Shared and distinct spontaneous brain activity pattern in crohn’s disease and ulcerative colitis: evidence from cortical surface‑based analysis

Abstract

Background

Crohn’s disease (CD) and ulcerative colitis (UC), the two major forms of inflammatory bowel disease (IBD), are associated with emotional disturbances, but their shared and distinct neurobiological substrates remain unclear. This neuroimaging study aimed to characterize shared and distinct patterns of spontaneous neural activity in CD and UC patients using resting-state functional MRI (rs-fMRI).

Methods

Using cortical surface-based analysis of rs-fMRI data, we compared intrinsic neural activity, measured by amplitude of low-frequency fluctuations (ALFF) and regional homogeneity (ReHo), in 248 patients with IBD (180 CD, 68 UC) and 190 healthy controls (HC), controlling for gray matter volume. Demographic, clinical, and neuropsychological data were collected. Group comparisons and correlation analyses were performed.

Results

Compared to HC, both CD and UC patients exhibited reduced ALFF and ReHo in bilateral somatosensory and motor cortices. Disease-specific patterns emerged: CD showed lower ReHo in lateral temporal cortices, while UC demonstrated higher ReHo in medial temporal and superior parietal regions. Correlation analyses revealed that in CD, motor cortex activity was linked to systemic symptoms and emotional function, whereas in UC, it correlated primarily with somatization.

Conclusion

This study identifies a common neural signature of sensory-motor dysfunction in IBD, alongside subtype-specific cortical patterns. By directly comparing CD and UC with a multimodal, surface-based approach and controlling for structural differences, our findings provide novel evidence for distinct brain-gut pathophysiology, highlighting neuroimaging as a potential tool for mechanistic insight and patient stratification.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12876-026-04615-w.

Key Message

• IBD, including CD and UC, is linked to gastrointestinal symptoms, emotional disorders, and brain-gut axis interactions, but neurobiological mechanisms remain unclear.

• Both CD and UC show shared brain activity reductions in sensory and motor regions, with CD specific lateral temporal cortex and UC specific medial temporal and parietal regions differences, correlated with clinical and psychological symptoms.

• The stronger association between abnormal brain activity and quality-of-life measures in CD suggests that clinical management should pay heightened attention to the neuropsychiatric well-being and overall life quality of CD patients.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12876-026-04615-w.

Introduction

Crohn’s disease (CD) and ulcerative colitis (UC) are chronic inflammatory bowel diseases (IBD), characterized by a long clinical course and high recurrence rate [1]. The clinical symptoms of patients with IBD are variable, yet there is a certain overlap in the main clinical manifestations, such as abdominal pain, fatigue, and systemic inflammation [2]. Additionally, a series of emotional disorders, such as anxiety and depression, are observed in some IBD patients [3]. The incidence of emotional disorders in IBD patients is higher than that in the general population, and is significantly associated with clinical recurrence [4]. However, in contrast to the clinical manifestations, the emotional disorders in patients with IBD show only a modest improvement during long-term follow-up [5]. The underlying mechanisms, especially those related to neurobiology, have not been effectively and comprehensively explored.

In recent years, an increasing body of evidence has suggested the existence of common neurobiological mechanisms between CD and UC [6]. However, the exact brain functions associated with these diseases remain poorly understood. Previous research has primarily focused on the gastrointestinal aspects of IBD, such as intestinal mucosal inflammation and abnormal immune responses [7, 8]. Nevertheless, the connection between the gut and the brain, known as the gut-brain axis, has received growing attention recently. In the context of IBD, intestinal inflammation can signal the brain through multiple pathways and alter central neurotransmission and behavior [9, 10]. The brain communicates with the gut via the gut-brain axis, which encompasses the sympathetic and parasympathetic components of the autonomic nervous system and the enteric nervous system [11]. The mediates the central regulation of the autonomic nervous system. Furthermore, as a crucial component of the stress response, the central autonomic network has its sympathetic component’s hyperactivity amplifying the emotional response to stress [12]. Despite the burgeoning interest, studies comprehensively investigating the patterns of functional brain activity in CD and UC, especially those that may account for the overlapping symptoms between the two diseases, are still lacking.

Resting-state functional magnetic resonance imaging (rs-fMRI) provides a non-invasive tool to investigate brain function. Subsequent analysis of rs-fMRI data can yield quantitative metrics, such as the amplitude of low-frequency fluctuations (ALFF) [13] and regional homogeneity (Reho) [14], which allow for the assessment of the strength of spontaneous neural activity and local functional coherence, respectively. We selected ALFF and Reho as markers of intrinsic neural activity for analysis. ALFF measures the amplitude of spontaneous low-frequency (0.01–0.08 Hz) blood-oxygen-level-dependent (BOLD) signal fluctuations, reflecting the intensity of regional spontaneous neural activity. In contrast, ReHo evaluates the temporal synchronization of the BOLD signal between a given voxel and its nearest neighbors, serving as an indicator of the local functional coherence or synchronization within a specific brain region. The combined use of these two metrics provides complementary information: ALFF captures the magnitude of intrinsic activity, while ReHo reflects the local functional integration, offering a more comprehensive characterization of spontaneous brain function. Several previous studies have confirmed alterations in the cortical structure and function of patients with IBD [15–19]. However, most of these studies had small sample sizes, investigated only patients with CD or UC in isolation, and mainly focused on patients in the remission phase. As a result, the research findings varied considerably. Meanwhile, traditional three-dimensional volumetric space-based analytical methods may lead to partial volume effects, whereby a single three-dimensional brain region may simultaneously contain signals from white matter, gray matter, and even cerebrospinal fluid. Cortical surface-based brain imaging processing algorithms reconstruct the gray matter of the brain into a two-dimensional cortical surface, which better reveals the gyral and sulcal structures of the brain and facilitates structural segmentation and registration. To date, no study has systematically integrated multi-dimensional metrics, including three-dimensional volumetric space and two-dimensional cortical surface data, to elucidate the reorganization patterns of functional networks in patients with IBD.

