For the purposes of the present analyses, we used data from a total of N = 96 adult volunteers (N = 31 UC, N = 23 IBS, N = 44 healthy controls), acquired within two comprehensive visceral pain studies conducted by our group between the years 2015 and 2020 at the University Hospital Essen, Germany. Primary studies involved different emotional learning/memory tasks (data will be presented elsewhere), all accomplished subsequent to the acquisition of data analyzed herein. Importantly, all participants underwent sociodemographic, psychological and clinical characterization and structural magnetic resonance imaging (MRI) prior to other experimental manipulations. Highly standardized and parallelized procedures were implemented for recruitment, screening and all other assessments that are part of this report, all accomplished within the same biomedical research setting using the same MR scanner. Work was conducted in accordance with The Declaration of Helsinki, and studies were approved by the local Ethics Committee of the University Hospital Essen (protocol numbers. 10-4493; 16-7237). All volunteers gave written informed consent and received monetary compensation for participation.

The screening process consisted of a standardized telephone screening, followed by an on-site visit with study staff, and completion of questionnaires (for details on questionnaires, see below). General exclusion criteria for all participants included age <18 or >65 years, body mass index <18 or >30, MRI-specific criteria like claustrophobia, pregnancy or ferromagnetic implants, and any evidence of structural brain abnormalities, verified by a neuroradiologist (author NT). Pregnancy was ruled out using a commercially available pregnancy test on the day of the MRI (Biorepair GmbH, Sinsheim, Germany, sensitivity 10 mIU/ml). For healthy controls, additional exclusion criteria included any known somatic or mental health condition, clinically-relevant anxiety or depression symptoms based on Hospital Anxiety and Depression Scale (HADS), or regular use of medications (except hormonal contraceptives, hormone replacement therapy, thyroid medication, irregular over-the-counter non-prescription drugs). For the UC group, only patients in clinical remission or with very low ongoing disease activity were included to avoid interference of active disease with study-related procedures, and to minimize putative effects of acute inflammation (or medical treatments required during phases of disease exacerbation) on study-related measures of interest acquired herein. Clinical disease activity was assessed based on symptom reports, initially evaluated in a structured screening interview, and then quantified with the Clinical Colitis Activity Index [CAI; (40)]. In addition, levels of fecal calprotectin were quantified, providing a non-invasive marker of intestinal inflammation (41), with an established reliable cut-off value indicating clinical remission below 150 μg (42). Treatment with systemic glucocorticoids within 4 weeks of the study were exclusionary. Other concomitant medications, which were continued as prescribed by the treating physician, were recorded. For IBS, symptom-based confirmation of diagnostic criteria was based on ROME IV criteria (43). Regular prescribed or non-prescribed IBS-related medications including low-dose treatment with antidepressants were not discontinued for the study. While minor and stable (or successfully treated) psychological symptoms, such as mild anxiety or depression symptoms (including elevated HADS scores) were not exclusionary, patients with diagnosed, more severe psychiatric comorbidities were excluded. Note that given frequent reporting of additional extraintestinal pain symptoms in IBS and IBD (44, 45), patients who reported such symptoms in addition to symptoms of their primary GI diagnoses were not excluded, but other types of chronic or recurring pain symptoms and chronic pain diagnoses were recorded. For all patients, an existing and confirmed diagnosis (of the respective GI disorder) established at least 1 year prior to recruitment for this study was required.

In all participants, GI symptoms were quantified with a standardized questionnaire that we routinely use in our group as it is applicable across visceral pain conditions as well as in healthy volunteers [who also experience such symptoms, albeit less frequently or intensely; (46)]. A range of typical GI symptoms (i.e., diarrhea, constipation, vomiting, nausea, lower abdominal pain, upper abdominal pain, heartburn, post-prandial fullness, bloating, loss of appetite) in the previous 3 months is assessed using a Likert-type response scale (0 = experience never, 1 = experience once or twice per month, 2 = experience once or twice per week, 3 = experience more than twice a week). The total sum score was calculated for analyses. Given the specific interest in visceral pain herein, individual responses on the items for upper and lower abdominal pain, respectively, are additionally provided for a more specific characterization of GI symptoms in each group (Table 1). For patients with IBS, current bowel alteration(s) and bowel symptom subtyping (i.e., diarrhea-predominant, constipation-predominant, mixed) were also accomplished based on the GI symptom questionnaire. Patients with UC additionally completed the CAI (40) to assess clinical disease activity. The CAI consists of 6 items capturing a range of typical UC symptoms (i.e., increase in stool frequency, bloody stools, abdominal pain, temperature due to colitis, extraintestinal manifestations, and the investigator's global assessment of symptomatic state) as well as 1 item concerning laboratory results (i.e., erythrocyte sedimentation rate and hemoglobin). Hemoglobin is relevant, as anemia is the most common complication in IBD associated with disease activity, and the erythrocyte sedimentation rate is a biomarker of inflammation. Based on the total sum score, disease activity can be classified into inactive (i.e., remission; ≤ 4), mild activity (5–10), moderate activity (11–17) and high activity (≥18) with a maximum score of 26 (47). However, laboratory assessments were not available for the entire sample of UC patients (missing for N = 13 patients), which is why we provide CAI average scores computed based on 6 items for all patients for consistency. In results, we refer to this measure as symptom-based CAI for clarity.

