFA data were then projected onto this skeleton and the resultingdata fed into voxel-wise cross-subject statistics. Modelling andinference using standard general linear model design set-up wasaccomplished using permutation-based cluster analysis (5000 per-mutation) [55] as implemented in FSL. The design encoded forgroup membership. Clusters were formed according to a definedthreshold (t = 2.3) and corrected for multiple comparisons (acrossspace) within the permutation framework by building up the nulldistribution of the maximum cluster size for each permutation(P < 0.05).

We also carried out a region of interest analysis on the whitematter regions whose FA values differed significantly betweengroups. We tested for differences in FA, MD, longitudinal, and per-pendicular diffusivity.

Connectivity of the identified differences in white matter integ-rity was defined by probabilistic tractography (FDT, part of FSL:www.fmrib.ox.ac.uk/fsl/fdt/). We fitted a multifibre diffusion mod-el [10] that estimates probability distributions on the direction of 1or more fibre populations at each brain voxel. Probabilistic tractog-raphy was then performed from any brain voxel by tracing stream-line samples through these probabilistic distributions on fibredirection. For tractography, we generated 5000 streamline samplesfrom each seed voxel to build up a connectivity distribution. Thenumber of these samples passing through each brain voxel is inter-preted as proportional to the probability of connection to the seedvoxel. By fitting a multifibre model to our diffusion data, we wereable to trace pathways through regions of fibre crossing [10]. Seedmasks were binary cluster-masks of the TBSS analysis.

3. Results

3.1. Focal white matter microstructure alterations

White matter microstructure as evaluated by group level voxel-wise FA differences in the centre of white matter fibre bundleswas significantly altered in migraine patients as compared to con-trols. Thedifferenceswereobserved in the right frontalwhitematter(maximal t-score at voxel location x = 25 mm, y = 24 mm, z = 5 mm

Table 1Patient demographics.

Migraine patients(n = 17)

Control group(n = 17)

Age, years ± SD (range) 34.65 ± 10.86 (21–54) 33.27 ± 11.32 (21–56)Gender Female FemaleWith aura 3 NADuration, years ± SD (range) 17.41 ± 9.59 (1–32) NAFrequency, attack/year ± SD

(range)33.1 ± 15.64 (12–60) NA

Right-handed 17 17

652 N. Szabó et al. / PAIN"153 (2012) 651–656

standard space coordinates; Fig. 1). Specifically, FAwas lower (Fig. 1.box-plot),whileMDandperpendicular diffusivitywere significantlyhigher in patients than in controls (P < 0.0088 and P < 0.0002,respectively, Fig. 2). Longitudinal diffusivity, on the other hand,was not different between groups (P > 0.101). In a whole brain anal-ysis, neither MD nor the longitudinal/perpendicular diffusivityshowed any significant difference between patients and controls.

In order to further characterise the microstructural alterationsfound in migraine, we carried out a correlation analysis betweendisease features and local FA by using an ROI approach. The resultswere negative for any association between the observed FA anddisease duration or attack frequency.

In order to exclude possible confounding effects of motion arte-facts groups, mean displacements were calculated from the resultof eddy current correction. There were no significant large motionartefacts, and absolute displacement was not different in the 2groups (P > 0.329).

3.2. Connectivity of focal FA changes

The probabilistic tractography indicated that fibres of the rightfrontal white matter showing the FA alteration (identified by theTBSS analysis) were connected to the ipsilateral prefrontal corticalregions, insula, thalamus, dorsal, and ventral midbrain. Fibres weretravelling in the direction of the occipital cortex through the puta-tive inferior fronto-occipital fasciculus. Some fibres also crossedthe midline through the corpus callosum (Fig. 3).

4. Discussion

Consistent with previous reports, in this study we observedbrain microstructural alterations as shown by diffusion-weightedMRI, in subjects with migraine. Our improved analysis approachrevealed that the reduced FA and the increased MD and transversaldiffusivity occurred in the right frontal white matter. A main find-ing of the study concerns the spatial pattern of connectivity of theaffected fibre bundle. The pattern was, in fact, found to be strik-ingly similar to the pain network described by Hadjipavlou andco-workers, which includes periaqueductal gray and cuneiform

nucleus, prefrontal cortex, amygdala, thalamus and hypothalamus,and rostroventral medulla [33].

