Abstract
Background
Migraine attacks manifest with hypersensitivities to light, sound, touch and odor. Some people with migraine have photosensitivity between migraine attacks, suggesting persistent alterations in the integrity of brain regions that process light. Although functional neuroimaging studies have shown visual stimulus induced “hyperactivation” of visual cortex regions in migraineurs between attacks, whether photosensitivity is associated with alterations in brain structure is unknown.
Methods
Levels of photosensitivity were evaluated using the Photosensitivity Assessment Questionnaire in 48 interictal migraineurs and 48 healthy controls. Vertex-by-vertex measurements of cortical thickness were assessed in 28 people with episodic migraine who had interictal photosensitivity (mean age = 35.0 years, SD = 12.1) and 20 episodic migraine patients without symptoms of interictal photosensitivity (mean age = 36.0 years, SD = 11.4) using a general linear model design.
Results
Migraineurs have greater levels of interictal photosensitivity relative to healthy controls. Relative to migraineurs without interictal photosensitivity, migraineurs with interictal photosensitivity have thicker cortex in several brain areas including the right lingual, isthmus cingulate and pericalcarine regions, and the left precentral, postcentral and supramarginal regions.
Conclusion
Episodic migraineurs with interictal photosensitivity have greater cortical thickness in the right parietal-occipital and left fronto-parietal regions, suggesting that persistent light sensitivity is associated with underlying structural alterations.
Keywords
Introduction
Hypersensitivity to light and light avoidance are hallmark features during a migraine attack, and bright, flickering lights often precipitate an attack (1).
Previous studies using functional magnetic resonance imaging (fMRI) and positron emission tomography (PET) have shown that migraineurs with and without aura have hyperactivation in striate and extrastriate regions (2–6) compared to healthy controls. Studies demonstrate stronger pattern-induced and light-induced visual discomfort in ictal and interictal migraine patients relative to healthy controls (7–11). Interictal migraine patients also report greater symptoms of photosensitivity compared to healthy controls using self-report questionnaires (5,12). Although visual hypersensitivities are well described in migraine, whether the presence and severity of interictal photosensitivity relates to brain structural alterations remains to be investigated.
In this study, we used a short self-report questionnaire to evaluate photosensitivity and to assess differences in cortical structure in migraineurs with interictal photosensitivity (MwiPh) compared to migraineurs without interictal photosensitivity (MwoiPh). The objectives of this study were to assess: 1) severity of photosensitivity in patients with episodic migraine compared to a healthy control cohort using the Photosensitivity Assessment Questionnaire (PAQ); 2) differences in cortical thickness between MwiPh and MwoiPh and 3) whether PAQ scores are correlated with alterations in brain structure in MwiPh.
Methods
Forty-eight episodic migraine patients (13 male, 35 female) and 48 healthy controls (13 male, 35 female) were included in this study. Individuals were recruited from the Mayo Clinic and from Washington University School of Medicine in St. Louis, Missouri. All participants gave informed consent prior to study participation according to institutional review board guidelines. Individuals were recruited from neurology clinics and from the community using a database of research volunteers and via advertisements. Exclusion criteria for migraineurs and healthy controls included a history of head trauma, acute or chronic pain from conditions other than migraine, chronic depression, anxiety and other psychiatric disorders, and neurological diseases other than migraine (for the migraine population). Patients were diagnosed with episodic migraine and were not overusing migraine-abortive medications according to the diagnostic criteria in the International Classification of Headache Disorders III beta (ICHD-III beta), and were not using migraine-prophylactic medications or opiates (1). All migraineurs had migraine for a minimum of three years and all patients reported symptoms of light sensitivity during a migraine attack (i.e. all patients answered “yes” to the question: “Do you experience light sensitivity during the attack?”).
Migraineurs were headache free at the time of imaging as well as for at least 48 hours prior to magnetic resonance imaging (MRI). All participants completed the PAQ (13), the Beck Depression Inventory (BDI-II) (14), and the State/Trait Anxiety Inventory (STAI, Form Y-1 and Form Y-2) (15). All migraine patients kept a headache diary for seven days after MRI in order to calculate the number of hours until the next migraine attack. For those migraineurs who did not have a migraine attack within the seven days after MRI (n = 7), 150 hours was used for determining correlations.
The PAQ is a 16-item self-report validated questionnaire (13) that assesses light-seeking behaviors (eight questions) and light-avoidance behaviors (eight questions) (the PAQ is provided as supplementary material). Participants were asked to read each statement and circle whether the statement applies to them (“agree,” score = 1) or does not apply to them (“disagree,” score = 0). The light-avoidance domain raw scores range from 0 to 8. We evaluated only domain questions relating to light avoidance, which we will continue to refer to within this paper as “photosensitivity.”
