The authors have declared that no competing interests exist.
Conceived and designed the experiments: TS CC. Performed the experiments: TS CC. Analyzed the data: TS CC. Contributed reagents/materials/analysis tools: TS CC. Wrote the paper: TS CC.
Migraineurs have atypical pain processing, increased expectations for pain, and hypervigilance for pain. Recent studies identified correlations between brain structure and pain sensation in healthy adults. The objective of this study was to compare cortical thickness-to-pain threshold correlations in migraineurs to healthy controls. We hypothesized that migraineurs would have aberrant relationships between the anatomical neurocorrelates of pain processing and pain thresholds.
Pain thresholds to cutaneously applied heat were determined for 31 adult migraineurs and 32 healthy controls. Cortical thickness was determined from magnetic resonance imaging T1-weighted sequences. Regional cortical thickness-to-pain threshold correlations were determined for migraineurs and controls separately using a general linear model whole brain vertex-wise analysis. A pain threshold-by-group interaction analysis was then conducted to estimate regions where migraineurs show alterations in the pain threshold-to-cortical thickness correlations relative to healthy controls.
Controls had
Unlike healthy control subjects who have a significant negative correlation between cortical thickness in a superior temporal/inferior parietal region with pain thresholds, migraineurs have a non-significant positive correlation between cortical thickness in a superior temporal/inferior parietal region with pain thresholds. Since this region participates in orienting and attention to painful stimuli, absence of the normal correlation might represent a migraineurs inability to inhibit pain sensation via shifting attention away from the painful stimulus.
Migraineurs have atypical processing of sensory stimuli, including painful stimuli, during and between migraine attacks. During a migraine attack, the majority of migraineurs are hypersensitive to visual, auditory, olfactory and somatosensory stimuli, leading to photophobia, phonophobia, olfactory hypersensitivity, and cutaneous allodynia.
Imaging experiments investigating brain
After Institutional Review Board approval, 31 consenting adult migraine subjects and 32 healthy control subjects were enrolled into this study. Subjects were 18-65 years of age, did not have contraindications to MRI scanning, did not have contraindications to or disorders that might directly affect measurement of cutaneous pain thresholds (e.g. peripheral neuropathy), did not have acute or chronic pain disorders other than migraine, and did not have other neurological disorders. All subjects had normal brain MRI scans according to routine clinical interpretation. Migraine was diagnosed according to International Classification of Headache Disorders II diagnostic criteria.
All study testing was conducted during one visit. All subjects were free of pain for at least 48 hours and free of pain medications and migraine medications for at least 48 hours. Subjects first completed questionnaires, then had quantitative sensory testing (QST) for determination of pain thresholds, and then underwent magnetic resonance imaging (MRI).
Structured interviews, the State-Trait Anxiety Inventory (STAI, Form Y-1 and Form Y-2), the Beck Depression Inventory (BDI-II), and the Allodynia Symptom Checklist 12 were utilized for determining the presence or absence of migraine, migraine frequency, number of years with migraine, presence of migraine aura, anxiety scores, depression scores, and cutaneous allodynia severity.
QST was used to determine pain thresholds to heat applied to the skin using a 30 mm × 30 mm thermode attached to a Medoc Pathway QST machine. Left ventral medial forearm pain thresholds were determined using the method of limits with a temperature increase of 1°C per second. Testing was performed 3 times and the mean of the 3 tests was considered the pain threshold.
All neuroimaging was performed on a Siemens MAGNETOM Trio 3 tesla scanner (Erlangen, Germany) with total imaging matrix technology using a 12-channel head matrix coil and FDA-approved sequences. Structural scans included a high-resolution T1-weighted sagittal magnetization-prepared rapid gradient echo (MP-RAGE) (TE = 3.16 ms, TR = 2.4 s, flip angle = 8°, 176 slices, 1 × 1 × 1 mm voxels, field-of-view 256 × 256 mm2) and an axial T2-weighted turbo spin echo (TE = 88 ms, TR = 6280 ms, flip angle = 120°, 36 slices, 1 × 1 × 4 mm voxels, field-of-view 256 × 256 mm2).