Our results are in line with previous reports describing struc-tural and functional alterations of the frontal cortex of subjectswith migraine. Frontal cortical atrophy was shown in migraine pa-tients [64,74], and similar frontal cortical gray matter densityreduction was correlated to T2-visible lesion load [57]. Schmitzand colleagues found altered cognitive set shifting in migraineurs,correlating with the reduced gray matter density in the frontal lobe[66].

On the other hand, we cannot exclude the possibility that thewhite matter alterations related to the chronic painful state ofthe patients and not to migraine itself. Accordingly, functionalalterations within the pain network have been extensively re-ported in chronic pain conditions. For instance, functional altera-tions were detected in the right ventromedial prefrontal area,suggesting neuronal reorganisation in patients suffering fromchronic complex regional pain syndrome [27]. Similar alterationsof fMRI activations were reported in chronic prostatic pain in theanterior insula and in the cingulate cortex [25]. Medial prefrontalactivations were found during sustained back pain, while in theincreasing phase of the pain, the activation involved the insula,anterior cingulate gyrus, and somatosensory cortex, areas involvedin acute pain processing [8].

Chronic pain-related morphological changes in regions involvedin pain processing are reported by an increasing number of studies.Gray matter loss of the bilateral dorsolateral prefrontal cortex andright thalamus was observed in chronic back pain [5]. In fibromy-algia in one study, cingulate, insular, and medial frontal cortices,and parahippocampal gyri [44]; in a further study, postcentral gyri,amygdalae, hippocampi, superior frontal gyri, and anterior cingu-late gyri showed reduced gray matter density [52].

Nevertheless, the lack of associations in our study betweenattack frequency or disease duration and intensity of the micro-structural changes, might suggest changes specific to migraineand be unrelated to the chronic, repeated painful condition.

In addition to being consistent with previous reports, our dataseem to suggest a potential role of the observed microstructuralchanges in the pathophysiology of the disease, as their spatial

Fig. 1. Tract-Based Spatial Statistic (TBSS) indicates reduced fractional anisotropy (FA) in the right frontal white matter in migraine patients. The mean FA skeleton isrendered in light blue. A thickened version of the significant cluster is used for easier visualisation (green). The t-scores are depicted in red-to-yellow colours within thesignificant cluster. Mean FA of the 2 groups within the cluster is depicted on the graph. On the box-plot, the central mark is the mean, the boxes represent the 25% and 75%percentiles, and outliers are depicted as red crosses.

N. Szabó et al. / PAIN"153 (2012) 651–656 653

distribution appears to coincide with the pain processing circuitry.We do in fact speculate about 2 alternative interpretations of thealteration: (i) the structural changes could be the consequence ofmaladaptive plastic changes; or (ii) they might indicate degenera-tive processes that could be either primarily migraine specific orsecondarily caused by migraine pathology.

(i) Gray matter morphological changes due to used-dependentplasticity have already been reported in adults [13,24]. Similaralterations were also found in the white matter with DTI [67]. Asan expression of similar mechanisms, repeated pain stimuli werealso reported to induce increase of gray matter density in pain pro-cessing regions, including the cingulate and the contralateralsomatosensory cortex [71]. One might hypothesise that repeatedpainful attacks induce used-dependent plastic changes also in thewhite matter.

Apart from the repeated pain stimuli in migraineurs, other fac-tors related to the pathomechanism of the disorder might also con-tribute to maladaptive plastic changes: the altered corticalexcitability [1,4,7,15,26] might also lead to such changes. Plasticchanges were reported in the central nervous system of animalsafter induction of cortical spreading depression (CSD) [22,34]. Apossible energy deficit suggested in migraine [63,78] might alsocontribute to plastic changes in the brain [46].

