Association between pain sensitivity and gray matter properties in the sensorimotor network in women with irritable bowel syndrome

Abstract Background Enhanced perception of visceral stimuli is an important feature of Irritable Bowel Syndrome (IBS), but it is not known whether visceral sensitivity is associated with regional structural brain properties in IBS. Methods Structural brain magnetic resonance imaging data from 216 women with IBS and 138 healthy women were parcellated with FreeSurfer to define regional gray matter morphometry (volume, cortical thickness, surface area and mean curvature) in the sensorimotor network. General linear models were used to detect group differences between IBS and health. In a second set of 48 female IBS patients, pain threshold, pain intensity ratings during rectal balloon distension, and reported levels of abdominal pain and bloating were correlated with brain regions that showed differences between IBS and health in the first data set. Key Results Several statistically significant differences between IBS patients and healthy controls were found, mainly higher gray matter volume and cortical thickness in primary somatosensory cortex, secondary somatosensory cortex, and subcortical regions, and lesser gray matter volume, surface area and cortical thickness in posterior insula and superior frontal gyrus. Pain intensity ratings during rectal distension were associated with left primary somatosensory cortical thickness, and pain threshold was associated with right nucleus accumbens volume. Conclusions and Inferences Regional gray matter differences in sensorimotor network are associated with visceral sensitivity and may represent neuroplastic changes in female IBS patients.


| INTRODUC TI ON
Irritable bowel syndrome (IBS) is one of the most common disorders of gut-brain interaction. 1 It is characterized by chronic, recurrent abdominal pain, and altered bowel habits 2 and is often accompanied by comorbid psychiatric and other chronic pain disorders. 3 Visceral hypersensitivity has consistently been identified as an important feature of IBS. 4 Dysregulation of the braingut axis 5 and in particular aberrant central sensory processing [6][7][8] have been found to be contributing mechanisms in IBS, although they remain incompletely understood. Among the proposed central networks relevant to IBS pathophysiology, the sensorimotor network has shown structural and functional differences between healthy controls (HCs) and IBS patients across several studies. [9][10][11] The sensorimotor network includes primary and secondary sensorimotor cortex, primary and supplementary motor cortex, posterior insula, basal ganglia, and thalamus. 9 Higher gray matter volume of the primary somatosensory cortex has been reported in IBS, 12 as well as in conditions such as interstitial cystitis and painful bladder syndrome. 13 Furthermore, the literature supports positive associations of cortical thickness in the primary sensorimotor regions with gastrointestinal symptom severity, whereas right anterior and middle insular cortex has shown negative associations with anxiety in women with IBS. 14 While evoked visceral pain has shown associations with regional gray matter properties in HCs 15 and resting state functional connectivity in IBS, 11 correlations of evoked visceral pain using rectal balloon distension and regional gray matter morphometric properties in IBS have not been reported previously.
In the current study, we aimed to identify the relationship between sensorimotor network gray matter morphometry and evoked and spontaneous pain measures in women with IBS.
First, we assessed differences in the gray matter properties of sensorimotor network between IBS and HCs in a large collection of subjects recruited from different studies. Thereafter, we examined correlations of these regional brain measurements in a second set of women with IBS with two types of pain assessments: experimentally evoked rectal pain and reported symptom severity.
Our hypothesis was that evoked pain, spontaneous symptoms of sensory origin, and psychological distress would show different associations with gray matter morphometry in regions of the sensorimotor network in IBS. More specifically, we hypothesized that symptoms of abdominal pain and bloating would show correlations with primary motor cortex and primary somatosensory cortex 14 ; visceral sensitivity measures during rectal barostat would show correlations with posterior insula 11 ; psychological distress would show no associations as it was previously shown to be more relevant to emotional-arousal, salience, and central autonomic networks than the sensorimotor network. 8 2 | ME THODS AND MATERIAL S 2.1 | Dataset 1: Analysis of a large archived brain imaging dataset 2.1.1 | Subjects and brain imaging Structural brain images from IBS and HCs were obtained from the NIH-funded Pain and Interoception Imaging Network (PAIN). 16 Parts of the dataset have been studied and published previously, references 10,12,14 are of relevance to the present study. The brain images were acquired using a 3.0 T Siemens Trio MRI (Siemens Healthineers AG, Erlangen, Germany) at the G. Oppenheimer Center for Neurobiology of Stress and Resilience at the University of California, Los Angeles between 2006 and 2017. IBS diagnoses were based on Rome III or Rome IV criteria depending on time of inclusion. Structural scans were obtained from six different acquisition sequences using a high-resolution 3-dimensional T1-weighted, sagittal magnetization-prepared rapid gradient echo protocol as described in Table 1. Acquisition protocols were only included if they were used with both IBS and HCs. A general linear model controlling for age indicated that protocol 1 and 2 were similar to each other but had lower total gray matter volumes than the remaining protocols. Fisher's two-sided exact test indicated that the distribution

