From localized to long-range: functional connectivity alterations in mild to moderate chronic low back pain patients Authors: Han-Jun Hu, Luo-Yu Wang, Jia-Yi Deng, Yi Lin, Zhong-Xiang Ding, Xue Tang, Yi-Di Zhu, Xiu-Hong Ge, Han Zhang, Liang Yu, Li-Qing Zhang Published: 2026 ABSTRACT BackgroundMild pain can influence brain function, yet its central neurobiological mechanisms in chronic low back pain (CLBP) remain insufficiently defined. This study aimed to quantitatively evaluate localized and long-range functional connectivity (FC) alterations in patients with mild to moderate CLBP using static/dynamic regional homogeneity (sReHo/dReHo) and seed-based FC, and to assess their statistical relationships with clinical pain and psychological metrics.MethodsA total of 55 mild to moderate CLBP patients and 39 healthy controls (HCs) underwent 3.0 T resting-state functional magnetic resonance imaging (rs-fMRI) and clinical assessments, including Visual Analogue Scale (VAS), Short-Form of the McGill Pain Questionnaire (SF-MPQ), Self-Rating Anxiety Scale (SAS), Self-Rating Depression Scale (SDS), and Pittsburgh Sleep Quality Index (PSQI). sReHo and dReHo (sliding window =50 TRs, step size =5 TRs) maps were computed to identify abnormal localized brain regions. Brain regions showing significant group differences were selected as seeds (6-mm spherical regions of interest, ROIs) for whole-brain FC analysis. Statistical comparisons were corrected using Gaussian random field (GRF) theory (voxel-level P<0.005, cluster-level P<0.05). Partial correlation analyses were used to control for age, sex, education, and head motion (Bonferroni-corrected threshold P<0.007).ResultsCompared with HCs, patients exhibited significantly reduced sReHo in the third cerebellar vermis (peak t=−4.253, GRF P<0.05), alongside reduced dReHo in the right parahippocampal gyrus (t=−4.796, GRF P<0.05) and left inferior temporal gyrus (t=−3.935, GRF P<0.05). Seed-based FC analysis revealed decreased connectivity between the third cerebellar vermis and right hippocampus (t=−4.678, GRF P<0.05), and increased connectivity between the left inferior temporal gyrus and left pallidus (t=4.441, GRF P<0.05). Patients showed significantly higher SAS anxiety scores (25.15±5.01 vs. 23.08±2.18, P=0.008). Correlation analyses demonstrated that sReHo in the third cerebellar vermis was negatively correlated with SAS scores (r=−0.384, P=0.005), whereas FC between the left inferior temporal gyrus and left pallidus was positively correlated with SF-MPQ sensory (r=0.381, P=0.006), affective (r=0.386, P=0.005), and total scores (r=0.445, P=0.001).ConclusionsPatients with mild to moderate CLBP exhibit distinct localized and long-range FC alterations within cerebello-hippocampal and cortico-basal ganglia pathways. These neurofunctional changes may contribute to the integration of subjective pain dimensions and comorbid anxiety. Combining static/dynamic ReHo with seed-based FC provides complementary neurobiological insight into how low-intensity chronic pain may affect the central nervous system and highlights the importance of early clinical recognition. Introduction While mild pain has been shown to influence brain activity (1), its impact has often been neglected in prior studies. Chronic low back pain (CLBP) is defined as pain occurring between the lower ribs and the sacrum, with a duration exceeding 3 months (2). It is now recognized as a multifactorial chronic pain condition rather than merely a symptom (3). CLBP is one of the most prevalent health issues globally. In 2020, it affected 619 million people worldwide, and this number is projected to rise to 843 million by 2050, imposing a significant burden on both individuals and society (4). Epidemiological studies indicate that most CLBP patients experience lower pain intensity (5), which can be classified using a standard 0–10-point linear Visual Analogue Scale (VAS), where scores of 1–3 indicate mild pain, scores of 4–6 indicate moderate pain, and scores of 7 or greater indicate severe, high-intensity pain (6). Although mild to moderate CLBP involves less severe pain than high-intensity conditions, it exhibits a high clinical incidence, a strong tendency for recurrence, prolonged symptom duration, and less-than-ideal treatment outcomes (7). Importantly, these persistent clinical characteristics may negatively affect mental health and increase the risk of depressive symptoms (8). These features, including prolonged symptom duration, recurrence tendency, and emotional vulnerability, may be linked to alterations in the underlying neural substrate. Given the adaptability of the human central nervous system, chronic pain can induce brain reorganization across multiple functional networks (9). Persistent pain can trigger central sensitization, and long-term pain-related emotional distress can disrupt relevant brain networks (10-12). Despite their lower pain intensity, patients with mild to moderate CLBP may also experience brain network alterations similar to those documented in more severe chronic pain states. Therefore, there is an urgent need to gain a deeper understanding of the underlying functional brain alterations in this population. Such knowledge may improve clinical awareness of the central mechanisms underlying low-intensity chronic pain, facilitate early intervention to prevent potential chronification, and provide neurobiological insights for optimizing individualized treatment strategies. Resting-state functional magnetic resonance imaging (rs-fMRI) is suitable for observing brain function changes in patients with mild to moderate CLBP. This advanced non-invasive imaging technique indirectly reflects neuronal activity by assessing spontaneous fluctuations of blood oxygen level-dependent (BOLD) signals and has been widely used in diverse diseases (13,14). Regional homogeneity (ReHo) is a widely used method for analyzing rs-fMRI data. It uses Kendall’s consistency coefficient (KCC) to measure the temporal synchronization of BOLD signals between a voxel and its neighbors, characterizing local neuronal activity at rest (15). Compared to other common indicators such as amplitude of low-frequency fluctuations (ALFF) and independent component analysis (ICA), ReHo has higher sensitivity in detecting local abnormalities and good test-retest reliability (16). Several studies have observed abnormal ReHo changes in brain regions of CLBP patients and their normalization after pain relief (17-19). Based on these studies, ReHo can effectively assess the local synchronization of spontaneous BOLD signals in CLBP patients and can further be applied to explore the central nervous system mechanisms of mild to moderate CLBP. Although participants are in a resting state during scanning, their brain activity still dynamically fluctuates over time (20), and conventional static metrics, such as static regional homogeneity (sReHo) and static amplitude of low-frequency fluctuations (sALFF), are prone to overlook these temporal dynamics (21). Dynamic regional homogeneity (dReHo) captures temporal neural-activity synchronization within shorter time windows than sReHo, making it more sensitive to inter-group differences in regional neural activity and better at detecting subtle brain-function changes in mild to moderate CLBP patients. dReHo has been used in various pain studies such as trigeminal neuralgia (22,23) and toothache (24), showing complementary results to static metrics. Unfortunately, it has not been applied to CLBP research yet. While sReHo and dReHo can only assess local adjacent brain-activity characteristics, they cannot detect functional connections between spatially separate brain regions. Seed-based functional connectivity (FC), another widely used rs-fMRI method, can address this. It correlates the resting-state time series of a selected region of interest (ROI, i.e., the seed) with those of the rest of the brain to explore the synchronous functional-activity relationships of remote neurons (25). Yu et al. found that FC of midbrain-cortical and midbrain-limbic pathways is reduced in CLBP patients (26). Other studies using seed-based FC also found long-range FC alterations in CLBP patients (27,28). For seed-based analysis, selecting a proper ROI is crucial. ReHo, a reliable data-driven method, can inform this choice (29). Therefore, we extracted seed points from brain regions with significant sReHo and dReHo changes for seed-based FC analysis. This framework combines local short-range synchronization and whole-brain long-range FC to better detect regional and global brain function alterations. This study includes two parts. First, we measured sReHo and dReHo to identify localized brain function differences between mild to moderate CLBP and healthy control (HC) groups, explored temporal variability of local brain activity, and studied the link between sReHo, dReHo changes and patients’ pain intensity and psychological assessments. Then, we investigated the role of sReHo and dReHo-different brain regions in whole-brain functional connections. We extracted seeds from brain regions with statistically significant sReHo and dReHo differences between the two groups, performed seed-based FC analysis, compared FC changes, and evaluated the relationship between patients’ FC, pain intensity, and psychological status. To our knowledge, this is the first study to jointly assess localized and long-range FC alterations in mild to moderate CLBP patients using static and dynamic ReHo and seed-based FC analysis. We hypothesized that: (I) sReHo and dReHo values in patients with mild to moderate CLBP would differ from those in HCs, with dynamic indicators providing complementary information to static ones; (II) whole-brain FC would be altered in patients