One puzzling fact that might question the above-stated mal-adaptive plasticity hypothesis, however, is that in chronic painconditions [5,21,28,44,72] as well as in migraine [40,57,64,65,74],reduction rather than increase of gray matter size or density has

been reported, as seen, for example, in learning [24]. Similarly,white matter alterations due to use-dependent plasticity-like pro-cesses were expected to appear in the form of increased FA, as itwas reported recently [67]. Explanation of such controversy maylie in a lack of noxious stimulus in chronic pain [53], chronificationof the pain condition, compensatory mechanisms [79] or affectivecomponents [36], and personality traits related to migraine [12].

(ii) An alternative hypothesis is that the reduced FA reflects adegenerative process in migraineurs. CSD shown to progressthrough the occipital cortex during the visual aura phase [32]might well constitute the noxious process [54] sufficient to causeneuroinflammation and the consequent pain generation [6,17,41]and cellular damage. The excessive release of glutamate [18] is alsoknown to induce excitotoxicity and cell death [51]. The activationof matrix metalloproteinases (MMP) can elicit the leakage ofblood–brain barrier and lead to inflammatory response and neuro-nal damage as well [30]. It is proven that CSD in animals causesupregulation of MMP-9 [31], and increased MMMP activity wasalso recently detected in human migraineurs [11]. In line with thishypothesis, a recent study by Yilmaz and co-workers found in-creased ictal level of S100B (a marker of glial damage) and neu-ron-specific enolase (a marker of neuronal damage) inmigraineurs without aura [80]. There seems, therefore, to be someevidence for biochemical changes potentially involved in the disin-tegration of white matter fibre bundles that might be reflected byreduction of FA, increase of MD, and augmented perpendicular dif-fusivity. Similar patterns of DTI abnormalities are most frequently

Controls Patients6.87

7.27.47.67.88

8.28.4

MD

Controls Patients10.811

11.211.411.611.812

12.212.412.612.8

L1

Controls Patients

4.5

5

5.5

6

6.5(L2+L3)/2

Fig. 2. Mean diffusivity (MD), longitudinal (L1), and perpendicular ((L2 + L3)/2) diffusivity in the cluster where reduced fractional anisotropy (FA) was found. MD was higherin migraine patients (P < 0.0088), which was explained by the increase in radial diffusivity (P < 0.0002). L1 did not differ between patients and controls (P > 0.101). On thebox-plot, the central mark is the mean, the boxes represent the 25% and 75% percentiles, and outliers are depicted as red crosses.

Fig. 3. Connectivity of the white matter cluster showing significantly lower fractional anisotropy (FA) in migraine patients than in controls. The binary cluster masks wereused as seed mask for each patient. Five thousand streamline samples from each seed voxel were drawn to build up a connectivity distribution that was thresholded for 2500particles for each subject and binarised. Population connectivity maps were derived for controls (upper row) and patients (lower row).

654 N. Szabó et al. / PAIN"153 (2012) 651–656

reported as a consequence of neurodegenerative processes [23,37].Reduction of FA and longitudinal diffusivity was reported in axonalloss [39,56,70], while increased perpendicular diffusivity seems tobe a sign of demyelination [39,70]. The serum markers of neuronaland glial damage reported recently [80] might indicate combineddamage.

Migraine was associated with appearance of T2 white matter le-sions [16], and the lesions have been widely considered of ischemicnature [43], but not without criticism and alternative hypotheses[60,61]. For instance, the coexistence of antineuronal antibody sug-gested the inflammatory origin of the altered MRI signal [73]. In arecent study by Rocca et al., reduction of frontal gray matter wasfound to be correlated with the T2 visible lesion load [58]. Theauthors explained this correlation with retrograde degenerationof axons passing through the macroscopic lesions. In our study,however, only one patient had a right frontal T2 visible white mat-ter lesion. Because of the close proximity of the lesion to our clus-ter of FA difference, we have repeated the analysis with theexclusion of this subject, but the results were essentially unaltered.Hence, it is likely that white matter microstructural alterations arenot directly related to the T2 visible lesions in migraine.