Key points
• Visceral sensitivity is of importance for IBS, and regional gray matter differences between IBS and health have been shown. There is no previous study investigating possible regional gray matter morphometry associations with visceral sensitivity in IBS.
• Right nucleus accumbens volume correlated with pain threshold, and left primary somatosensory cortical thickness correlated with pain intensity ratings during rectal distension.
• Regional gray matter differences in sensorimotor network were partly associated with visceral sensitivity and may represent neuroplastic changes in female IBS patients. we controlled for protocol in the subsequent statistical analyses.

| Structural MRI analysis
After quality control, FreeSurfer 17,18 was used for brain segmenta-

| Regions of interest (ROIs)
Regions comprising the sensorimotor network includes thalamus, basal ganglia the posterior insula, and the primary and secondary sensorimotor cortices. 9 In total, 34 regions of interest (17 ROIs, 2 hemispheres) were examined (see Table 2).

| Statistical analysis of Dataset 1
Using Matlab, R2017b, contrast analysis within the framework of the general linear model (GLM) was used to detect group differences in gray matter morphometry between IBS and HCs, controlling for age, total intracranial volume, and acquisition protocol. ROIs demonstrating significant group differences (p < 0.05, Bonferroni corrected for the number of ROIs tested) were used as ROIs for correlational analysis in a second independent data, Dataset 2.

| Rectal barostat procedure
The rectal barostat protocol described by Cremonini et al 23

| Symptom questionnaires
The clinical symptoms of IBS were measured with the IBS severity scoring system (IBS-SSS) 24 and the Gastrointestinal Symptom Rating Scale for IBS (GSRS-IBS). 25 In this study, we used IBS-SSS only descriptively, that is to characterize the overall severity in the IBS population. The GSRS-IBS is a validated questionnaire of gastrointestinal symptom severity in IBS consisting of 13 questions, divided into 5 domains: abdominal pain, bloating, diarrhea, constipation, and satiety. For the present study, we focus specifically on abdominal pain and bloating, since these presumably are lower GI symptoms of sensory origin.
The Hospital Anxiety and Depression scale (HADS) is a questionnaire used to screen for symptoms of depression and anxiety, and assess their severity. The score is usually calculated separately for anxiety and depression, and higher scores reflect more severe symptoms. 26 The separation of the subscales has been questioned, 27 and in this study, we used total HADS score as a measurement of overall psychological distress.

| Brain imaging acquisition
The brain images were acquired on a 3 Tesla Philips Achieva (Philips Healthcare, Best, the Netherlands) using the standard 8 channel head coil. For the high-resolution scan, a T1-weighted 3D TFE gradient echo scan, that is, a magnetization-prepared rapid acquisition gradient echo (MP-RAGE) sequence was used, with TR = 7.0 ms, TE = 3.2 ms, flip angle = 9 degrees, and a voxel size of 1 × 1 × 1 mm 3 .
The same quality control and structural MRI analysis as described for Dataset 1 was used also in Dataset 2 (see sections 2.1.2 for details).

| Statistical analysis
IBM SPSS statistics version 24 was used for all analyses of Dataset 2. Correlations between regional brain measurements and clinical pain threshold and pain intensity ratings; HADS total score) were examined. Partial correlations were run, controlling for age and total gray matter volume (TGMV). Symptoms of pain, bloating, and HADS score were normally distributed, and parametric partial correlations were used. Rectal pain threshold was positively skewed, but the distribution could be normalized using a natural logarithmic transformation, the transformed values were used for the partial correlations. The pain ratings during 24 mm Hg rectal distension were non-normally distributed and could not be normalized after transformation; hence, non-parametric partial correlations were used for the correlational analyses for this measurement. The significance level for these analyses was set to p < 0.05 Bonferroni corrected for the five sets of analyses performed.

| Dataset 1: Subjects
The analyzed population consisted of 138 healthy women and 216 women with IBS. There was no significant difference in age between the two groups (28.

| Dataset 2: Subjects
We included 55 female IBS patients. Of these, 7 were excluded because of pathologies found on MRI scan (n = 1), another GI disease discovered during the study (n = 2), drop-out (n = 3; completed the brain imaging part but not the rectal barostat testing), and use of probiotics (n = 1). No subject was excluded due to poor image quality.
The analyzed population hence consisted of 48 female IBS patients, see Table 3

| Dataset 1: Group differences between IBS and HCs
As shown in Table 4 and Figure 2, differences between IBS and HCs were found in several regions of the sensorimotor network.