with mild to moderate CLBP; and (III) localized brain activity and long-range FC alterations would be related to patients’ clinical symptoms and mental-psychological changes. We present this article in accordance with the STROBE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2026-0906/rc). Methods Participants Between April 2024 and January 2025, 61 mild to moderate CLBP patients with lumbar disc herniation and 40 HCs were recruited from The Affiliated Hangzhou First People’s Hospital, Westlake University School of Medicine. HCs were not recruited using a strict 1:1 matching protocol; however, recruitment targeted comparable distributions of age, sex, and education, and group balance was statistically evaluated. This prospective study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of The Affiliated Hangzhou First People’s Hospital, Westlake University School of Medicine (No. KY-20240321-0088-02). All participants were right-handed and provided written informed consent. All subjects underwent comprehensive assessments, including medical history collection, lab tests, physical exams, routine brain magnetic resonance imaging (MRI) scans, and scale evaluations. Inclusion criteria for patients with mild to moderate CLBP were as follows: (I) diagnosis of lumbar disc herniation supported by consistent clinical signs and MRI findings, characterized by predominant CLBP; (II) pain lasting or recurring for ≥3 months; (III) VAS score between 1 and 6, representing mild to moderate pain intensity based on a standard 0–10 point linear VAS and excluding severe pain scores of 7 or greater; (IV) no other chronic pain conditions; (V) aged 20–75 years; (VI) right-handed; (VII) no motor, auditory, or visual system diseases, language barriers, or vision problems (corrected vision was normal); (VIII) willingness to participate. Inclusion criteria for HCs: (I) no history of chronic pain, including minor pain; (II) aged 20–75 years; (III) right-handed; (IV) normal routine cranial MRI; (V) good health, no cancer or psychiatric history; (VI) no motor, auditory, or visual system diseases, language barriers, or vision problems (corrected vision was normal); (VII) willingness to participate. Exclusion criteria for all participants: (I) severe hypertension, diabetes, or other brain-function-affecting diseases; (II) history of organic brain lesions, head trauma, or invasive brain surgery; (III) contraindications to MRI scans; (IV) history of alcohol dependence or drug abuse. After image quality checks, invalid data were excluded (Figure 1). Finally, 55 mild to moderate CLBP patients and 39 HCs were included. Flowchart of participant inclusion and exclusion. CLBP, chronic low back pain; FD, framewise displacement; HC, healthy control. Pain and psychological assessment Within 2 hours after the MRI scan, all participants completed a series of mental health questionnaires. Given that all participants were native Chinese speakers, officially validated Chinese versions of these scales were administered to ensure optimal comprehension and psychometric reliability. The study used the validated Chinese versions of the Self-Rating Anxiety Scale (SAS) and Self-Rating Depression Scale (SDS) to assess anxiety and depression (30), and the Pittsburgh Sleep Quality Index (PSQI) to evaluate sleep quality (31). Patients with mild to moderate CLBP also completed a standard 0–10-point linear VAS and the validated Chinese version of the Short-Form of the McGill Pain Questionnaire (SF-MPQ) (6). The VAS measured current pain intensity, with 0 representing no pain and 10 representing the worst imaginable pain. Patients were instructed to select the integer that best reflected their average pain intensity over the past month. The SF-MPQ sensory items assess pain’s physical nature, and the affective items evaluate its emotional impact. The same physician administered all scales. MRI data acquisition All participants underwent MRI data acquisition using a 3.0 T MRI scanner (SIGNA Premier, GE Healthcare, Waukesha, WI, USA) with a 48-channel phased-array head coil. Routine brain MRI scans were performed to exclude relevant intracranial diseases, including T1-weighted imaging (T1WI), T2-weighted imaging (T2WI), T2 fluid-attenuated inversion recovery (FLAIR), and diffusion-weighted imaging (DWI) sequences. Anatomical images were acquired using 3D T1-weighted imaging (3D-T1WI) with these parameters: repetition time (TR) =2,617 ms, echo time (TE) =2.95 ms, inversion time =1,000 ms, flip angle =8°, field of view (FOV) =256 mm × 256 mm, slice thickness/gap=1/1 mm, matrix =256×256, 176 sagittal slices. The fMRI data were acquired using the following scanning parameters: TR =800 ms, TE =30 ms, slice thickness/gap =3/3 mm, FOV =220 mm × 220 mm, flip angle =56°, with Hyperband acceleration (factor of 4), and 600 time points, in the phase encoding direction. Additionally, fMRI