Our results, however, have to be handled with care; currentlywe have no direct evidence to prove that migraine is a progressivedisease that results in definitive change in brain structure. The factthat migraine remits with age, and that white matter lesions wereshown to be reversible [2], raise the possibility that our findingshave no direct role in migraine pathogenesis. Also, it must be takeninto account that these findings might be only an epiphenomenonand reflect only personality traits [12] or changes related to comor-bidity [47] (regarding this latter, we excluded the confounding ef-fect of depression in the current study). To exclude the possibilityof transient changes, longitudinal studies are needed.

The results of our study suggest that diffusion imaging in re-search settings has potentiality for becoming a biomarker tool inmigraine. However, given the heterogeneity of the disease andthe level of between-subject variability of diffusion parameters,in the clinical setting white matter integrity measured by DTI can-not currently be a specific and sensitive marker of the disease.

While altered frontal white matter integrity seems to be a clueto the pathogenesis of migraine, further investigations into the dif-ferent phases of the disease would help to elucidate the pathoge-netic importance of our findings.

Conflict of interest statement

The authors declare no conflicts of interest.

Acknowledgements

The research was supported by the ‘‘Neuroscience ResearchGroup of the Hungarian Academy of Sciences and University ofSzeged’’ grant. Dr. Vécsei was supported by the Medical ResearchCouncil (ETT 026-04). Dr. Zsigmond Tamás Kincses was supportedby the Bolyai Scholarship Programme of the Hungarian Academy ofSciences. Dr. Kincses, Dr. Szabó, Dr. Párdutz, and Dr. Vécsei weresupported by ‘‘TÁMOP-4.2.1/B-09/1/KONV-2010-0005—Creatingthe Center of Excellence at the University of Szeged’’ grant. Dr.Bomboi was supported by the European Federation of NeurologicalSocieties Department to Department Program. We thank Dr. Johan-nes Klein for his useful discussion on tractography results.

References

[1] Afra J, Proietti Cecchini A, Sandor PS, Schoenen J. Comparison of visual andauditory evoked cortical potentials in migraine patients between attacks. ClinNeurophysiol 2000;111:1124–9.

[2] Agarwal S, Magu S, Kamal K. Reversible white matter abnormalities in a patientwith migraine. Neurol India 2008;56:182–5.

[3] Andersson JLR, Jenkinson M, Smith S. Non-linear optimisation: FMRIB technicalreport. Oxford, UK: FMRIB; 2007.

[4] Antal A, Temme J, Nitsche MA, Varga ET, Lang N, Paulus W. Altered motionperception in migraineurs: evidence for interictal cortical hyperexcitability.Cephalalgia 2005;25:788–94.

[5] Apkarian AV, Sosa Y, Sonty S, Levy RM, Harden RN, Parrish TB, Gitelman DR.Chronic back pain is associated with decreased prefrontal and thalamic graymatter density. J Neurosci 2004;24:10410–5.

[6] Arnold G, Reuter U, Kinze S, Wolf T, Einhaupl KM. Migraine with aura showsgadolinium enhancement which is reversed following prophylactic treatment.Cephalalgia 1998;18:644–6.

[7] Aurora SK, Ahmad BK, Welch KM, Bhardhwaj P, Ramadan NM. Transcranialmagnetic stimulation confirms hyperexcitability of occipital cortex inmigraine. Neurology 1998;50:1111–4.

[8] Baliki MN, Chialvo DR, Geha PY, Levy RM, Harden RN, Parrish TB, Apkarian AV.Chronic pain and the emotional brain: specific brain activity associated withspontaneous fluctuations of intensity of chronic back pain. J Neurosci2006;26:12165–73.

[9] Beaulieu C. The basis of anisotropic water diffusion in the nervous system—atechnical review. NMR Biomed 2002;15:435–55.