For cortical reconstruction and segmentation, T1-weighted MP-RAGE sequences were processed using FreeSurfer (
Subject demographics, pain thresholds, total cortical thickness, cutaneous allodynia severity scores, anxiety and depression scores were reported using descriptive statistics and compared amongst subject groups using independent sample t-tests (two-tailed) or chi-squared tests, as appropriate. Cortical thickness-to-pain threshold correlations were analyzed separately for migraine subjects and for control subjects. Correlations were analyzed by applying a general linear model (GLM) whole brain vertex-wise analysis within FreeSurfer. Age, Beck Depression Inventory scores, trait and state anxiety scores, and estimated total intracranial volume were entered into the GLM as covariates of no interest. Cortical thickness-to-pain threshold correlations surviving p<.01 uncorrected and those surviving Monte Carlo cluster correction of p<.01 were determined for migraine and control groups separately.
In a post-hoc analysis, the mean cortical thickness of regions that had correlations with pain thresholds that differed in migraineurs vs. controls and survived Monte Carlo cluster correction with p<0.01 were calculated and exported into SPSS 22.0 (SPSS Inc., Chicago, IL USA). Mean cortical thickness of these regions in migraineurs was compared to mean cortical thickness in controls using independent sample t-tests. Correlations between cortical thickness in these regions and headache frequency, number of years with migraine, and allodynia scores were calculated for the migraineurs. In these post-hoc analyses, comparisons and correlations with p<.05 were considered significant.
Migraine (n = 31) | Control (n = 32) | p-value | |
34.9 (12.1) | 35.3 (11.6) | .88 | |
26/5 | 25/7 | .57 | |
25.4 (6.1) | 25.2 (5.8) | .88 | |
31.0 (9.9) | 29.4 (9.3) | .51 | |
4.8 (5.0) | 2.8 (4.9) | .12 | |
.52 (1.6) | .13 (.34) | .19 | |
43.5°C (3.7) | 43.7°C (3.4) | .81 | |
7.8 (5.8) | NA | NA | |
16.0 (9.2) | NA | NA | |
10/21 | NA | NA |
All continuous variables are reported as means with standard deviation in parentheses. Categorical variables are reported as counts. State and Trait anxiety was determined via the Spielberger State-Trait Anxiety questionnaire. Depression scores were determined via the Beck Depression Inventory (BDI-II). Allodynia scores were determined via the Allodynia Symptom Checklist 12.
Average age was 34.9 years (+/− 12.1) in the migraineurs (n = 31) and 35.3 years (+/− 11.6) in the controls (n = 32). Approximately 80% of all subjects were women, consistent with the predominance of migraine amongst women. There were no significant differences in age, gender distribution, anxiety scores, or depression scores between the migraine and control groups. Mean pain threshold and allodynia symptom severity scores did not differ between the two subject cohorts. There were no significant differences between groups for total cortical thickness over both hemispheres (migraineurs 2.48 mm vs. controls 2.51 mm, p = .35). Migraineurs averaged 7.8 days with headache per month (+/− 5.8) and had migraine for 16 years (+/− 9.2). Ten migraineurs (32.3% of all migraineurs) had migraine with aura attacks.
Correlation between the cortical thickness of a left superior temporal/inferior parietal region with pain thresholds differed in healthy control subjects vs. migraine subjects. Healthy control subjects demonstrated the expected negative correlation between pain thresholds and cortical thickness in this region while migraineurs lacked this negative correlation.
Cortical Region | Region Size (mm2) | X | Y | Z | |
Left Superior Temporal/Inferior Parietal | 412 | −43 | −39 | 12 | |
Right Precuneus | 47 | 7 | −64 | 55 | |
Right Superior Temporal/Inferior Parietal | 39 | 60 | −42 | 16 | |
Left Inferior Parietal | 14 | −43 | −54 | 10 | |
Left Posterior Cingulate/Precuneus | 224 | −17 | −29 | 37 | |
Right Superior Temporal | 176 | 59 | −35 | 6 | |
Right Inferior Parietal | 152 | 32 | −77 | 16 | |
Left Inferior Temporal | 66 | −52 | −34 | −19 | |
Left Inferior Temporal | 10 | −44 | −42 | −17 | |
Right Superior Temporal/Inferior Parietal | 484 | 61 | −40 | 15 | |
Right Precentral | 248 | 58 | 0 | 33 | |
Left Posterior Cingulate/Precuneus | 105 | −11 | −37 | 41 | |
Right Inferior Parietal | 8 | 32 | −74 | 17 |
Correlations between cortical thickness and pain thresholds in migraineurs and controls are listed, as well as those correlations that differed when comparing the migraineurs to controls. All correlations in this table had uncorrected p<.01, while those bolded and starred survived Monte Carlo correction (p<.01). X, Y, and Z coordinates are based on the Talairach and Tournoux Atlas.