Two previously published normative studies assessed light-avoidance scores using the PAQ for healthy people. Both studies reported similar means and standard deviations (SD) (mean = 0.12; SD = 0.13 (16) versus mean = 0.11; SD = 0.13 (17)). In accordance with these data, we classified raw scores of 0 and 1 to be within the “normal” range and to reflect no photosensitivity and raw scores 2 SD above the mean (raw score of ≥ 2) to reflect “abnormal,” or high levels of photosensitivity.
Imaging parameters
All imaging was acquired on Siemens (Erlangen, Germany) 3-Tesla whole-body MRI scanners (Siemens Trio at Washington University and Siemens Skyra at Mayo Clinic). Study participants were scanned during a single session and demographic information as well as questionnaire data were collected on the same day. During the time period of data acquisition, all imaging sequences remained the same and no updates were performed on the scanners. Three-dimensional (3D) imaging included T1-weighted sagittal magnetization-prepared-rapid acquisition with gradient echo (MP-RAGE), echo time (TE) = 3.03 ms, repetition time (TR) = 2400 ms, flip angle = 8 degrees, 1 × 1 × 1.3 mm3 voxels, 256 mm2 field of view (FOV) at Mayo Clinic or TE = 3.16 ms, TR = 2400 ms, flip angle = 8 degrees, 1 × 1 × 1 mm3 voxels, 256 mm2 FOV at Washington University. Axial T2-weighted imaging included either TE = 84 ms, TR = 6800 ms, flip angle = 150 degrees, 38 slices, 1 × 1 × 4 mm3 voxels, 256 mm2 FOV at Mayo Clinic or TE = 88 ms, TR = 6280 ms, flip angle = 120 degrees, 36 slices, 1 × 1 × 4 mm3 voxels, 256 mm2 FOV at Washington University. Twenty-seven migraine patients and 21 healthy controls were imaged at Washington University and 21 migraine patients and 16 healthy controls were scanned at Mayo Clinic.
All imaging was reviewed by a board-certified neuroradiologist. Data were excluded from the analysis upon detection of gross structural abnormalities as identified on T1-weighted imaging or occurrence of white matter hyperintensities as identified on T2-weighted imaging.
Cortical thickness estimation
T1 MP-RAGE imaging data were reconstructed using FreeSurfer, version 5.3 (http://surfer.nmr.mgh.harvard.edu/). To avoid image processing inconsistencies (18), all data were analyzed by a single technician using a single Mac workstation running operating system (OS) X Lion 10.75 software. FreeSurfer uses an automated brain segmentation and parcellation technique that includes non-brain removal, Talairach transformation, gray and white matter parcellation and segmentation, intensity normalization and deformation of the brain surface structure (19–22). Cortical thickness was estimated by calculating the distance between the gray and white matter boundary to the boundary of the gray matter and cerebrospinal fluid at each vertex for the right and left hemisphere separately. Automated output from brain segmentations and parcellations for each participant was manually checked for motion artifacts and for accuracy in delineating brain boundary lines.
Statistical analysis
Group differences (MwiPh versus MwoiPh) were estimated using a whole-brain vertex-by-vertex general linear model approach. Several neuroimaging studies have demonstrated that cortical thickness is negatively affected by advanced aging and severe depression (23,24). Although our participant cohort was relatively young (18–46 years of age) and only three individuals had mild depression according to the BDI-II criteria, we included age and depression (i.e. scores on the BDI-II) as covariates in the design matrix to account for potential variance in our results. Significance levels for all vertex-by-vertex cortical thickness comparisons were set at p < 0.025 to account for multiple comparisons over both hemispheres (p = 0.05 × 2) and cluster-corrected using Monte Carlo simulations (1000 iterations) (25). Cortical thickness maps were smoothed with a 10 mm full width at half maximum (FWHM) kernel. Group (MwiPh and MwoiPh) demographic data were assessed using two-tailed T-tests or Fisher’s exact tests as appropriate. Pearson’s correlations or two-tailed T-tests were used for post-hoc data analyses.
Results
Photosensitivity scores
Participant characteristics for episodic migraine patients with interictal photosensitivity and healthy controls.
f: female; m: male; BDI-II: Beck Depression Inventory (II); TRAIT: Trait Anxiety Inventory; PAQ: Photosensitivity Assessment Questionnaire.