[10] Behrens TE, Berg HJ, Jbabdi S, Rushworth MF, Woolrich MW. Probabilisticdiffusion tractography with multiple fibre orientations: what can we gain?Neuroimage 2007;34:144–55.

[11] Bernecker C, Pailer S, Kieslinger P, Horejsi R, Moller R, Lechner A, Wallner-Blazek M, Weiss S, Fazekas F, Truschnig-Wilders M, Gruber HJ. Increasedmatrix metalloproteinase activity is associated with migraine and migraine-related metabolic dysfunctions. Eur J Neurol 2011;18:571–6.

[12] Blankstein U, Chen J, Diamant NE, Davis KD. Altered brain structure in irritablebowel syndrome: potential contributions of pre-existing and disease-drivenfactors. Gastroenterology 2010;138:1783–9.

[13] Boyke J, Driemeyer J, Gaser C, Buchel C, May A. Training-induced brainstructure changes in the elderly. J Neurosci 2008;28:7031–5.

[14] Cao Y, Aurora SK, Nagesh V, Patel SC, Welch KM. Functional MRI-BOLD ofbrainstem structures during visually triggered migraine. Neurology2002;59:72–8.

[15] Chadaide Z, Arlt S, Antal A, Nitsche MA, Lang N, Paulus W. Transcranial directcurrent stimulation reveals inhibitory deficiency in migraine. Cephalalgia2007;27:833–9.

[16] Cooney BS, Grossman RI, Farber RE, Goin JE, Galetta SL. Frequency of magneticresonance imaging abnormalities in patients with migraine. Headache1996;36:616–21.

[17] Cui Y, Takashima T, Takashima-Hirano M,Wada Y, Shukuri M, Tamura Y, Doi H,Onoe H, Kataoka Y, Watanabe Y. 11C-PK11195 PET for the in vivo evaluation ofneuroinflammation in the rat brain after cortical spreading depression. J NuclMed 2009;50:1904–11.

[18] D’Andrea G, Cananzi AR, Joseph R, Morra M, Zamberlan F, Ferro Milone F,Grunfeld S, Welch KM. Platelet glycine, glutamate and aspartate in primaryheadache. Cephalalgia 1991;11:197–200.

[19] DaSilva AF, Granziera C, Snyder J, Hadjikhani N. Thickening in thesomatosensory cortex of patients with migraine. Neurology 2007;69:1990–5.

[20] DaSilva AF, Granziera C, Tuch DS, Snyder J, Vincent M, Hadjikhani N. Interictalalterations of the trigeminal somatosensory pathway and periaqueductal graymatter in migraine. Neuroreport 2007;18:301–5.

[21] Davis KD, Pope G, Chen J, Kwan CL, Crawley AP, Diamant NE. Cortical thinningin IBS: implications for homeostatic, attention, and pain processing. Neurology2008;70:153–4.

[22] Dehbandi S, Speckmann EJ, Pape HC, Gorji A. Cortical spreading depressionmodulates synaptic transmission of the rat lateral amygdala. Eur J Neurosci2008;27:2057–65.

[23] Della Nave R, Ginestroni A, Tessa C, Giannelli M, Piacentini S, Filippi M,Mascalchi M. Regional distribution and clinical correlates of white matterstructural damage in Huntington disease: a tract-based spatial statistics study.AJNR Am J Neuroradiol 2010;31:1675–81.

[24] Draganski B, Gaser C, Busch V, Schuierer G, Bogdahn U, May A. Neuroplasticity:changes in grey matter induced by training. Nature 2004;427:311–2.

[25] Farmer MA, Chanda ML, Parks EL, Baliki MN, Apkarian AV, Schaeffer AJ. Brainfunctional and anatomical changes in chronic prostatitis/chronic pelvic painsyndrome. J Urol 2011;186:117–24.

[26] Gawel M, Connolly JF, Rose FC. Migraine patients exhibit abnormalities in thevisual evoked potential. Headache 1983;23:49–52.