Chronic stress was assessed by the 12-item screening scale of the Trier Inventory of Chronic Stress [TICS-SSCS; (48)]. The scale evaluates individual experiences with chronic stressors in everyday life and provides a reliable global measure of perceived stress during the last 3 months (49). Likert-scale response options are “never” (0), “rarely” (1), “sometimes” (2), “often” (3), and “very often” (4), with a total score ranging from 0 to 48, and higher scores indicating greater perceived presence and frequency of chronic stressors. Note that we chose this questionnaire specifically for its applicability not only to research in clinical populations but also in healthy volunteers, expanding on our early work on the role of chronic stress in the context of visceral pain (50).

In addition, the Hospital Anxiety and Depression Scale [HADS; (51)] was used as screening tool, and to provide a clinically-relevant and widely-used measure suitable for a characterization of patient groups with respect to psychological distress. The HADS consists of two subscales with 7 items measuring anxiety (HADS_A) and depression (HADS_D), respectively. For each subscale, available cut-off values differentiate between non-cases (subscale score <8), potential cases (subscale score 8–10), and probable cases (subscale score ≥11) of anxiety and depression (52). For the purposes of sample characterization, in addition to the two subscale scores, we provide mean total scores (HADS Total), which can range from 0 to 42 with higher scores indicating higher levels of overall psychological distress.

All questionnaire data and other self-report variables were analyzed using IBM SPSS Statistics 27 (IBM Corporation, Armonk, NY). Group comparisons were accomplished using independent-samples t-tests, and data are reported as mean ± standard deviation, unless indicated otherwise.

Structural images were acquired on a 3 Tesla MR scanner using a 32-channel head coil (Skyra, Siemens Healthcare, Erlangen, Germany). All data were acquired on the identical scanner, and used one of two 3D-MPRage T1-weighted sequences with very similar yet not identical acquisition parameters: sequence 1 [repetition time (TR) 1,900 ms, echo time (TE) 2.13 ms, flip angle 9°, field of view (FOV) 239 × 239 mm², voxel size 0.9 × 0.9 × 0.9 mm³]; sequence 2 [TR 1,770 ms, TE 3.24 ms, flip angle 8°, FOV 256 × 256 mm², voxel size 1 × 1 × 1 mm³]. All group analyses were accomplished after a matching of healthy controls (based on the entire sample of N = 44) to each individual patient group, providing dedicated control groups for UC and IBS, respectively, referred to subsequently as HC_UC and HC_IBS. The matching procedure was based on MR scanning sequence and age, accounting for the slightly different acquisition parameters of the two sequences. Note that the data included for analyses of IBS vs. HC_IBS were all acquired with sequence 1; analyses of UC vs. HC_UC had equal number of patients and healthy controls measured with sequence 1 (N = 13 UC, N = 13 HC_UC) and 2 (N = 18 UC, N = 18 HC_UC). The acquired images were pre-processed and analyzed with the CAT12 toolbox (Structural Brain Mapping group, Jena University Hospital, Jena, Germany) and SPM12 (Statistical Parametric Mapping, Wellcome Center for Human Neuroimaging, UCL Queen Square Institute of Neurology, London, UK) implemented in Matlab R2020a (MathWorks Inc., Natick, MA, USA). The analysis followed the standard protocol for this toolbox (http://www.neuro.uni-jena.de/cat12/CAT12-Manual.pdf) using default settings and parameters, unless otherwise specified. The main processing steps included the segmentation of voxels into gray matter (GM), white matter (WM) and cerebrospinal fluid (CSF), and normalization using optimized shooting registration. After pre-processing, the homogeneity of the sample was checked by inspecting the correlation between all volumes to ensure data quality. As all images showed high correlation values (>0.86), no images were excluded from further analysis. Modulated normalized GM maps were smoothed with a Gaussian kernel of 8 mm (FWHM). The smoothed images were used for further analysis to test for regional GMV differences between groups. Atlas labeling was based on MRI scans originating from the Open Access Series of Imaging Studies (OASIS) project. The labeled data were provided by Neuromorphometrics under academic subscription (Neuromorphometrics, Inc., Somerville, MA, USA). The total intracranial volume was determined for each subject, as it is an important confounding variable in voxel-based morphometry.