In this context, by recruiting a larger and more comprehensive sample encompassing both CD and UC, we specifically aimed to investigate the shared and distinct patterns of spontaneous brain activity between these two IBD subtypes. Unlike prior studies that examined CD or UC in isolation, or focused solely on volume-based metrics, the present study employed a dual-metric (ALFF and ReHo), cortical surface-based analytical framework. This approach allows for a more precise spatial localization of cortical changes and facilitates a direct comparison between CD and UC, which remains underexplored. We hypothesized that this refined methodology would uncover both common neural correlates of core IBD symptomatology and subtype-specific functional alterations linked to their distinct clinical profiles. Moreover, we measured markers of intestinal permeability and inflammatory cytokines to explore the interactions within the gut-brain axis, which could further uncover the intricate relationship between the gut and the brain in CD and UC.

Materials and methods

Participants

Between January 1, 2020, and December 31, 2024, consecutive adult patients with IBD who were hospitalized at our institution were prospectively enrolled. Meanwhile, healthy control (HC) subjects were recruited via advertisements. All IBD diagnoses were established by experienced gastroenterologists through a comprehensive assessment, integrating clinical manifestations, laboratory tests, endoscopic findings, histopathological evaluations, and imaging examinations. The inclusion criteria for the patient group were as follows: (1) adult individuals newly diagnosed with IBD (CD/UC) at our hospital and not used biologicals before; (2) right-handed; (3) native proficiency in the Chinese language; (4) finished MRI examinations. All enrolled patients were newly diagnosed, with a short disease duration (median < 6 months), and were in an active disease stage at the time of scanning, as confirmed by standard clinical activity indices and elevated inflammatory markers. We also collected clinical data including disease duration and disease activity status. Disease activity for CD patients was assessed using the Crohn’s Disease Activity Index (CDAI) (CDAI; remission: CDAI < 150, active: CDAI ≥ 150), and for UC patients using the Mayo score/Partial Mayo Score (remission: Mayo score ≤ 2 with no sub-score > 1; active: Mayo score > 2).

For the HC group, participants had to meet the following criteria: (1) absence of digestive system abnormalities, including abdominal pain, diarrhea, and hematochezia; (2) right-handed; (3) native Chinese speakers; (4) finished MRI examinations. Exclusion criteria for patients: (1) poor quality of MRI images; (2) use of psychotropic drugs within the past three months; (3) history of abdominal surgery; (4) history of head trauma, brain tumor, or loss of consciousness; (5) presence of chronic conditions (e.g., hypertension, diabetes) that may affect brain structure. Exclusion criteria of HC: (1)-(5) identical to those for patients; (6) presence of any mental disorder.

A total of 17 IBD patients and 2 HC were excluded for the following reasons: 6 had poor image quality, 2 had incomplete questionnaire completion, and 9 patients and 2 HC showed head movement exceeding a mean framewise displacement (FD) of 0.5–2.0 mm in translation or 2.0° in rotation in any direction. Finally, a total of 248 patients, including 180 with CD and 68 with UC, along with 190 HC, were included in the analysis. The flow diagram of the enrolled patients and HC is shown in Fig. 1. A subset of participants in the CD group (n = 12) and HC group (n = 92) were drawn from a cohort previously reported in a study focused on disease-stage differentiation in CD using volume-based ALFF analysis by Huang et al. [13]. The present study represents a significant expansion of that work, featuring: (a) a substantially larger and independent cohort of UC patients (n = 68) not included in any prior publication from our group; (b) the inclusion of all newly recruited CD patients (bringing the total CD cohort to n = 180); and (c) a correspondingly enlarged HC group. This expanded cohort was specifically assembled to address the novel research question of shared and distinct neural signatures between CD and UC.

Fig. 1.

Flow diagram of the enrolled participants. Note: IBD, inflammatory bowel diseases; CD, Crohn’s disease; UC, ulcerative colitis; HC, healthy control

The present study represents a significant expansion of that work, featuring: (a) a substantially larger and independent cohort of UC patients not included in any prior publication from our group; (b) the inclusion of all newly recruited CD patients; and (c) a correspondingly enlarged HC group. This expanded cohort was specifically assembled to address the novel research question of shared and distinct neural signatures between CD and UC. The study was conducted in accordance with the Declaration of Helsinki, and approved by the institutional ethics committee. All participants were signed informed consent forms.

All participants completed the Chinese version of Inflammatory Bowel Disease Quality of Life Questionnaire (IBDQ) to assess quality of life, and Symptom Checklist-90 (SCL-90) for assessment of symptoms of mental disorders. The IBDQ is a validated [20], disease-specific instrument designed to assess the health-related quality of life (HRQoL) in patients with IBD. Comprising 32 items, the IBDQ comprehensively evaluates four distinct domains: intestinal symptoms (10 items), systemic symptoms (5 items), social function (5 items), and emotional function (7 items). Scores on the IBDQ are directly proportional to HRQoL, such that higher scores indicate a better quality of life among individuals with IBD. The SCL-90 is a comprehensive and widely-utilized self-assessment scale for the screening of mental disorders and psychological illnesses. It conducts screenings across ten distinct domains, including sensory perception, emotions, thinking patterns, consciousness, and interpersonal relationships. This scale demonstrates excellent discriminative ability in identifying individuals who may have psychological disorders. More importantly, its reliability and validity have withstood rigorous verification in cohorts of patients with IBD [20]. According to the Chinese norm scores, a mean score ranging from 2 to 2.9 for each item indicates mild severity, a score between 3 and 3.8 suggests moderate severity, and a score of 3.9 or higher denotes severe severity.

We also collected demographic data of the participants, including gender, age, body mass index and marital status. We further collected a comprehensive set of patients’ intestinal inflammation and nutritional parameters, including fecal calprotectin, C-reactive protein, erythrocyte sedimentation rate, hemoglobin, and albumin.

MRI data acquisition

All participants underwent MRI scans acquired with a 3.0 Tesla MRI scanner (SIEMENS SKYRA) and a 32-channel head coil. An imaging session consisted of MRI for ALFF and Reho analysis and high-spatial resolution anatomical T1-weighted imaging for anatomical registration.