[27] Geha PY, Baliki MN, Harden RN, Bauer WR, Parrish TB, Apkarian AV. The brainin chronic CRPS pain: abnormal gray–white matter interactions in emotionaland autonomic regions. Neuron 2008;60:570–81.

[28] Gerstner G, Ichesco E, Quintero A, Schmidt-Wilcke T. Changes in regional grayandwhitematter volume in patientswithmyofascial-type temporomandibulardisorders: a voxel-based morphometry study. J Orofac Pain 2011;25:99–106.

[29] Granziera C, DaSilva AF, Snyder J, Tuch DS, Hadjikhani N. Anatomicalalterations of the visual motion processing network in migraine with andwithout aura. PLoS Med 2006;3:e402.

[30] Gupta VK. CSD, BBB and MMP-9 elevations: animal experiments versus clinicalphenomena in migraine. Expert Rev Neurother 2009;9:1595–614.

[31] Gursoy-Ozdemir Y, Qiu J, Matsuoka N, Bolay H, Bermpohl D, Jin H, Wang X,Rosenberg GA, Lo EH, Moskowitz MA. Cortical spreading depression activatesand upregulates MMP-9. J Clin Invest 2004;113:1447–55.

N. Szabó et al. / PAIN"153 (2012) 651–656 655

[32] Hadjikhani N, Sanchez Del Rio M, Wu O, Schwartz D, Bakker D, Fischl B, KwongKK, Cutrer FM, Rosen BR, Tootell RB, Sorensen AG, Moskowitz MA. Mechanismsof migraine aura revealed by functional MRI in human visual cortex. Proc NatlAcad Sci USA 2001;98:4687–92.

[33] Hadjipavlou G, Dunckley P, Behrens TE, Tracey I. Determining anatomicalconnectivities between cortical and brainstem pain processing regions inhumans: a diffusion tensor imaging study in healthy controls. Pain2006;123:169–78.

[34] Haghir H, Kovac S, Speckmann EJ, Zilles K, Gorji A. Patterns of neurotransmitterreceptor distributions following cortical spreading depression. Neuroscience2009;163:1340–52.

[35] Hamilton M. A rating scale for depression. J Neurol Neurosurg Psychiatry1960;23:56–62.

[36] Hsu MC, Harris RE, Sundgren PC, Welsh RC, Fernandes CR, Clauw DJ, WilliamsDA. No consistent difference in gray matter volume between individuals withfibromyalgia and age-matched healthy subjects when controlling for affectivedisorder. Pain 2009;143:262–7.

[37] Ibarretxe-Bilbao N, Junque C, Marti MJ, Valldeoriola F, Vendrell P, Bargallo N,Zarei M, Tolosa E. Olfactory impairment in Parkinson’s disease and whitematter abnormalities in central olfactory areas: a voxel-based diffusion tensorimaging study. Mov Disord 2010;25:1888–94.

[38] Jenkinson M, Smith S. A global optimisation method for robust affineregistration of brain images. Med Image Anal 2001;5:143–56.

[39] Kim JH, Budde MD, Liang HF, Klein RS, Russell JH, Cross AH, Song SK. Detectingaxon damage in spinal cord from a mouse model of multiple sclerosis.Neurobiol Dis 2006;21:626–32.

[40] Kim JH, Suh SI, Seol HY, Oh K, Seo WK, Yu SW, Park KW, Koh SB. Regional greymatter changes in patients with migraine: a voxel-based morphometry study.Cephalalgia 2008;28:598–604.

[41] Knotkova H, Pappagallo M. Imaging intracranial plasma extravasation in amigraine patient: a case report. Pain Med 2007;8:383–7.

[42] Kruit MC, van Buchem MA, Hofman PA, Bakkers JT, Terwindt GM, Ferrari MD,Launer LJ. Migraine as a risk factor for subclinical brain lesions. JAMA2004;291:427–34.

[43] Kruit MC, Launer LJ, Overbosch J, van Buchem MA, Ferrari MD. Ironaccumulation in deep brain nuclei in migraine: a population-based magneticresonance imaging study. Cephalalgia 2009;29:351–9.