Previous studies indicate lower thresholds of light sensitivity for depressed patients and a relationship between light sensitivity and anxiety in psychiatric populations (16,26). Henceforth, the potential relationships between photosensitivity and measures of anxiety and depression as well as migraine-specific characteristics such as headache frequency and years lived with migraine were investigated using post-hoc Pearson’s correlations. Results indicated no significant correlations between the PAQ with measures of depression (r = 0.13; p = 0.52), state anxiety (r = 0.08; p = 0.69), trait anxiety (r = 0.20; p = 0.30), headache frequency (r = –0.01; p = 0.89) and years lived with migraine (r = 0.32, p = 0.099) in MwiPh. A post-hoc analysis of MwiPh showed no difference in photosensitivity scores (p = 0.32) between MwiPh with aura (mean photosensitivity raw score = 4.2, SD = 2.3) and MwoiPh with out aura (mean photosensitivity raw score = 3.46, SD = 1.3).
Participant characteristics for episodic migraine patients with interictal photosensitivity (MwiPh) and episodic migraine patients without interictal photosensitivity (MwoiPh).
f: female; m: male; BDI-II: Beck Depression Inventory (II); STATE: State Anxiety Inventory; TRAIT: Trait Anxiety Inventory; Headache frequency: days with headache per month; Time to next attack (hours): onset of migraine after scanning.
Cortical thickness differences between MwiPh and MwoiPh
MwiPh had several areas of greater cortical thickness compared to MwoiPh (see Figure 1). Areas of greater cortical thickness in MwiPh included regions in the: right lingual gyrus, right isthmus cingulate, right pericalcarine area, left precentral, left postcentral, and left supramarginal region. A post-hoc analysis showed no cortical thickness differences for any of these regions when comparing MwiPh patients who had migraine auras compared to those without aura. There were no regions for which MwiPh had less cortical thickness compared to MwoiPh.
Vertex-by-vertex whole brain cortical thickness differences between migraineurs with interictal photosensitivity and migraineurs without interictal photosensitivity are displayed on an average right medial (a) and left lateral (b) pial brain surface. The color red indicates greater cortical thickness in MwiPh relative to MwoiPh in right isthmus cingulate, pericalcarine and lingual regions and left precentral, postcentral and supramarginal regions. Significance thresholds were set at p < 0.025 and were cluster-corrected using Monte Carlo corrections (1000 permutations). Translucent colors show FreeSurfer-specific regional parcellations and labels were assigned according to FreeSurfer terminology. MwiPh: migraineurs with interictal photosensitivity; MwoiPh: migraineurs without interictal photosensitivity.
Next, using a region-of-interest post-hoc analysis, we calculated the average cortical thickness (in mm) over left and right cluster-corrected regions shown in Figure 1 (left fronto-parietal, right parietal-occipital regions) in order to also assess whether the average cortical thickness for these two cluster-corrected areas differed between MwiPh and healthy controls, between healthy controls and MwoiPh, and between MwiPh and MwoiPh. The rationale for this comparison was mainly illustrative and to demonstrate regional cortical thickness alterations in MwiPh and MwoiPh relative to a cohort of healthy controls.
For this comparison, we included only healthy controls without photosensitivity (n = 37) based on the same normative criteria for photosensitivity that were used for MwoiPh.
Results indicated no cortical thickness differences between MwoiPh and healthy controls for the left fronto-parietal region (controls = 2.57 mm, SD = 0.31; MwoiPh = 2.44 mm, SD = 0.28, p = 0.21) or right parietal-occipital region (controls = 1.76 mm, SD = 0.23; MwoiPh = 1.63 mm, SD = 0.28; p = 0.07) but significant group differences between MwiPh and healthy controls for the left fronto-parietal region (controls = 2.57 mm, SD = 0.31; MwiPh = 2.80 mm, SD = 0.33 p = 0.006) and the right parietal-occipital region (controls = 1.76 mm, SD = 0.23; MwiPh = 1.87 mm, SD = 0.22, p = 0.047). As expected, there were significant group differences between MwiPh and MwoiPh for the left fronto-parietal region (p < 0.0001) and the right parietal-occipital region (p = 0.001).