The parameters for the rs-fMRI imaging were as follows: time of repetition (TR) = 2,000ms, echo time (TE) = 30 ms, flip angle (FA) = 90°, field of view (FOV) = 230 × 230 mm, number of slices = 60, voxel size = 2.4 × 2.4 × 2.4 mm, slice thickness = 2.4 mm (no gaps), time points = 240, and the total scan time was 8’13’’. The parameters for the T1-weighted imaging were as follows: TR = 5,000ms, TE = 2.98ms, FA = 9°, FOV = 256 × 240 mm, number of slices = 176, voxel size = 1 × 1 × 1 mm, slice thickness = 1.0 mm (no gaps); and the total scan time was 8’22’’. Before the scan, participants were instructed not to take caffeine containing substances. During the scan, they were asked to lie in a supine position, keep their eyes closed but remain awake. Earplugs and sponge pads were used to reduce scanning noise to protect the participants’ hearing and prevent head movement.

MRI data processing

The preprocessing and analysis pipeline consisted of three main stages: (1) structural data processing and cortical surface reconstruction, (2) functional data preprocessing and nuisance regression, and (3) calculation of cortical surface-based metrics.

Structural data processing and cortical surface reconstruction: high-resolution T1-weighted images were processed using FreeSurfer via the standard, fully automated recon-all pipeline. This process included motion correction, non-brain tissue removal, Talairach transformation, intensity normalization, subcortical structure segmentation, automatic topology correction, and gray/white matter boundary tessellation. The pipeline constructed smooth, topologically correct cortical surface models for each hemisphere. Individual surfaces were then registered to the fsaverage standard spherical template to enable group-level analysis.

Resting-state fMRI data were preprocessed using fMRIPrep 21.0.2 [21], a robust, standardized pipeline. Key steps included: removal of the first 10 time points to allow for signal stabilization, slice-timing correction, head motion realignment, and co-registration of the functional images to the individual’s T1-weighted image. The preprocessed functional data were then projected onto the individual’s native cortical surface. Nuisance covariate regression was performed on the surface-projected time series to mitigate non-neural signals. The following regressors were removed via linear regression: 24 head motion parameters (6 rigid-body motion parameters, their first-order temporal derivatives, and the squares of these 12 terms). The mean signals from white matter (WM) and cerebrospinal fluid (CSF) masks, defined from the FreeSurfer segmentation. The global signal was not regressed out, as its removal remains controversial in resting-state fMRI studies and may introduce spurious anti-correlations. Subsequently, the residual time series were band-pass filtered to focus on spontaneous low-frequency neural oscillations.

For each vertex, the filtered time series was transformed to the frequency domain using a Fast Fourier Transform. The ALFF value was defined as the average square root of the power spectrum within the 0.01–0.08 Hz range. ReHo was calculated using the Kendall’s coefficient of concordance to measure the temporal synchronization between the time series of a given vertex and those of its 26 nearest neighboring vertices on the cortical surface mesh. Critically, ReHo was computed before any spatial smoothing to preserve the intrinsic local synchronization characteristics and avoid artificial homogenization. Finally, to improve the signal-to-noise ratio and meet the Gaussian field theory assumptions for group-level inference, both the vertex-wise ALFF and ReHo maps were smoothed on the surface using a Gaussian kernel with a full-width at half-maximum (FWHM) of 6 mm. All surface-based computations and group-level statistical analyses were performed using the toolbox for Data Processing & Analysis for Brain Imaging on Surface (DPABISurf) [22] in MATLAB R2021a, which integrates fMRIPrep [21], FreeSurfer [23], and other essential neuroimaging tools [24–26].

To account for potential confounding effects due to structural differences in gray matter volume, we also performed a whole-brain voxel-based morphometry analysis comparing GM volume between the IBD (and its CD/UC subgroups) and HC groups (Supplementary Tables 1 and Supplementary Fig. 1). The detailed preprocessing steps were as follows, which were conducted using the CAT12 toolbox within SPM software: segmentation of T1-weighted images into GM, WM, and CSF; normalization to the Montreal Neurological Institute standard space with resampling to an isotropic 1.5 mm³ voxel size; modulation to preserve the total amount of GM tissue; and smoothing with an 8 mm full-width at half-maximum Gaussian kernel. All images underwent rigorous quality checks prior to analysis. This pipeline also calculated total intracranial volume for use as a nuisance covariate in the statistical models.

Statistical analysis

For continuous variables that met assumptions of normality and homogeneity of variance, a one-way ANOVA was performed across the HC, CD, and UC groups. Subsequently, a post-hoc test between the two samples was carried out. For continuous variables that violated these assumptions, the Kruskal-Wallis H test was used. For categorical variables, the Pearson’s chi-square test was applied across the three groups. Following a significant omnibus test, post-hoc pairwise comparisons were conducted with appropriate correction for multiple comparisons. The sample size for this study was determined by the availability of consecutively enrolled, treatment-naive patients meeting the inclusion criteria, resulting in a cohort that is substantial for a neuroimaging study in this clinical population. For significant group comparisons, effect sizes are reported to facilitate the interpretation of the findings. All statistical analyses were performed using R (version 4.4.3). p < 0.05 was considered statistically significant.

To identify brain regions exhibiting altered intrinsic neural activity in IBD, we compared ALFF and Reho values between HC and IBD patients, as well as between CD and UC subgroups. To control for the potential confounding effect of local gray matter volume, the preprocessed gray matter maps were incorporated as nuisance covariates in the group-level General Linear Model for all subsequent ALFF and ReHo analyses [27]. Group comparisons of ALFF and ReHo maps were conducted using two-sample t-tests in the DPABISurf statistical module, controlling for age, gender and gray matter maps as covariates. The statistical significance of the resulting statistical maps was assessed using a cluster-level family-wise error (FWE) correction approach implemented via Monte Carlo simulation. Multiple comparisons were corrected at the cluster level using Monte Carlo simulations (10,000 iterations). The cluster-forming threshold was set at a vertex-wise p < 0.001, and clusters surviving a cluster-level corrected p < 0.025 (two-tailed) were considered significant. This approach effectively controls the family-wise error rate.