sequences with identical parameters were acquired in the anterior-posterior direction for spatial distortion correction. Image preprocessing We checked the quality of all 3D-T1WI and rs-fMRI images, discarding any incomplete or artifact-contaminated ones. We used the FMRIB Software Library (FSL) and Advanced Normalization Tools (ANTs) to preprocess rs-fMRI data. The steps were as follows: (I) discarding the first 12 volumes (~10 s) for subject acclimatization and signal stabilization; (II) correcting head motion, excluding subjects with displacement >3 mm, rotation >3°, or average frame displacement >0.5 mm; (III) using opposite-phase encoded EPI data for spatial distortion correction to fix susceptibility artifacts; (IV) aligning functional MRI with structural MRI for precise localization via tissue boundary maps; (V) resampling in each subject’s space to merge conversions, avoiding repeated interpolation and smoothing; (VI) using ICA-AROMA to identify and detect motion-related noise; (VII) regressing out nuisance covariates in a single statistical step, including the 24 head motion parameters derived from the Friston 24-parameter model, linear trends, and average time series extracted from white matter and cerebrospinal fluid; (VIII) normalizing individual brain images to the Montreal Neurological Institute (MNI) standard space, based on tissue segmentation by FSL and resampling them to 3×3×3 mm3 voxels to ensure anatomical consistency across subjects; (IX) applying a bandpass filter (0.01–0.08 Hz). sReHo calculation ReHo values were obtained by calculating the KCC between a given voxel and its 26 nearest neighbors’ time series. For standardization, the ReHo value of each voxel was divided by the global mean ReHo. Spatial smoothing [full width at half maximum (FWHM) =6 mm] was then applied to reduce registration imperfections and boost the image signal-to-noise ratio. All these procedures were performed using the Data Processing & Analysis for (Resting-State) Brain Imaging (DPABI, version 8.1) software (32). The anatomical identification and labeling of localized brain regions were determined based on the Automated Anatomical Labeling (AAL) atlas. dReHo calculation In this study, we used a sliding window approach to calculate dReHo and detect temporal brain-activity fluctuations. The analysis was performed using the Dynamic Analysis Toolbox in DPABI 8.1. Window length is crucial in dynamic metrics; too short or too long a window may fail to capture dynamic activity or robustly assess changes (33). Based on prior research (33), we used a sliding window of 50 TRs (40 s) with a step size of 5 TRs (4 s) to calculate ReHo in each window. To assess ReHo variability, we generated a coefficient of variation (CV) map for each participant by calculating the CV as where σ represents the standard deviation and μ represents the mean of ReHo values across time windows, producing individual dReHo maps. Spatial smoothing with a Gaussian kernel (FWHM =6 mm) was then applied. To guarantee that varying sliding-window lengths wouldn’t impact results, we validated the dReHo results using a sliding window of 125 TRs (100 s) and a step size of 5 TRs (4 s) (Figure S1, Table S1). FC calculation based on ROI Seed-based FC was estimated using DPABI 8.1 software. Regions with significant group differences in both static and dynamic ReHo analysis between the mild to moderate CLBP and HC groups were used as seeds for further resting-state FC analysis. Specifically, the center voxels of these seed regions were defined by the peak MNI coordinates of statistically significant clusters derived from inter-group comparisons, as detailed in the Results section. Seed regions were spheres with a radius of 6 mm around the center voxels, and the reference time series for seeds were obtained by averaging the time series of all voxels within the seed region. Correlation analysis was then performed between the seeds and the remaining voxels. Finally, the correlation coefficients were converted into Fisher z values to obtain a z-FC map for further statistical analysis. Statistical analysis We used SPSS 25.0 (IBM Corp., Armonk, NY, USA) to analyze clinical data, pain, and mental health scores for mild to moderate CLBP and HC groups. Continuous variables (age, education, head motion, SAS, SDS, PSQI, VAS, SF-MPQ scores) were expressed as means ± standard deviations (SDs) and compared using independent-samples t-tests. Sex, a categorical variable, was compared using the chi-square test. Statistical significance was set at P<0.05. In DPABI 8.1, independent two-sample t-tests compared sReHo, dReHo, and FC values in gray matter between mild to moderate CLBP patients and HCs to detect group differences. We regressed out four covariates: age, sex, education, and head motion. Using Gaussian random field (GRF) theory, we corrected for multiple comparisons in