The robustness of the imaging software FreeSurfer has been established in prior publications, and data showed that brain segmentations and volumetric measurements are not affected by the use of different scanners (27). To ensure no scanner differences on volumetric measurements in our data, the estimated total intracranial volumes (eTIV) and total gray matter volumes were calculated for individuals imaged at one scanner (scanner 1: n = 48; MwiPh = 14; MwoiPh = 13; healthy controls = 21) versus the other scanner (scanner 2: n = 37; MwiPh = 14; MwoiPh = 7; healthy controls = 16). There were no significant differences in eTIV (scanner 1 = 1401 cm3 scanner 2 = 1462 cm3; p = 0.11) or total gray matter volume (scanner 1 = 601 cm3, scanner 2 = 619 cm3; p = 0.19) between images collected on the two scanners and there were no differences in the ratios of MwiPh and MwoiPh between both scanners (p = 0.38). In order to ensure that the use of two different scanners did not affect our findings, cortical thickness differences were estimated between individuals scanned on scanner 1 versus individuals scanned on scanner 2 (with “scanner” as a variable of interest) using a vertex-by-vertex general linear model design using the same significant thresholds as in the group comparisons. Results indicated no regions where cortical thickness differed between individuals scanned on scanner 1 versus individuals scanned on scanner 2, thus indicating that the use of two different scanners did not affect cortical thickness measurements.
Correlations between severity of photosensitivity and cortical thickness
In order to assess a potential relationship between the severity of photosensitivity and cortical thickness, correlations between photosensitivity scores and the cortical thickness of regions that showed increased cortical thickness in MwiPh were determined. No correlations were found between photosensitivity scores and the left precentral gyrus, left postcentral gyrus, left supramarginal region, right lingual gyrus and right pericalcarine region. However, a significant positive correlation was found between photosensitivity score and the right isthmus cingulate region (r = 0.43; p = 0.023) thus indicating that in MwiPh, higher scores of light sensitivity correlate with increased cortical thickness in the right isthmus cingulate.
Discussion
Photosensitivity and migraine
Photosensitivity during and between migraine attacks is a well-known and common characteristic of migraine. As such, several studies to date have investigated the presence and severity of light sensitivity in interictal migraine patients. For example, several studies reported lower discomfort thresholds to visual stimuli for migraine patients relative to controls (11,28,29). Other studies have used standardized questionnaires as proxy estimates for investigating visual or light sensitivity and a recently published study by Cucchiara and colleagues, (5) who used a self-report questionnaire (visual discomfort score) to assess levels of photosensitivity, reported increased photosensitivity in interictal migraine patients relative to controls. Similarly, Shepherd and colleagues reported increased visual discomfort in migraine patients with and without aura relative to controls using the Conlon et al. visual discomfort scale (29), and Mulleners et al. reported more photophobic symptoms in migraineurs with and without aura relative to controls using a short photophobia questionnaire (12). In the present study, we used the PAQ to assess the severity of photosensitivity in interictal migraine patients.
Our results corroborate previous findings and show that interictal episodic migraine patients have increased levels of photosensitivity relative to healthy control subjects. In addition, our results show that the severity of photosensitivity does not correlate with frequency of migraine attacks, years lived with migraine, or to levels of depression or anxiety (state and trait). These results suggest that psychological factors and disease history are unrelated to the severity of photosensitivity in migraine.
Cortical substrates of photosensitivity
Several studies have investigated the neural substrates of visual sensitivity using fMRI or PET imaging in healthy individuals (30) and in patients with migraine (2–6,31). Our study adds to the existing literature by investigating the association between self-reported photosensitivity and cortical structure. Although our results need to be validated using larger sample sizes, our results suggest that the presence of photosensitivity could be mediated by somatomotor and visual processing regions while the severity of visual discomfort might be mediated by affective regions. As such, we believe that our study is novel since it describes brain structural alterations in migraine patients with interictal photosensitivity compared to migraine patients without interictal photosensitivity and healthy controls without photosensitivity. Our results indicate that MwiPh have greater cortical thickness in the left precentral, postcentral and supramarginal regions and the right extrastriate regions including the isthmus cingulate, lingual and pericalcarine regions relative to MwoiPh. In addition, for MwiPh, higher scores of photosensitivity are associated with greater cortical thickness in the right isthmus cingulate region.
Additionally, the calculation of average cortical thickness estimates for left fronto-parietal and right parietal-occipital regions revealed that MwiPh had greater average cortical thickness for these bilateral regions compared to MwoiPh and healthy controls without photosensitivity. Yet, there were not significant differences in the average cortical thickness between MwoiPh and healthy controls without photosensitivity thus indicating that region-specific alterations in the left fronto-parietal and right parietal-occipital regions are specifically related to the presence of photosensitivity.
Recent fMRI data showed proprioceptive representation of the eye muscle to localize to the sensorimotor cortex (32) and Moulton and colleagues showed increased fMRI activation in the primary somatosensory cortex in a patient with corneal pain and photophobia (33). These data suggest that oculomotor proprioception and photophobic pain are related to primary somatosensory and primary motor activation. The supramarginal region is known to modulate multisensory processing including receiving input from visual and somatosensory areas. The supramarginal region also has reciprocal connections with the pulvinar (34,35), a thalamic region that has been implicated in migraine-related pain and allodynia (36,37).