Subsequently, the average signal values of brain regions with ALFF and Reho differences were extracted using DPABI software (Utilities, ROI Signal Extractor). We then calculated the Spearman’s rank correlation coefficients between these values and the clinical as well as neuropsychological parameters. Spearman’s method was chosen for its robustness to non-normal distributions and potential outliers in clinical scores. To evaluate the specificity of the observed correlations, supplementary partial correlation analyses were conducted controlling for scores on the anxiety and depression subscales of the SCL-90. Given the exploratory nature of these correlation analyses and to avoid overly conservative correction given the moderate number of pre-specified clinical variables tested against the identified regions of interest, a significance threshold of p < 0.05 (uncorrected) was applied for these specific correlations. All correlation tests were two-tailed.

Results

Analyses of demographic and neuropsychological data

Demographic and clinical characteristics of the participants are summarized in Table 1. No significant differences in age, gender, or marital status were observed between the HC and IBD groups. However, post-hoc comparisons indicated that patients with CD were significantly younger than those with UC (p < 0.05). Furthermore, both CD and UC patients had a lower BMI compared to HC subjects (both p < 0.05). The distribution of marital status also differed significantly between CD and UC patients (p < 0.05). Regarding quality of life, significant overall group differences were observed across all domains of the Inflammatory IBDQ, including bowel symptoms, systemic symptoms, emotional function, and social function. Post-hoc analyses confirmed that both CD and UC groups reported significantly more severe symptoms and poorer functioning in all these domains compared to the HC group (all p < 0.05). However, no significant differences in IBDQ scores were detected between CD and UC patients. Similarly, for neuropsychological assessment using the SCL-90, significant overall differences were found among the three groups across all symptom domains. Post-hoc tests demonstrated that both CD and UC patients had significantly elevated scores on all SCL-90 subscales compared to HCs (all p < 0.05). No significant differences in psychological symptom scores were observed between the two patient subgroups. No significant difference in disease duration was observed between CD and UC groups (p = 0.657). Clinical inflammatory and nutritional markers showed significantly higher median C-reactive protein levels in CD compared with UC (p = 0.016). In contrast, erythrocyte sedimentation rate, fecal calprotectin, hemoglobin, and albumin did not differ significantly between subtypes. The detailed results of all post-hoc pairwise comparisons are provided in Supplementary Table 2.

Table 1.

Demographics and psychological characteristics parameters

Parameters	HC	IBD	CD	UC	p value
sample size	190	248	180	68	-
age (year)	31.35 ± 12.04	32.43 ± 12.16	30.79 ± 11.25	36.78 ± 13.45#^	0.354
gender (male/female)	128/62	183/65	140/40	43/25^	0.087
body mass index	21.01 ± 2.02	20.26 ± 3.47	20.31 ± 3.51*	20.14 ± 3.40#	0.003
marital status (S/M/W)	94/95/1	121/123/4	95/81/4	26/42/0	0.223
bowel symptoms	64.87 ± 4.74	57.13 ± 8.43	57.42 ± 8.30*	56.35 ± 8.79#	< 0.001
systemic symptoms	27.55 ± 3.69	25.25 ± 5.55	25.56 ± 5.42*	24.41 ± 5.84#	< 0.001
emotional function	72.09 ± 6.37	62.80 ± 10.69	62.57 ± 10.42*	63.43 ± 11.43#	< 0.001
social function	33.66 ± 2.13	26.02 ± 6.69	26.14 ± 7.00*	25.69 ± 6.91#	< 0.001
somatization	14.24 ± 2.98	17.18 ± 4.72	17.09 ± 4.48*	17.43 ± 5.31#	< 0.001
obsessive compulsive	14.02 ± 4.07	16.16 ± 4.69	16.15 ± 4.42*	16.18 ± 5.38#	< 0.001
interpersonal sensitivity	10.81 ± 2.38	13.00 ± 4.46	13.17 ± 4.38*	12.56 ± 4.65#	< 0.001
depression	16.05 ± 4.27	20.40 ± 7.05	20.38 ± 6.80*	20.44 ± 7.74#	< 0.001
anxiety	11.90 ± 2.63	14.04 ± 4.37	14.03 ± 4.00*	14.06 ± 5.27#	< 0.001
hostility	7.07 ± 1.71	8.89 ± 3.33	8.90 ± 3.36*	8.85 ± 3.28#	< 0.001
phobic anxiety	7.68 ± 1.51	8.37 ± 2.14	8.40 ± 2.20*	8.28 ± 1.98#	< 0.001
paranoid	6.75 ± 1.47	7.84 ± 2.50	7.99 ± 2.61*	7.46 ± 2.13#	< 0.001
psychoticism	11.13 ± 2.16	13.19 ± 3.76	13.24 ± 3.64*	13.04 ± 4.08#	< 0.001
other	8.57 ± 1.93	10.78 ± 3.14	10.72 ± 2.90*	10.94 ± 3.72#	< 0.001
disease duration	-	0.13 ± 0.08	0.13 ± 0.09	0.12 ± 0.08	0.657
C-reactive protein	-	4.74 (2.98,19.50)	6.48 (2.98, 22.13)	3.31 (2.98,9.96)a	-
erythrocyte sedimentation rate	-	12.00 (5.00,27.00)	12.00 (5.00,28.00)	8.00 (4.00,23.00)	-
fecal calprotectin	-	214.31 (79.50,392.35)	249.92 (85.18,395.56)	169.19 (71.01,358.38)	-
hemoglobin	-	126.00 (114.00,140.00)	126.50 (114.00,140.00)	125.00 (114.00,141.00)	-
albumin	-	41.00 (36.90, 44.80)	40.75 (36.60, 44.37)	41.75 (38.58,45.23)	-

HC Healthy controls, IBD inflammatory bowel diseases, CD Crohn’s disease, UC ulcerative colitis, S single, M married/cohabitating, W widowed/divorced. p, p value for a T-test (HC vs. IBD)

a Significant difference of Mann-Whitney test between CD and UC

* Significant difference of post-hoc comparison between HC and CD (p < 0.05)

# Significant difference of post-hoc comparison between HC and UC (p < 0.05)

^ Significant difference of post-hoc comparison between CD and UC (p < 0.05)