All whole-brain statistical analyses were performed within the CAT12 environment. To increase sensitivity and to avoid the arbitrary choice of an initial cluster-forming threshold, the Threshold Free Cluster Enhancement [TFCE; (53, 54)] toolbox was used for all analyses (Structural Brain Mapping group, Jena University Hospital, Jena, Germany). In a first step, two analyses of covariance (ANCOVA) were run to compare the GMV between each patient group and matched healthy controls with total intracranial volume and age as covariates of no interest. Note that T1-sequence was additionally included as covariate of no interest for the comparison of UC vs. HC_UC. For all ANCOVAs, we report significant results corrected for multiple comparisons [using family-wise error (FWE) correction of alpha, set at p < 0.05].

In a second step, four multiple regressions were calculated (i.e., two for each patient group vs. controls) to test for group interactions in the correlation of GMV with GI symptoms and chronic stress, respectively, controlling for total intracranial volume, age (and sequence where appropriate) as covariates of no interest. For these analyses, we report FWE-corrected results as well as results without correction applying an alpha level of p < 0.001. For each cluster identified by multiple regression analysis, the estimated GMV of its peak brain region was extracted and transferred to SPSS. As exploratory analysis, we examined correlations of GMV within brain regions identified by multiple regression analyses and GI symptoms and chronic stress, respectively, within each group using partial correlation analyses. To this end, the extracted tissue volumes within anatomical regions and GI symptoms and chronic stress were regressed based on total intracranial volume, age, and sequence (where appropriate). Correlational analyses were then accomplished and plotted using RStudio (version 1.2.5001, RStudio PBC).

Supplemental analyses were carried out as follows (all results reported within Supplementary Material): Firstly, as sex was not equally distributed in patients with UC and HC_UC, all analyses were re-computed in a subsample comprising only women, i.e., after exclusion of 5 male patients and their 5 age-matched female controls. Secondly, to indirectly address whether patterns of GMV alterations in patients with UC and IBS are disease-specific, further data and results are provided (details on approach provided in Supplementary Material). The approach included extracting and plotting GMV of the clusters identified in the comparison of one patient group and matched healthy controls in the other patient group and matched healthy controls, and using these clusters as regions of interest (ROI) in ROI-based analyses.

As per matching of healthy controls to patient group based on age and T1-scanning sequence, the final samples we report upon consisted of N = 31 UC vs. N = 31 HC_UC and N = 23 IBS vs. N = 23 HC_IBS (with N = 2 healthy controls excluded during matching and an overlap of N = 12 healthy controls in both control groups). As intended by matching and consistent with stringent screening for abnormal BMI, no differences between the patient and control groups were evident in age or BMI (Table 1). In both patient groups, GI symptoms were expectedly significantly increased compared to healthy controls, as were reports of abdominal pain, especially in the lower abdominal region. For the UC patient group, inclusion of patients in remission or with only mild disease activity was successful, as confirmed by a symptom-based average CAI of 1.48 (SD = 1.99), and a median fecal calprotectin concentration of 41.88 μg (IQR = 83.74 μg). IBD-related medications continued as prescribed by the treating physician included aminosalicylates (N = 20), local corticosteroids (N = 2), TNF-α blocker (N = 2), and azathioprine (N = 2). Few patients reported additional extraintestinal pain symptoms (fibromyalgia, N = 2; migraine, N = 2; arthritis, N = 1). Patients with IBS reported different bowel habit disturbances, as is typical for this condition, with diarrhea-predominant (N = 9), constipation-predominant (N = 4), mixed IBS (N = 9), or unspecified (N = 1). IBS-related medications included selective serotonin reuptake inhibitors (N = 1), muscarine receptor antagonists (N = 2), and loop diuretics (N = 1). Extraintestinal pain symptoms were reported by some patients (fibromyalgia, N = 3; migraine, N = 1, arthritis, N = 2). Regarding psychological variables, significantly higher levels of psychological distress based on HADS total score were observed in both patient groups, while only patients with IBS reported significantly more chronic stress when compared to controls (Table 1).

For the comparison between patients with UC and healthy controls, the ANCOVA identified two clusters in which GMV was significantly lower in patients with UC (Table 2). These clusters comprised the left middle frontal gyrus and left anterior insula, respectively. In addition to a rendered view (Figure 1A), the two clusters are visualized on axial slices to enable a more precise localization (Supplementary Figure 1). Each cluster's extracted GMV was plotted for patients with UC and matched control groups to provide data on the single-subject level (Supplementary Figure 2). However, it should be kept in mind that these plots cannot visualize the correction for total intracranial volume, age, and sequence that was applied in the ANCOVA. In the reversed contrast, no clusters demonstrating higher GMV in patients with UC compared to healthy controls yielded significance.