the resulting t maps, setting the voxel-level threshold at P<0.005 and the cluster-level threshold at P<0.05. To explore the relationships between sReHo, dReHo, and Seed-based FC metrics with pain and mental health scores in mild to moderate CLBP patients, we used the ROI signal extractor command in DPABI 8.1. We extracted sReHo and dReHo values from seed regions and FC values from brain areas with significant group differences after Seed-based FC analysis. Then, using SPSS 25.0, we performed partial correlation analyses between these extracted values and the patients’ pain and mental health scores, controlling for sex, age, education, and head motion. Finally, we applied a Bonferroni-corrected multiple testing threshold of P<0.007 (0.05/7) to identify statistically significant correlations. Results Demographic information and clinical characteristics A total of 55 patients with mild to moderate CLBP and 39 HCs were included in the data analysis. Notably, no significant differences were found between the patient and control groups in sex, age, education, head motion, SDS, or PSQI scores (all P>0.05). However, the patient group had a higher SAS score than the control group (P=0.008, Table 1). Table 1Demographics and clinical characteristics of the participantsCharacteristicsMild to moderate CLBP (N=55)HCs (N=39)PvalueSex (male/female)27/2814/250.204†Age (years)52.29±12.9348.49±14.920.191‡Education (years)11.16±4.5912.78±5.910.157‡Head motion (average FD)0.07±0.040.06±0.030.448‡SAS (scores)25.15±5.0123.08±2.180.008‡SDS (scores)23.16±4.1023.79±2.850.381‡PSQI (scores)6.38±3.065.72±3.800.369‡VAS (scores)3.60±1.10––SF-MPQ sensory (scores)3.80±2.53––SF-MPQ affective (scores)1.55±1.93––SF-MPQtotal (scores)5.35±3.83––†, chi-square test; ‡, independent sample t-test. CLBP, chronic low back pain; FD, framewise displacement; HC, healthy control; PSQI, Pittsburgh Sleep Quality Index; SAS, Self-Rating Anxiety Scale; SDS, Self-Rating Depression Scale; SF-MPQ, Short-Form of the McGill Pain Questionnaire; VAS, Visual Analogue Scale. Demographics and clinical characteristics of the participants †, chi-square test; ‡, independent sample t-test. CLBP, chronic low back pain; FD, framewise displacement; HC, healthy control; PSQI, Pittsburgh Sleep Quality Index; SAS, Self-Rating Anxiety Scale; SDS, Self-Rating Depression Scale; SF-MPQ, Short-Form of the McGill Pain Questionnaire; VAS, Visual Analogue Scale. Group differences in sReHo There were significant group differences in the sReHo values of the third cerebellar vermis (Vermis_3) in patients with mild to moderate CLBP. Compared with the control group, the sReHo values of the patient group were lower (voxel-level P<0.005, cluster-level P<0.05, corrected with GRF; Table 2, Figure 2). Table 2Brain areas changed by sReHo across the two groupsAnatomical locationPeak MNI coordinatesCluster size (voxels)Peak intensityXYZVermis_36−39−9255−4.253MNI, Montreal Neurological Institute; sReHo, static regional homogeneity; Vermis_3, third cerebellar vermis. Brain areas changed by sReHo across the two groups MNI, Montreal Neurological Institute; sReHo, static regional homogeneity; Vermis_3, third cerebellar vermis. Brain regions with significant sReHo differences between mild to moderate CLBP and HC. Statistical parametric maps demonstrate the brain region with significantly lower sReHo in the mild to moderate CLBP group compared to the HC group. The cold color indicates the spatial extent of the decrease, localized in the Vermis_3. The violin plot compares the mean sReHo values extracted from the Vermis_3 between the two groups. **, P<0.005. CLBP, chronic low back pain; HC, healthy control; sReHo, static regional homogeneity; Vermis_3, third cerebellar vermis. Group differences in dReHo In the right parahippocampal gyrus (ParaHippocampal_R) and the left inferior temporal gyrus (Temporal_Inf_L), significant group differences were found. In these regions, the dReHo values of the patient group were lower than those of the control group (GRF correction, voxel-level P<0.005, cluster-level P<0.05; Figure 3, Table 3). Brain regions with significant dReHo differences between mild to moderate CLBP and HC. Statistical parametric maps demonstrate brain regions with significantly lower dReHo in the mild to moderate CLBP group compared to the HC group. The cold color indicates the spatial extent of the decrease, primarily localized in the (A) ParaHippocampal_R and (B) Temporal_Inf_L. Violin plots compare the mean dReHo values extracted from these two regions between the CLBP and HC groups. **, P<0.005. CLBP, chronic low back pain; dReHo, dynamic regional homogeneity; HC, healthy control; ParaHippocampal_R, right parahippocampal gyrus; Temporal_Inf_L, left inferior temporal gyrus. Table 3Brain areas changed by dReHo across the two groupsAnatomical locationPeak MNI coordinatesCluster size (voxels)Peak