*, #, ^ Denote variables with statistically significant pairwise differences as detailed in Supplementary Table 2

Intrinsic neural activities differences between IBD and HC

Whole-brain cortical surface-based analysis, revealed significant alterations in both the ALFF and ReHo in patients with IBD compared to HC. Relative to HC, the IBD group demonstrated a regionally specific pattern of altered intrinsic activity intensity (Table 2; Fig. 2, voxel p < 0.001, cluster p < 0.025, two-tailed, Monte Carlo Simulation correction). Significantly increased ALFF was observed in visual processing regions, including the bilateral primary visual cortex (Left V1: T = 4.614; Right V1: T = 4.417) and the right early visual cortex, encompassing the second (V2: T = 4.951) and third (V3: T = 4.199) visual areas. Conversely, significantly decreased ALFF was identified in the left primary motor cortex (somatosensory and motor cortex: T=-5.233) and the right anterior cingulate and medial prefrontal cortex (T=-4.837).

Table 2.

Significant differences in brain regions between IBD and healthy controls

	Region	Area Description	Cortical Division	X	Y	Z	Cluster Size	T value
ALFF LH	L_V1	Primary Visual Cortex	Primary Visual Cortex	26	-94	-15	198	4.614
	L_4	Primary Motor Cortex	Somatosensory and Motor Cortex	-36	20	14	81	-5.233
ALFF RH	R_V1	Primary Visual Cortex	Primary Visual Cortex	-21	-67	-22	34	4.417
	R_V3	Third Visual Area	Early Visual Cortex	-13	-102	-2	29	4.199
	R_V2	Second Visual Area	Early Visual Cortex	-18	-103	-4	25	4.951
	R_9m	Area 9 Middle	Anterior Cingulate and Medial Prefrontal Cortex	-24	96	13	46	-4.837
Reho LH	L_MIP	Medial IntraParietal Area	Superior Parietal Cortex	6	-66	42	56	4.484
	L_4	Primary Motor Cortex	Somatosensory and Motor Cortex	-32	18	23	81	-5.124
	L_STSdp	Area STSd posterior	Auditory Association Cortex	-40	-26	-26	35	-4.682
Reho RH	R_V1	Primary Visual Cortex	Primary Visual Cortex	-24	-74	-6	92	4.679
	R_4	PrimaryMotor Cortex	Somatosensory and Motor Cortex	36	20	14	116	-4.925
	R_SFL	SuperiorFrontal Language Area	Dorsolateral Prefrontal Cortex	-25	50	54	97	-5.772
	R_STSdp	Area STSd posterior	Auditory Association Cortex	39	-17	-31	43	-4.922

IBD inflammatory bowel diseases, HC healthy controls, L left hemisphere, R right hemisphere, MNI Montreal Neurological Institute standard coordinate space (coordinates X, Y, Z in mm), ALFF amplitude of low-frequency fluctuations, ReHo regional homogeneity. The statistical threshold was set at a vertex-wise p < 0.001 and cluster-level corrected p < 0.025 (two-tailed, Monte Carlo simulation correction)

Fig. 2.

Whole-brain differences in ALFF and ReHo between IBD patients and HC. Note: Cortical surface maps display vertices with statistically significant differences. Red colors indicate vertices where patients had greater activity/coherence compared to HC (positive T-values). Blue colors indicate vertices where patients had lower activity/coherence compared to HC (negative T-values). The color bar represents the range of T-statistic values from the two-sample t-tests. HC, healthy controls; IBD, inflammatory bowel diseases; ALFF, amplitude of low-frequency fluctuations; Reho, regional homogeneity; L, left hemisphere; R, right hemisphere. The statistical threshold was set at a vertex-wise p<0.001 and cluster-level corrected p<0.025 (two-tailed, Monte Carlo simulation with 10000 iterations)

Analysis of Reho also revealed significant between-group differences (Table 2; Fig. 2). Patients with IBD exhibited higher ReHo in the left superior parietal cortex (medial intraparietal area: T = 4.484) and the right primary visual cortex (T = 4.679). In contrast, significantly reduced ReHo was found in the bilateral primary motor cortex (Left: T=-5.124; Right: T=-4.925), the right dorsolateral prefrontal cortex (T=-5.772), and the bilateral auditory association cortex (Left: T=-4.682; Right: T=-4.922).

Disease-specific assessment of intrinsic neural activities

To delineate subtype-specific neural signatures, we conducted an omnibus ANOVA across the HC, CD, and UC groups, followed by post-hoc pairwise comparisons for ALFF and ReHo. Post-hoc comparisons revealed significant ReHo differences both between the patient groups and when each was compared to HC (Table 3; Fig. 3). Direct comparison between CD and UC patients showed that CD patients exhibited significantly lower ReHo in the bilateral lateral temporal cortex (Left: T=-4.640; Right: T=-4.833). When compared to HC, both CD and UC patients shared a significant reduction in ReHo within the bilateral primary motor cortex (somatosensory and motor cortex). Beyond this commonality, subtype-specific patterns emerged. In patients with CD, increased ReHo was observed in the bilateral primary visual cortex (Left V1: T = 4.041; Right V1: T = 4.371) and bilateral superior parietal cortex (Left MIP: T = 4.315; Right VIP: T = 4.669). Additionally, CD patients showed decreased ReHo in the right dorsolateral prefrontal cortex (T=-5.945) and the right auditory association cortex (T=-4.985) relative to HC. Conversely, in patients with UC, the most prominent finding was significantly elevated ReHo in the bilateral lateral temporal cortex compared to HC (Left TGv: T = 5.395; Right TGv: T = 5.968). They also exhibited reduced ReHo in the bilateral primary motor cortex, the right auditory association cortex (T=-4.098), and the right somatosensory cortex (Area 1: T=-4.443).

Table 3.