[44] Kuchinad A, Schweinhardt P, Seminowicz DA, Wood PB, Chizh BA, BushnellMC. Accelerated brain gray matter loss in fibromyalgia patients: prematureaging of the brain? J Neurosci 2007;27:4004–7.

[45] Leemans A, Jones DK. The B-matrix must be rotated when correcting forsubject motion in DTI data. Magn Reson Med 2009;61:1336–49.

[46] Levy M, Faas GC, Saggau P, Craigen WJ, Sweatt JD. Mitochondrial regulation ofsynaptic plasticity in the hippocampus. J Biol Chem 2003;278:17727–34.

[47] Li XL, Fang YN, Gao QC, Lin EJ, Hu SH, Ren L, Ding MH, Luo BN. A diffusiontensor magnetic resonance imaging study of corpus callosum from adultpatients with migraine complicated with depressive/anxious disorder.Headache 2011;51:237–45.

[48] Lipton RB, Stewart WF. Prevalence and impact of migraine. Neurol Clin1997;15:1–13.

[49] Lipton RB, Stewart WF, von Korff M. Burden of migraine: societal costs andtherapeutic opportunities. Neurology 1997;48:S4–9.

[50] Lipton RB, Bigal ME, Steiner TJ, Silberstein SD, Olesen J. Classification ofprimary headaches. Neurology 2004;63:427–35.

[51] Longoni M, Ferrarese C. Inflammation and excitotoxicity: role in migrainepathogenesis. Neurol Sci 2006;27:S107–10.

[52] Lutz J, Jager L, de Quervain D, Krauseneck T, Padberg F, Wichnalek M, Beyer A,Stahl R, Zirngibl B, Morhard D, Reiser M, Schelling G. White and gray matterabnormalities in the brain of patients with fibromyalgia: a diffusion-tensorand volumetric imaging study. Arthritis Rheum 2008;58:3960–9.

[53] May A. Morphing voxels: the hype around structural imaging of headachepatients. Brain 2009;132:1419–25.

[54] Moskowitz MA. Pathophysiology of headache—past and present. Headache2007;47:S58–63.

[55] Nichols TE, Holmes AP. Nonparametric permutation tests for functionalneuroimaging: a primer with examples. Hum Brain Mapp 2002;15:1–25.

[56] Pierpaoli C, Barnett A, Pajevic S, Chen R, Penix LR, Virta A, Basser P. Waterdiffusion changes in Wallerian degeneration and their dependence on whitematter architecture. Neuroimage 2001;13:1174–85.

[57] Rocca MA, Colombo B, Pagani E, Falini A, Codella M, Scotti G, Comi G, Filippi M.Evidence for cortical functional changes in patients with migraine and whitematter abnormalities on conventional and diffusion tensor magneticresonance imaging. Stroke 2003;34:665–70.

[58] Rocca MA, Ceccarelli A, Falini A, Colombo B, Tortorella P, Bernasconi L, Comi G,Scotti G, Filippi M. Brain gray matter changes in migraine patients with T2-visible lesions: a 3-T MRI study. Stroke 2006;37:1765–70.

[59] Rocca MA, Pagani E, Colombo B, Tortorella P, Falini A, Comi G, Filippi M.Selective diffusion changes of the visual pathways in patients with migraine: a3-T tractography study. Cephalalgia 2008;28:1061–8.

[60] Rozen TD. Vanishing cerebellar infarcts in a migraine patient. Cephalalgia2007;27:557–60.

[61] Rozen TD. Images from headache: white matter lesions of migraine are notstatic. Headache 2010;50:305–6.

[62] Rueckert D, Sonoda LI, Hayes C, Hill DL, Leach MO, Hawkes DJ. Nonrigidregistration using free-form deformations: application to breast MR images.IEEE Trans Med Imaging 1999;18:712–21.

[63] Sas K, Pardutz A, Toldi J, Vecsei L. Dementia, stroke and migraine—somecommon pathological mechanisms. J Neurol Sci 2010;299:55–65.