intensityXYZParaHippocampal_R18−24−18694−4.796Temporal_Inf_L−39−27−15245−3.935dReHo, dynamic regional homogeneity; MNI, Montreal Neurological Institute; ParaHippocampal_R, right parahippocampal gyrus; Temporal_Inf_L, left inferior temporal gyrus. Brain areas changed by dReHo across the two groups dReHo, dynamic regional homogeneity; MNI, Montreal Neurological Institute; ParaHippocampal_R, right parahippocampal gyrus; Temporal_Inf_L, left inferior temporal gyrus. Group differences in seed-based FC As shown in Table 4 and Figure 4, when seeds were placed in the third cerebellar vermis and the left inferior temporal gyrus, patients with mild to moderate CLBP exhibited decreased FC between the third cerebellar vermis and the right hippocampus (Hippocampus_R), and increased FC between the left inferior temporal gyrus and the left pallidus (Pallidum_L). FC values from seeds placed in the right parahippocampal gyrus showed no significant difference between the two groups (voxel-level P<0.005, cluster-level P<0.05, GRF corrected). Table 4Brain areas changed by seed-based FC across the two groupsROIAnatomical locationPeak MNI coordinatesCluster size (voxels)Peak t valueXYZVermis_3Hippocampus_R24−27−12221−4.678Temporal_Inf_LPallidum_L−24−6−91164.441FC, functional connectivity; Hippocampus_R, right hippocampus; MNI, Montreal Neurological Institute; Pallidum_L, left pallidus; ROI, region of interest; Temporal_Inf_L, left inferior temporal gyrus; Vermis_3, third cerebellar vermis. Brain areas changed by seed-based FC across the two groups FC, functional connectivity; Hippocampus_R, right hippocampus; MNI, Montreal Neurological Institute; Pallidum_L, left pallidus; ROI, region of interest; Temporal_Inf_L, left inferior temporal gyrus; Vermis_3, third cerebellar vermis. Brain regions with significant seed-based FC differences between mild to moderate CLBP and HC. Statistical parametric maps demonstrate altered FC in patients with mild to moderate CLBP compared to HCs. (A) Using the Vermis_3 as a seed, the CLBP group shows significantly lower FC in the Hippocampus_R. The cold color indicates a decrease in FC. (B) Using the Temporal_Inf_L as a seed, the CLBP group shows significantly higher FC in the Pallidum_L. The warm color indicates an increase in FC. Violin plots compare the mean z-FC values extracted from these target regions between the two groups. **, P<0.005. CLBP, chronic low back pain; FC, functional connectivity; HC, healthy control; Hippocampus_R, right hippocampus; Pallidum_L, left pallidus; ROI, region of interest; Temporal_Inf_L, left inferior temporal gyrus; Vermis_3, third cerebellar vermis. Correlation analysis between seed region sReHo, dReHo values and pain, mental health scale scores In this study, partial correlation analysis was used to analyze the correlation between the sReHo and dReHo values of the seed regions and the pain and mental health scores of patients with mild to moderate CLBP. Results showed that sReHo values in the third cerebellar vermis were negatively correlated with SAS scores (P=0.005), and no other correlations were found (all P>0.007; Figure 5, Table 5). The correlation between the sReHo value of the third cerebellar vermis and the SAS score in the patient group. Scatter plot illustrating the relationship between the sReHo value in Vermis_3 and the SAS score in patients with mild to moderate CLBP. Each dot represents an individual patient. The solid line represents the linear regression fit, and the shaded area indicates the 95% confidence interval. This significant negative correlation (r=−0.384, P=0.005) suggests that reduced regional synchronicity in the cerebellar vermis is associated with higher anxiety levels in these patients. CLBP, chronic low back pain; SAS, Self-Rating Anxiety Scale; sReHo, static regional homogeneity; Vermis_3, third cerebellar vermis. Table 5Correlation analysis between brain regions with ReHo differences and pain/psychological scoresBrain regionSASSDSPSQIVASSF-MPQ sensorySF-MPQ affectiveSF-MPQ totalVermis_3−0.384/0.005*−0.251/0.075−0.191/0.179−0.018/0.900−0.048/0.737−0.238/0.093−0.150/0.295ParaHippocampal_R−0.109/0.446−0.210/0.139−0.075/0.599−0.049/0.7300.087/0.5420.029/0.8410.073/0.612Temporal_Inf_L0.128/0.3710.093/0.5150.026/0.857−0.122/0.395−0.044/0.762−0.052/0.719−0.055/0.703Data are presented as r/P. *, significant level of P<0.007. ParaHippocampal_R, right parahippocampal gyrus; PSQI, Pittsburgh Sleep Quality Index; ReHo, regional homogeneity; SAS, Self-Rating Anxiety Scale; SDS, Self-Rating Depression Scale; SF-MPQ, Short-Form of the McGill Pain Questionnaire; Temporal_Inf_L, left inferior temporal gyrus; VAS, Visual Analogue Scale; Vermis_3, third cerebellar vermis. Correlation analysis between brain regions with ReHo differences and pain/psychological scores Data are presented as r/P. *, significant level of P<0.007. ParaHippocampal_R, right parahippocampal