| Dataset 2: Correlations between sensorimotor regions and rectal barostat measurements
In IBS patients, increased volume of right nucleus accumbens was

| Dataset 2: Correlations between sensorimotor regions and reported symptoms
In IBS patients, there were no correlations between brain morphom-

| DISCUSS ION
Differences in regions of the sensorimotor network were identified in the biggest cohort published to date comparing regional gray matter properties between female IBS patients and HCs.
Higher gray matter measurements were observed in primary

| Gray matter morphometry measurements as indicators of neuroplasticity
Neuroplasticity refers to structural and functional changes in neural circuits in response to experience. 28,29 Experiences and learning can cause regional gray matter alterations identifiable with MRI. 29,30 There is mounting evidence for structural plasticity and reorganisation in human chronic pain in general, [31][32][33][34][35][36] as well as in IBS specifically. 10,12,[37][38][39][40] The cellular correlates that underlie the macroscopically detected gray matter plastic changes are so far incompletely understood. They are thought to be composed of a combination of adult neurogenesis (for certain regions such as the hippocampus), synaptic changes (such as changes in dendritic length and branching or in the number of dendritic spines per neuron), and changes in the number and morphology of glial cells. 30 It could also reflect a change in cell sizes, as well as changes in blood flow or interstitial fluid. 29 In one study of fibromyalgia using multimodal neuroimaging, gray matter decrease was not evidently related to neurodegeneration, but possibly to reduced blood flow, whereas increased gray matter volume was F I G U R E 2 Group differences between IBS patients and HCs in dataset 1. Regions larger in IBS than HCs are shown in red, and regions smaller in IBS than HCs are shown in blue F I G U R E 3 Association between the rectal pain threshold and the volume of right nucleus accumbens. Partial regression scatter plot illustrating the association between the rectal pain threshold transformed using natural logarithm and the volume of nucleus accumbens, controlled for age and TGMV. The x-axis represents the residuals from predicting pain threshold from age and TGMV. The y-axis represents the residuals from predicting right nucleus accumbens volume from age and TGMV associated with increased GABA A -receptor concentration and water content, indicative of neural plasticity. 41 The correlations with visceral sensitivity in this study were in regions with increased gray matter in IBS, which using the results of Pomares et al 41 supports the interpretation of neural plasticity underlying these findings.

| Group differences between IBS and HCs
The major group differences between health and IBS in this study were gray matter increases in primary somatosensory cortex, secondary somatosensory cortex and subcortical regions, as well as decreases in posterior insula and superior frontal gyrus in IBS patients (see Table 4). Postcentral gyrus has in several studies been described as being larger in IBS than health, 10,12,14 and posterior insula smaller in IBS. 12,14,37 Three of these studies were done using parts of the same cohort as the present study. 10,12,14 However, this study is larger than previous studies, combining several cohorts and therefore contributes with increased robustness to the previous findings.

| Primary somatosensory cortex
Primary somatosensory cortex (S1) has a role in pain processing, in particular regarding the localization of the stimuli, 42 and it has been argued that neuronal activity in primary somatosensory cortex participates in producing awareness of the sensory/discriminative aspects of pain. 43 There are also indications of primary somatosensory F I G U R E 4 Association between pain ratings during rectal distension and cortical thickness of left postcentral gyrus. Partial regression scatter plot illustrating the association between the pain ratings at 24 mm Hg above BOP rectal distension and the cortical thickness of the left postcentral gyrus, controlled for age and TGMV. The x-axis represents the residuals from predicting pain intensity ratings at 24 mm Hg from age and TGMV. The y-axis represents the residuals from predicting left postcentral gyrus cortical thickness from age and TGMV cortex being involved in modulation of the affective and attentional component of pain. 44 Visceral stimuli, both innocuous and noxious, have been found to elicit activation in primary somatosensory cortex in HC, as well as in IBS subjects. 45 In this study, IBS patients had increased cortical thickness of the primary somatosensory cortex compared to healthy controls, and we found an association between left primary somatosensory cortex thickness and rectal pain intensity ratings during rectal distension. This is in line with the notion that gray matter increase in primary somatosensory cortex is likely related to repeated exposure to pain. 46

| Nucleus accumbens
The nucleus accumbens is important for motivational processes, reward, and avoidance, directing attention and behavior toward a goal or away from a threat. 47 Nucleus accumbens is also involved in the transition from acute to chronic pain. 48,49 The increased volume of right nucleus accumbens in IBS shown in this study has not been reported before. In the second data set, the right nucleus accumbens was associated with rectal pain sensitivity, as well as reported severity of abdominal pain and bloating (albeit the latter two not significant after correction for multiple com-

| Associations between reported symptoms and brain structure in IBS
Contrary to previous studies 12, 14 and our hypothesis, we did not prefrontal cortex. 50 The lack of associations in the current study supports the notion that the results regarding visceral sensitivity is not related to psychological distress, but represents another pathophysiological mechanism evident at the central nervous system level, possibly neuroplasticity related to the experience of recurrent visceral pain.

| Limitations
Hypersensitivity is present only in a portion of patients with IBS. 51,52 In this study, we did not differentiate IBS patients as normosensitive or hypersensitive, and it is likely that the associations between regional brain morphometry measures and visceral sensitivity measures would be different between these IBS subgroups. Future studies investigating correlations between visceral sensitivity and structural brain properties should consider taking the sensitivity status into account.
We did not include early life trauma or psychological distress in the group comparison, which in previous studies have been shown to influence the results. 12

| CON CLUS ION
This study consolidates previous findings of gray matter alterations sexes, both IBS and HCs using the same visceral sensitivity testing protocol, and assess additional networks to obtain a more complete picture of these associations relevant for long-standing GI symptoms.