Significant differences in brain regions among CD, UC and HC regional homogeneity

Region	Area Description	Cortical Division	X	Y	Z	Cluster Size	T value
CD vs. UC, LH
L_TGd	Area TG dorsal	Lateral Temporal Cortex	-20	23	-70	91	-4.640
CD vs. UC, RH
R_TGv	Area TG Ventral	Lateral Temporal Cortex	7	22	-70	106	-4.833
CD vs. HC, LH
L_V1	Primary Visual Cortex	Primary Visual Cortex	24	-79	-20	141	4.041
L_MIP	Medial Intra Parietal Area	Superior Parietal Cortex	9	-68	44	37	4.315
L_4	Primary Motor Cortex	Somatosensory and Motor Cortex	-32	18	23	62	-5.094
CD vs. HC, RH
R_V1	Primary Visual Cortex	Primary Visual Cortex	-24	-74	-6	78	4.371
R_VIP	Ventral Intraparietal Complex	Superior Parietal Cortex	-11	-55	57	46	4.669
R_4	Primary Motor Cortex	Somatosensory and Motor Cortex	25	19	32	63	-4.712
R_SFL	SuperiorFrontal Language Area	Dorsolateral Prefrontal Cortex	-24	48	56	83	-5.945
R_STSdp	Area STSd posterior	Auditory Association Cortex	39	-17	-34	35	-4.985
UC vs. HC, LH
L_TGv	Area TG dorsal	Lateral temporal	-16	14	-72	172	5.395
L_4	Primary Motor Cortex	Somatosensory and Motor Cortex	-31	17	25	195	-5.058
UC vs. HC, RH
R_TGv	Area TG Ventral	Lateral Temporal Cortex	10	20	-72	187	5.968
R_4	Primary Motor Cortex	Somatosensory and Motor Cortex	29	17	28	142	-5.166
R_STSdp	Area STSd posterior	Auditory Association Cortex	39	-18	-29	43	-4.098
R_1	Area1	Somatosensory and Motor Cortex	17	-13	46	76	-4.443

CD Crohn’s disease, UC ulcerative colitis, HC Healthy controls, L left, R right, Reho regional homogeneity, MNI Montreal Neurological Institute standard coordinate space (coordinates X, Y, Z in mm); The statistical threshold was set at a vertex-wise p < 0.001 and cluster-level corrected p < 0.025 (two-tailed, Monte Carlo simulation correction)

Fig. 3.

Whole-brain differences in ALFF and ReHo among CD, UC and HC. Cortical surface maps display vertices with statistically significant differences. Red colors indicate vertices where patients had greater activity/coherence (positive T-values). Blue colors indicate vertices where patients had lower activity/coherence (negative T-values). The color bar represents the range of T-statistic values from the two-sample t-tests. CD, Crohn’s disease; UC, ulcerative colitis; HC, healthy controls; ALFF, amplitude of low-frequency fluctuations; Reho, regional homogeneity; L, left hemisphere; R, right hemisphere. The statistical threshold was set at a vertex-wise p<0.001 and cluster-level corrected p<0.025 (two-tailed, Monte Carlo simulation with 10000 iterations)

Post-hoc analysis of ALFF revealed distinct alterations in each patient group compared to HC, but no significant differences were observed in the direct comparison between CD and UC patients (Table 4). Patients with CD demonstrated a complex pattern of ALFF alterations relative to HC. Significantly increased ALFF was identified in the bilateral primary visual cortex (Left V1: T = 5.529; Right V1: T = 5.278), the right early auditory cortex (T = 4.640), and the right early visual cortex (V2: T = 4.734). In contrast, significantly reduced ALFF was found in the left primary motor cortex (T=-5.272), bilateral dorsolateral prefrontal cortex (Left 9a: T=-5.205; Right SFL: T=-4.512), and the right anterior cingulate/medial prefrontal cortex (T=-6.040). Patients with UC, compared to HC, showed a more spatially restricted ALFF alteration, characterized solely by significantly decreased activity in the left primary motor cortex (T=-4.917). No brain regions showed significantly increased ALFF in UC patients relative to HC.

Table 4.

Significant differences in brain regions among CD, UC and HC amplitude of low-frequency fluctuations

Region	Area Description	Cortical Division	X	Y	Z	Cluster Size	T value
CD vs. HC, LH
L_V1	Primary Visual Cortex	Primary Visual Cortex	26	-94	-15	169	5.529
L_9a	Area 9 anterior	Dorsolateral Prefrontal Cortex	23	103	1	122	-5.205
L_4	Primary Motor Cortex	Somatosensory and Motor Cortex	-29	18	26	63	-5.272
L_8BL	Area 8B Lateral	Dorsolateral Prefrontal Cortex	18	75	39	21	-3.815
CD vs. HC, RH
R_V1	Primary Visual Cortex	Primary Visual Cortex	-24	-74	-6	89	5.278
R_52	Area 52	Early Auditory Cortex	19	5	-11	35	4.640
R_V2	Second Visual Area	Early Visual Cortex	-19	-102	-1	35	4.734
R_9m	Area 9 middle	Anterior Cingulate and Medial Prefrontal	-23	100	6	141	-6.040
R_SFL	Superior Frontal Language Area	Dorsolateral Prefrontal Cortex	-24	52	54	28	-4.512
UC vs. HC, LH
L_4	Primary Motor Cortex	Somatosensory and Motor Cortex	-28	15	29	53	-4.917

CD Crohn’s disease, UC ulcerative colitis, HC Healthy controls, L left, R right, ALFF amplitude of low-frequency fluctuations. MNI Montreal Neurological Institute standard coordinate space (coordinates X, Y, Z in mm), The statistical threshold was set at a vertex-wise p < 0.001 and cluster-level corrected p < 0.025 (two-tailed, Monte Carlo simulation correction)

Correlations among shared atypical intrinsic neural activities, neuropsychological and intestinal inflammation markers

To investigate the relationship between shared atypical intrinsic neural activities, neuropsychological symptoms, and intestinal inflammatory markers, we conducted correlation analyses based on regions of interest (ROI) derived from cortical regions commonly altered in both CD and UC patients compared to HC (Fig. 4A). These analyses were performed for the entire IBD cohort as well as separately for the CD and UC subgroups.

Fig. 4.