gyrus; PSQI, Pittsburgh Sleep Quality Index; ReHo, regional homogeneity; SAS, Self-Rating Anxiety Scale; SDS, Self-Rating Depression Scale; SF-MPQ, Short-Form of the McGill Pain Questionnaire; Temporal_Inf_L, left inferior temporal gyrus; VAS, Visual Analogue Scale; Vermis_3, third cerebellar vermis. Correlation analysis between seed-based FC metrics with between-group difference brain regions and pain, psychological scale scores Partial correlation analysis showed that in the patient group, the FC values between the left inferior temporal gyrus and the left pallidus were positively correlated with the VAS score (P=0.038), the SF-MPQ sensory score (P=0.006), the SF-MPQ affective score (P=0.005), and the SF-MPQ total score (P=0.001). After Bonferroni correction (threshold P<0.007), the correlation between FC values and the VAS score was not significant (P>0.007), but significant positive correlations remained with the SF-MPQ sensory, affective, and total scores (all P<0.007). FC values between the third cerebellar vermis and the right hippocampus in the patient group were not significantly correlated with pain or mental health scores (all P>0.007). See Table 6 and Figure 6 for details. Table 6Correlation analysis between brain regions with seed-based FC differences and pain/psychological scoresBrain regionSASSDSPSQIVASSF-MPQ sensorySF-MPQ affectiveSF-MPQ totalHippocampus_R−0.052/0.717−0.007/0.9620.102/0.4750.226/0.1110.085/0.5510.115/0.4200.114/0.425Pallidum_L0.198/0.1640.244/0.0850.149/0.2960.291/0.0380.381/0.006*0.386/0.005*0.445/0.001*Data are presented as r/P. *, significant level of P<0.007. FC, functional connectivity; Hippocampus_R, right hippocampus; Pallidum_L, left pallidus; PSQI, Pittsburgh Sleep Quality Index; SAS, Self-Rating Anxiety Scale; SDS, Self-Rating Depression Scale; SF-MPQ, Short-Form of the McGill Pain Questionnaire; VAS, Visual Analogue Scale. Correlation analysis between brain regions with seed-based FC differences and pain/psychological scores Data are presented as r/P. *, significant level of P<0.007. FC, functional connectivity; Hippocampus_R, right hippocampus; Pallidum_L, left pallidus; PSQI, Pittsburgh Sleep Quality Index; SAS, Self-Rating Anxiety Scale; SDS, Self-Rating Depression Scale; SF-MPQ, Short-Form of the McGill Pain Questionnaire; VAS, Visual Analogue Scale. Correlations between left inferior temporal gyrus-left pallidus FC and VAS/SF-MPQ scores in the patient group. Scatter plots illustrate the positive correlations between the FC values between the Temporal_Inf_L and the Pallidum_L and various pain-related metrics, including the VAS score, SF-MPQ sensory score, SF-MPQ affective score, and SF-MPQ total score. Each dot represents an individual patient, with solid lines indicating linear regression fits and shaded areas representing 95% confidence intervals. Notably, while all four metrics show significant uncorrected correlations, the associations with SF-MPQ sensory, affective, and total scores remained statistically significant after Bonferroni correction (P<0.007). These findings suggest that stronger functional connectivity involving the left pallidus is associated with heightened pain intensity and more severe negative pain-related experiences. FC, functional connectivity; Pallidum_L, left pallidus; SF-MPQ, Short-Form of the McGill Pain Questionnaire; Temporal_Inf_L, left inferior temporal gyrus; VAS, Visual Analogue Scale. Discussion Mild to moderate CLBP is highly prevalent yet often neglected due to its mild symptoms; it remains unclear whether mild to moderate CLBP alters the nervous system and if such changes relate to its high recurrence rate and increased depression risk. To explore this, we used dynamic and static ReHo metrics from rs-fMRI to extract local synchronous features of whole-brain spontaneous BOLD signals. We selected seeds in significantly abnormal brain regions to study long-range functional-connection changes in mild to moderate CLBP patients. Finally, we analyzed the interaction between these changes and questionnaire results related to pain, anxiety, depression, and sleep quality. To our knowledge, this study is the first to jointly apply dynamic and static ReHo and seed-based FC to evaluate changes in local and remote functional connectivity patterns in mild to moderate CLBP patients, and the first to notice subtle brain-activity changes under a low-pain state. Differences in clinical characteristics between groups Previous studies have indicated that patients with mild to moderate CLBP are at an increased risk of depression. When comparing patients with mild to moderate CLBP to HCs, we found that patients had higher SAS scores than the HC group, but no difference in SDS scores. We considered that this might be related to anxiety being a precursor to depression (34). Additionally, there was no significant difference in PSQI scores between