The lingual and pericalcarine regions are both vision-related areas that are important for visual encoding (38,39) and memory storage for visual scenes (40).
The posterior cingulate is an integral region of the default mode network and is important for self-reflection and appraisal (41). Additionally, the posterior cingulate is commonly associated with visual-spatial perception and memory (42) as well as with orientation toward painful stimuli (43). In our study, MwiPh had greater isthmus (posterior cingulate) thickness that was related to symptom severity, thus indicating that photosensitivity modulates integrity of the isthmus cingulate cortex and potentially suggests that greater cortical thickness in the posterior cingulate region in MwiPh might underlie cognitive-affective components of photosensitivity.
Similar to our findings of greater cortical thickness in right parietal-occiptial and left fronto- parietal regions in MwiPh, PET imaging results indicate that photophobic patients had increased activation in the right occipital and precentral region during the premonitory phase of migraine (3), and patients suffering from “visual snow” have hypermetabolism in the right lingual gyrus (44). A correlation between visual sensitivity and primary visual cortex blood oxygenation level-dependent (BOLD) signal change was also reported in migraine patients with aura, yet not for migraine patients without aura (5). In addition, a recent study investigating healthy people who were hypersensitive to visual glare showed increased bilateral lingual and parietal (parietal lobule) activation (30), and a study investigating the relationship between functional activation patterns and self-described hypersensitivity showed increased right lingual activation in patients with fibromyalgia (45). Similar to our findings, other structural studies have found increased cortical thickness in somatosensory and visual regions for migraine patients relative to healthy controls (46–48). Although the precise link between cortical thickness and migraine symptomatology awaits further investigation, altered cortical thickness in these regions might be reflective of a “reinforcement” of visual-affective pathways in MwiPh. Whether this “strengthening” is specific to migraine or present in other disease populations with comorbid photosensitivity remains to be investigated.
While our results indicate that the presence of photosensitivity is associated with greater cortical thickness in bilateral parietal, left frontal areas, and right occipital brain regions, only the right isthmus cingulate region appears to directly correlate with the severity of photosensitivity by demonstrating increased cortical thickness in patients with more severe photosensitivity. Although our results need to be validated using larger sample sizes, our results suggest that the presence of photosensitivity could be mediated by somatomotor and visual processing regions while the severity of visual discomfort might be mediated by affective regions.
Limitations
In order to maintain an adequate sample size, patients with and without aura were not assessed as separate groups in this study. However, a post-hoc analysis of MwiPh who had aura symptoms compared to those without aura symptoms showed no difference in cortical thickness for regions that showed greater cortical thickness in MwiPh relative to MwoiPh. However, the potential interaction between aura symptoms and photosensitivity remains inconclusive and needs to be investigated further in future studies with larger sample sizes. Additionally, the percentage of patients with migraine aura was higher in the MwiPh group relative to the MwoiPh group although MwiPh who had aura symptoms did not have higher scores of photosensitivity relative to MwiPh without aura.
It is of importance to note that in this study photosensitivity was measured solely via self-report. Thus, it cannot be ascertained whether participants who reported photosensitivity also had abnormal visual discomfort thresholds. Future studies will need to establish the relationship between visual photometric measures and self-report questionnaires.
Summary
Episodic migraine patients with photosensitivity have greater cortical thickness than episodic migraine patients without photosensitivity in areas including the right lingual, pericalcarine, and isthmus cingulate region and the left precentral, postcentral and supramarginal region. In addition, more severe photosensitivity was associated with greater cortical thickness in the right isthmus cingulate, indicating that photosensitivity and brain structure might modulate each other although the direction of the interaction remains to be explored.
Key findings
Compared to interictal migraineurs without photosensitivity, migraineurs with self-reported symptoms of photosensitivity have greater cortical thickness in right parietal-occipital and left fronto-parietal regions. For migraineurs with photosensitivity, higher scores on the photosensitivity questionnaire are positively correlated with greater cortical thickness in the right isthmus cingulate region. Findings suggest that the presence of photosensitivity might be mediated by somatomotor and visual processing regions while the severity of photosensitivity might be mediated by affective regions.
Footnotes
Funding
This study was supported by a grant from the National Institutes of Health (K23NS070891) to TJS.
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Acknowledgments
We would like to express our gratitude to the MR technicians for their efforts and assistance in scheduling and scanning study participants, and we would like to thank all participants for their commitment and time dedication to this study.
All authors contributed significantly to the study design and were involved in writing and revising the manuscript. The published version of this manuscript was approved by all authors.
References
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.