Correlation among intrinsic neural activities, neuropsychological and intestinal inflammation markers. Note: A) The amplitude of low-frequency fluctuations (ALFF) and regional homogeneity (Reho) values of common cortical differential regions observed in both Crohn’s disease (CD) and ulcerative colitis (UC) relative to and healthy controls (HC). B) Correlation among intrinsic neural activities and neuropsychological markers in patients with inflammatory bowel disease. C) Correlation among intrinsic neural activities and neuropsychological markers in CD. D) Correlation among intrinsic neural activities, neuropsychological and intestinal inflammation markers in UC

In the combined IBD group, ReHo in the left primary motor cortex was positively correlated with emotional function (r = 0.136, p = 0.033) and somatization (r = 0.128, p = 0.044). ALFF in the left primary motor cortex was also positively associated with systemic symptoms (r = 0.130, p = 0.041). In the right primary motor cortex, ReHo showed a significant positive correlation with somatization (r = 0.157, p = 0.013). Conversely, ReHo in the right superior temporal sulcus dorsal posterior region was negatively correlated with obsessive-compulsive scores (r=-0.139, p = 0.029), anxiety (r=-0.144, p = 0.024), and paranoid (r=-0.128, p = 0.045). Anxiety were negatively correlated with fecal calprotectin (r=-0.129, p = 0.043) and C-reactive protein (r=-0.167, p = 0.009). Erythrocyte sedimentation rate was negatively correlated with emotional function (r=-0.218, p = 0.001) and social function (r=-0.256, p = 0.001). Albumin levels were positively correlated with social function (r = 0.149, p = 0.019) (Fig. 4B).

Subgroup analyses revealed distinct correlation patterns between CD and UC. In CD patients, ReHo in the right primary motor cortex was positively correlated with phobic anxiety (r = 0.164, p = 0.027), while ReHo in the right superior temporal sulcus dorsal posterior region was negatively correlated with obsessive compulsive symptoms (r=-0.178, p = 0.017) and paranoid (r=-0.160, p = 0.032). ALFF in the left primary motor cortex was positively correlated with systemic symptoms (r = 0.180, p = 0.015). Fecal calprotectin was negatively correlated with somatization (r=-0.190, p = 0.010), and C-reactive protein was negatively associated with social function (r=-0.192, p = 0.010) and anxiety (r=-0.148, p = 0.047). Erythrocyte sedimentation rate was negatively correlated with emotional function (r=-0.273, p = 0.001), while albumin was positively correlated with bowel symptoms (r = 0.163, p = 0.029) (Fig. 4C).

In UC patients, somatization scores were positively correlated with ReHo in the primary motor cortex bilaterally (left: r = 0.257, p = 0.034; right: r = 0.298, p = 0.014). ALFF in the left primary motor cortex was positively correlated with somatization (r = 0.242, p = 0.047) but negatively correlated with fecal calprotectin (r=-0.262, p = 0.031) (Fig. 4D).

To assess the potential confounding influence of emotional symptoms, partial correlation analyses controlling for anxiety and depression scores were performed (Supplementary Table 3). The magnitudes and significance levels of the reported correlations between neural activity and primary clinical measures remained largely unchanged, supporting the robustness of these associations independent of comorbid anxiety and depression.

Discussion

The present study extends the current neuroimaging literature on IBD in several key aspects. First, it provides the first direct, concurrent comparison of intrinsic brain activity between CD and UC using a cortical surface‑based approach, identifying both convergent and divergent neural patterns. Second, by employing complementary metrics of ALFF and ReHo, it offers a more nuanced characterization of functional alterations than studies relying on a single metric. Third, the analysis of correlations with a broad range of clinical and psychological variables within each subtype provides new insights into the differential brain‑gut‑behavior relationships in CD versus UC. Critically, by incorporating voxel‑wise GM volume as an anatomical covariate in our models, we have strengthened the inference that the observed functional changes are not merely secondary to structural differences, but likely reflect intrinsic neurophysiological disturbances. Two primary findings were identified. First, compared with HC, both CD and UC groups exhibited significantly decreased ReHo values in the bilateral somatosensory and motor cortices and the right auditory association cortex. Additionally, significantly reduced ALFF values were observed in the left somatosensory and motor cortex of both patient groups relative to HC. Importantly, these core alterations remained statistically significant after controlling for local GM volume, underscoring their robustness as functional signatures of IBD. Second, in the overall IBD cohort, shared atypical intrinsic neural activity clusters correlated with HRQoL, emotional symptoms, and markers of intestinal inflammation. Subgroup analyses further revealed that CD showed stronger and more extensive associations between these neural alterations and HRQoL domains compared to UC.

The identification of bilateral somatosensory and motor cortex hypoactivity in both CD and UC relative to HC underscores a common neural signature in IBD. These regions are critical for processing visceral pain and motor control. Their hypoactivity may reflect maladaptive neural plasticity in response to persistent gut inflammation, potentially contributing to central sensitization and altered pain perception, a hallmark of IBD, related abdominal pain [18, 28, 29]. Notably, the reduced activity in the right auditory association cortex further suggests disruptions in multisensory integration, which could exacerbate emotional distress by impairing the brain’s ability to modulate stress responses via the central autonomic network [30, 31]. The observed hypoactivity in the dorsolateral prefrontal cortex and anterior cingulate cortex aligns with prior studies linking IBD to deficits in emotion regulation [32, 33]. These regions are pivotal for top‑down control of the autonomic nervous system, and their dysfunction may explain the heightened anxiety and depression in IBD patients by compromising stress resilience [34, 35]. The shared pattern across CD and UC supports the hypothesis of a unified neurobiological pathway for emotional and somatic symptoms in IBD. Deep brain stimulation for pain relief is under investigation, and the dorsal anterior cingulate cortex is considered another potential target for neural modulation [36]. Although it has not been validated as a specific therapeutic target for CD, this may represent a promising new direction for treatment.

The correlations between somatosensory and motor cortex activity and intestinal inflammation markers highlight gut-brain interactions. Elevated fecal calprotectin was negatively correlated with motor cortex activity, suggesting that systemic inflammation may drive neuroinflammation via cytokines crossing the blood‑brain barrier, leading to synaptic pruning or glial activation in motor‑related regions regions [37, 38]. It is noteworthy that we did not find significant correlations between albumin levels, C‑reactive protein and intrinsic neural activity. This may reflect that these markers capture immediate or recent intestinal inflammatory activity, whereas the neural alterations we observed might be more closely linked to the cumulative effects of chronic inflammation or to neural remodeling processes that persist even during inflammatory remission.