For the comparison between patients with IBS and healthy controls, the ANCOVA identified seven clusters with significantly lower GMV in patients with IBS (Table 2). These clusters encompassed the left postcentral gyrus, left precuneus, right lateral orbital gyrus, right inferior temporal gyrus, right middle temporal gyrus, and right inferior occipital gyrus, respectively. Clusters are depicted in a rendered view (Figure 1B, blue color scale) as well as axial slices (Supplementary Figure 3). Again, each cluster's GMV was extracted and plotted for patients with IBS as well as matched controls (Supplementary Figure 4A). In contrast, GMV was significantly higher in patients with IBS compared to healthy controls in 12 clusters including the bilateral temporal pole, bilateral superior frontal gyrus, left middle cingulate gyrus, left middle frontal gyrus, left opercular part of the inferior frontal gyrus, left occipital pole, and right superior parietal lobule, respectively. These results are visualized using a rendered view (Figure 1B, red color scale) and axial slices (Supplementary Figure 3), and GMV of these clusters was extracted and plotted for IBS patients as well as matched controls (Supplementary Figure 4B).

Multiple regression analysis testing correlations of GMV and GI symptoms in patients with UC and HC_UC revealed a significant interaction effect between group and GI symptoms in two clusters located in the right superior frontal gyrus (Table 3). Supplemental partial correlational analyses between extracted GMV for this region and GI symptoms revealed a negative correlation in UC and a positive correlation in HC_UC, suggesting that greater GI symptoms correlated with reduced GMV in superior frontal gyrus only in patients (details and partial correlation plots in Supplementary Figure 8). Note that multiple regression analysis performed without correction for multiple comparisons revealed an interaction effect between group and GI symptoms in nine clusters (at p < 0.001), comprising additional frontal and occipital regions (Figure 2A, Table 3). These results are additionally visualized on axial slices to enable a more precise localization (Supplementary Figure 5).

In patients with IBS and controls, multiple regression analysis testing correlations between GMV and GI symptoms resulted in a significant interaction effect between group and GI symptoms in one cluster in the right occipital pole (Table 3). Supplemental partial correlational analyses between extracted GMV for this region and GI symptoms revealed a negative correlation in IBS and a positive correlation in HC_IBS, suggesting that greater GI symptoms correlated with reduced GMV in the occipital pole only in patients (details and correlation plots in Supplementary Figure 9). Note that multiple regression analysis performed without correction for multiple comparisons revealed an interaction effect between group and GI-symptoms in seven clusters (at p < 0.001, uncorrected for multiple comparisons), comprising additional frontal and temporal regions as well as the left posterior cingulate gyrus (Figure 2B, Table 3). These results are additionally visualized on axial slices (Supplementary Figure 6).

Multiple regression analysis evaluating correlations of GMV with chronic stress did not yield significant interaction effects for the analysis including patients with UC and controls (neither with FWE-correction nor with a more liberal threshold). Conversely, the same analysis in patients with IBS and controls revealed a significant interaction effect between group and chronic stress in a total of 10 clusters, encompassing the bilateral angular gyrus, bilateral temporal pole, left superior frontal gyrus (medial segment), right middle frontal gyrus, right inferior frontal gyrus (triangular part), right postcentral gyrus, right thalamus, and right inferior occipital gyrus (Table 4). Supplemental partial correlational analyses between extracted GMV for regions identified by multiple regression and chronic stress suggested that associations were consistently negative in IBS, supporting that more stress was related to lower GMV, while correlations were overall positive in healthy controls (details and correlation plots in Supplementary Figure 10). Note that multiple regression analysis performed without correction for multiple comparisons resulted in an interaction effect between group and chronic stress in 15 clusters at p < 0.001, comprising additional frontal regions, left middle cingulate gyrus, right anterior insula, left basal forebrain, and left caudate (Figure 2C, Table 4). These results are additionally visualized on axial slices (Supplementary Figure 7).

For details on methods and results, see Supplementary Material (Chapter 2). In sum, the first set of supplemental analyses in the subsample of female UC and controls revealed no significant differences in GMV between patients and controls, but largely unchanged results of multiple regression analyses. For GI symptoms, significant interaction effects were observed in comparable clusters (Supplementary Table 1). For chronic stress, no significant interaction effects were demonstrated (neither with FWE-correction nor with a more liberal threshold), in line with the original analysis. The second set of analyses on GMV plots and results of ROI-based analyses indirectly addressing the question whether the observed GMV alterations are disease-specific are presented in the (Supplementary Figures 11, 12, respectively). Results revealed disease-specific GMV changes, especially for IBS, together with shared GMV alterations for a small subset of brain regions for both disorders.