In healthy controls there were
In migraineurs, there were
Cortical thickness to pain threshold correlations differed (p<.01) between migraineurs and controls for left superior temporal/inferior parietal, right superior temporal/inferior parietal, right precentral, left posterior cingulate/precuneus, and right inferior parietal regions. Cortical thickness to pain threshold correlations with the left superior temporal/inferior parietal region were the only correlations that significantly differed between migraineurs and controls after Monte Carlo correction with p<.01. For this region, healthy controls had a significant negative correlation between cortical thickness and pain thresholds while migraineurs had a non-significant positive correlation between cortical thickness and pain thresholds.
Although cortical thickness in this superior temporal/inferior parietal cortex was numerically thinner in migraineurs compared to controls, there was no significant difference in cortical thickness of this region between subject cohorts (2.35 mm +/− 0.39 mm in migraineurs vs. 2.52 mm +/− 0.32 mm in controls, p = .071).
Amongst the migraineurs, there were no significant correlations between cortical thickness in the left superior temporal/inferior parietal region and headache frequency, number of years with migraine, or allodynia symptom severity.
The main finding of this study is that migraineurs lacked the normal correlation between cortical thickness of a left superior temporal/inferior parietal region and cutaneous pain thresholds (migraineurs had a non-significant positive correlation), while healthy controls demonstrated the expected negative correlation. Amongst migraineurs, superior temporal/inferior parietal cortical thickness did not correlate with headache frequency, number of years with migraine, or allodynia symptom severity suggesting that the lack of cortical thickness-to-pain threshold correlation might be an underlying trait of the migraine brain as opposed to being a result of recurrent migraine attacks.
A study by Erpelding and colleagues published in 2012 was reportedly the first to investigate correlations between cortical thickness and thermal pain thresholds.
Correlations between pain sensitivity (i.e. individual subject pain intensity ratings in response to 49 degrees Celsius stimulation of the skin) and regional gray matter density were studied in 116 healthy volunteers.
The underlying mechanism by which cortical thickness correlates with pain thresholds is yet to be completely understood. However, several studies have found associations between the presence of pain disorders and atypical cortical thickness, including several migraine studies.
The inferior parietal and superior temporal lobes participate in cognitive aspects of processing sensory stimuli, including pain.
Abnormal structure and function of superior temporal and inferior parietal regions have previously been demonstrated in migraineurs. A voxel-based morphometry study found that migraineurs have lower gray matter volume in the right superior temporal gyrus with extension to the parietal operculum compared to controls.
A lack of correlation between pain thresholds and superior temporal/inferior parietal cortex cortical thickness found in our study and the atypical structure and function of these regions found in other migraine studies could represent a lack of the normal attention orienting role of the superior temporal/inferior parietal cortex in pain processing and a resulting inability to adequately direct attention away from pain. Since distraction from painful stimuli is a powerful method of reducing the pain experience, inability to direct attention away from pain could associate with migraineurs having higher ratings of pain intensity and pain unpleasantness, somatosensory amplification, and hypervigilance. Prior work has shown that migraineurs have enhanced expectations for pain and hypervigilance regarding pain.
Since cortical thickness might be fluid, changing over time in response to pain and other factors, the cross-sectional design of our study is a limitation. Longitudinal studies that correlate changes in migraine patterns, pain thresholds, and cortical thickness would help to better define the direction of the relationship between cortical thickness, migraine, and pain thresholds. Although the cortical thickness of the superior temporal/inferior parietal region was numerically thinner in migraineurs compared to controls, there was no significant difference in cortical thickness of this region between subject cohorts. Cortical thickness differences between migraineurs and controls might be identified for this region if this study had a larger sample size and if the analysis accounted for age.
In this study, migraineurs had absence of the normal pain threshold-to-cortical thickness correlation in a region of the superior temporal and inferior parietal cortex. Since this region is typically involved in attention and orienting to sensory stimuli, loss of such a correlation might imply that the migraine brain has less ability to re-orient attention away from painful stimuli. Since distraction from painful stimuli lessens pain sensation, lack of ability to distract oneself from pain could lead to an enhancement of the pain experience for migraineurs.
Researchers wishing to access our data should send their request via e-mail to the corresponding author (