Our study reinforces previous reports indicating a specific association between CD and HRQoL metrics. This may reflect that CD patients are more prone to systemic symptoms such as fatigue, affective dysfunctions like anxiety and depression, and social stigma due to intestinal lesions or perianal diseases [39, 40]. Conversely, albumin levels positively correlated with bowel symptoms and social function, aligning with prior studies [13, 41] and suggesting that malnutrition in IBD could exacerbate neural dysfunction, possibly through energy deficit or oxidative stress. The stronger brain‑behavior correlations observed in CD suggest that gastroenterologists should pay heightened attention to the neuropsychiatric well‑being and overall quality of life when managing CD, potentially integrating psychological assessments and interventions more proactively into care plans.

From a clinical perspective, the identification of shared and distinct neural signatures in IBD may inform future diagnostic and therapeutic strategies. For instance, monitoring spontaneous neural activity via rs-fMRI could serve as an objective biomarker for predicting treatment response or disease relapse, complementing traditional inflammatory markers. Furthermore, the shared motor, somatosensory abnormalities suggest that non‑invasive neuromodulation techniques, such as transcranial magnetic stimulation (TMS) targeting these regions, could be explored as potential adjunctive therapies for managing chronic pain and emotional comorbidities in IBD patients.

While our analysis identified robust alterations in somatosensory, motor, and prefrontal regions, it is noteworthy that we did not detect significant group differences in several other networks frequently implicated in chronic visceral conditions and emotional processing. Most notably, core nodes of the default mode network (DMN), such as the posterior cingulate cortex/precuneus and medial prefrontal cortex, did not show altered ALFF or ReHo.

This contrasts with some previous studies in chronic pain syndromes [42] and suggests that the intrinsic neural dysfunction in newly diagnosed, active IBD may be more specific to circuits governing sensory integration, motor planning, and cognitive‑emotional regulation rather than generalized DMN dysregulation. Furthermore, key limbic structures directly involved in fear and anxiety processing, such as the amygdala, did not emerge as significantly different in our whole-brain analysis. The absence of significant findings in these areas could reflect the specific methodology, the clinical stage of our cohort, or indicate that alterations in these regions may be more subtle or relate more specifically to sustained disease burden or particular symptom clusters.

The study’s large sample size and cortical surface-based analysis, enhancing spatial precision, mitigate limitations of prior small-scale IBD neuroimaging research. However, this study has several limitations. First, its cross‑sectional design precludes causal inference between intrinsic neural activity, inflammation, and mood disorders. Longitudinal studies are needed to track neural activity changes across active and remission phases and to validate temporal associations. Second, while excluding patients using biologic agents or psychotropic medications ensured the purity of neural signals, it limits the generalizability of our findings to treated populations. Third, the neuropsychological assessment focused on emotional and quality‑of‑life dimensions; we did not include specific cognitive tests, which limits our understanding of potential cognitive alterations in IBD.

Furthermore, we did not collect data on educational level, a factor that could influence neuropsychological performance. Fourth, although covarying for GM volume significantly strengthens the inference regarding functional specificity, the cross‑sectional nature of our study precludes definitive conclusions about structure‑function dynamics over time. Fifth, our cohort primarily comprised patients in the active disease stage. While this reduces heterogeneity, it precludes analysis of state‑dependent neural changes across remission phases. Future longitudinal studies tracking patients from active to remission are needed to disentangle state from trait neural markers. Finally, while we employed a standard cluster‑based correction method to balance Type I and Type II error rates, the multiple group comparisons performed increase the likelihood of findings. The results should therefore be interpreted in this context and await replication in independent cohorts.

Conclusions

In conclusion, this study establishes a common atypical brain activity pattern in CD and UC, centered on motor, somatosensory, and prefrontal regions, which correlates with inflammation, nutritional status, and psychological symptoms. The use of a cortical surface‑based approach, combined with stringent control for anatomical differences, provides robust evidence for distinct functional neurophysiology in IBD. These findings advance our understanding of the gut‑brain axis in IBD pathogenesis and highlight the potential of neuroimaging as a tool for mechanistic insight and therapeutic monitoring.

Abbreviations

CD

Crohn’s disease

IBD

Inflammatory bowel diseases

UC

ulcerative colitis

HC

healthy controls

ALFF

amplitude of low-frequency fluctuations

Reho

regional homogeneity

IBDQ

Inflammatory Bowel Disease Quality of Life Questionnaire

Authors’ contributions

M.H.: Conceptualization, Methodology, Software, Validation, Formal analysis, Data curation, Funding, Writing-original draft, Writing-review & editing, Visualization; Q.L.: Methodology, Software, Validation, Formal analysis, Writing-review & editing, Visualization; S.H.: Software, Validation, Writing-review & editing; Y.Z.: Methodology, Validation, Formal analysis, Resources, Writing-review & editing; Y.H.: Formal analysis, Visualization, Writing-review & editing; L.Z.: Methodology, Resources, Data curation, Writing-review & editing; P.H.: Conceptualization, Methodology, Validation, Resources, Writing-review & editing, Supervision; J.W.: Conceptualization, Methodology, Validation, Resources, Writing-review & editing, Supervision; H.S.: Conceptualization, Methodology, Funding, Validation, Formal analysis, Resources, Data curation, Writing-review & editing, Supervision, Research administration.

Funding

This study was supported by National Natural Science Foundation of China (grant no. 82071921) and Fundamental Research Funds for the Central Universities (YCJJ20252426).

Data availability

The materials and codes used and/or analyzed during the current study are available from the corresponding author upon reasonable request after the entire data collection procedure and project are completed.

Declarations

Ethics approval and consent to participate

This study was approved by the institutional ethics committee of Union Hospital of Tongji Medical College of Huazhong University of Science and Technology (0940-01). All participants were signed informed consent forms.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.