[64] Schmidt-Wilcke T, Ganssbauer S, Neuner T, Bogdahn U, May A. Subtle greymatter changes between migraine patients and healthy controls. Cephalalgia2008;28:1–4.

[65] Schmitz N, Admiraal-Behloul F, Arkink EB, Kruit MC, Schoonman GG, FerrariMD, van Buchem MA. Attack frequency and disease duration as indicators forbrain damage in migraine. Headache 2008;48:1044–55.

[66] Schmitz N, Arkink EB, Mulder M, Rubia K, Admiraal-Behloul F, Schoonman GG,Kruit MC, Ferrari MD, van Buchem MA. Frontal lobe structure and executivefunction in migraine patients. Neurosci Lett 2008;440:92–6.

[67] Scholz J, Klein MC, Behrens TE, Johansen-Berg H. Training induces changes inwhite-matter architecture. Nat Neurosci 2009;12:1370–1.

[68] Smith SM, Jenkinson M, Woolrich MW, Beckmann CF, Behrens TE, Johansen-Berg H, Bannister PR, De Luca M, Drobnjak I, Flitney DE, Niazy RK, Saunders J,Vickers J, Zhang Y, De Stefano N, Brady JM, Matthews PM. Advances infunctional and structural MR image analysis and implementation as FSL.Neuroimage 2004;23:S208–19.

[69] Smith SM, Jenkinson M, Johansen-Berg H, Rueckert D, Nichols TE, Mackay CE,Watkins KE, Ciccarelli O, Cader MZ, Matthews PM, Behrens TE. Tract-basedspatial statistics: voxelwise analysis of multi-subject diffusion data.Neuroimage 2006;31:1487–505.

[70] Sun SW, Liang HF, Trinkaus K, Cross AH, Armstrong RC, Song SK. Noninvasivedetection of cuprizone induced axonal damage and demyelination in themouse corpus callosum. Magn Reson Med 2006;55:302–8.

[71] Teutsch S, Herken W, Bingel U, Schoell E, May A. Changes in brain gray matterdue to repetitive painful stimulation. Neuroimage 2008;42:845–9.

[72] Tu CH, Niddam DM, Chao HT, Chen LF, Chen YS, Wu YT, Yeh TC, Lirng JF, HsiehJC. Brain morphological changes associated with cyclic menstrual pain. Pain2011;150:462–8.

[73] Turkoglu R, Tuzun E, Icoz S, Birisik O, Erdag E, Kurtuncu M, Akman-Demir G.Antineuronal antibodies in migraine patients with white matter lesions. Int JNeurosci 2011;121:33–6.

[74] Valfre W, Rainero I, Bergui M, Pinessi L. Voxel-based morphometry revealsgray matter abnormalities in migraine. Headache 2008;48:109–17.

[75] Weiller C, May A, Limmroth V, Juptner M, Kaube H, Schayck RV, Coenen HH,Diener HC. Brain stem activation in spontaneous human migraine attacks. NatMed 1995;1:658–60.

[76] Welch KM, Ramadan NM. Mitochondria, magnesium and migraine. J Neurol Sci1995;134:9–14.

[77] Welch KM, Cao Y, Aurora S, Wiggins G, Vikingstad EM. MRI of the occipitalcortex, red nucleus, and substantia nigra during visual aura of migraine.Neurology 1998;51:1465–9.

[78] Welch KM, Nagesh V, Aurora SK, Gelman N. Periaqueductal gray matterdysfunction in migraine: cause or the burden of illness? Headache2001;41:629–37.

[79] Woolf CJ, Salter MW. Neuronal plasticity: increasing the gain in pain. Science2000;288:1765–9.

[80] Yilmaz N, Karaali K, Ozdem S, Turkay M, Unal A, Dora B. Elevated S100B andneuron specific enolase levels in patients with migraine-without aura:evidence for neurodegeneration? Cell Mol Neurobiol 2011;31:579–85.

656 N. Szabó et al. / PAIN"153 (2012) 651–656