the two groups, indicating that mild to moderate pain did not significantly affect patients’ sleep. Abnormal brain region in sReHo Neural homogeneity is regarded as a crucial and stable characteristic of brain activity at the microscopic level. ReHo indirectly measures it at the voxel level. In this study, patients with mild to moderate CLBP had lower sReHo in the third cerebellar vermis than the HC group, indicating reduced synchronicity and possible neuronal metabolic issues. The cerebellum is famous for its role in motor functions, including planning, coordination, and learning. But with scientific progress, other functions like cognition, emotion, and autonomic processing have been discovered (35). Now, much evidence shows the cerebellum is also involved in nociception and pain processing (36). The cerebellar vermis, part of the spinocerebellum, receives somatosensory inputs from the head and proximal body via ascending spinal pathways (37). Contemporary clinical research indicates that, unlike the cerebellar anterior lobe involved in motor control and the posterior lobe in complex cognition, the cerebellar vermis is more related to emotional processing (38). This is supported by findings such as Bobée et al.’s observation of autism-like symptoms in rats with cerebellar midline lesions (39) and the enhanced functional connection between the right parahippocampal gyrus and vermis in temporomandibular disorders patients, which correlates with depressive symptoms (40). We also believe that local cerebellar vermis dysfunction may be linked to pain-induced negative emotions. Specifically, the cerebellar vermis may act as a potential emotional regulatory node that modulates affective states through reciprocal connections with limbic structures via cerebello-thalamo-limbic pathways; disruption in this intrinsic modulatory circuitry may contribute to the exacerbation of emotional distress (36,38). In our study, the patient group had significantly higher anxiety scores than the HC group. Moreover, there was a negative correlation between the sReHo values of the third cerebellar vermis in the patient group and the SAS scores. We speculate that the reduced local activity in the third cerebellar vermis of patients with mild to moderate CLBP may be associated with emotional disorders such as anxiety caused by long-term chronic pain. Abnormal brain regions in dReHo The brain is highly dynamic, with unstable neural activity and rapid interactions. sReHo cannot fully capture the temporal changes in brain activity, while dReHo can better provide new information on the dynamic nature of neural activity. In this study, we identified two brain regions with significant dReHo changes: the right parahippocampal gyrus and the left inferior temporal gyrus, which were not found in sReHo studies and thus complement the static results well. The parahippocampal gyrus, located in the medial cortical region of the temporal lobe, serves as a pivotal hub bridging the limbic system and the default mode network (DMN). It plays a central role in memory encoding, emotional integration, and spatial cognition (41). A growing body of evidence highlights its involvement in pain processing (42). Studies have consistently demonstrated reduced gray matter volume in the parahippocampal gyrus across patients with CLBP, fibromyalgia, and postherpetic neuralgia (43-45). Using machine learning approaches, Kotikalapudi et al. (46) and Zhang et al. (47) identified that alterations in parahippocampal gyrus cortical thickness and ALFF strongly correlate with pain intensity, suggesting these metrics as predictive biomarkers for pain sensitivity. Furthermore, Zhou et al. observed significantly decreased ReHo in the left parahippocampal gyrus of lumbar disc herniation patients (48). Although their finding involved static ReHo in the left hemisphere, whereas ours involved reduced dReHo in the right parahippocampal gyrus, both studies converge in implicating parahippocampal dysfunction in this population. Collectively, these findings support the inference that intrinsic functional alterations within the parahippocampal gyrus may be implicated in the formation of pain-related memory traces. As a key component of the limbic system, the parahippocampal gyrus likely integrates and modulates affective and cognitive dimensions of pain. Notably, this region is involved in the cognitive evaluation of adverse stimuli by integrating sensory context with emotional responses (49). Reduced ReHo in this region may impair its inhibitory control over pain-associated aversiveness, potentially exacerbating negative emotional experiences such as depression and anxiety. The inferior temporal gyrus, located inferior to the lateral sulcus, contributes to the orbitofrontal limbic system, which primarily performs self-control of social emotions and management of emotional behaviors. Studies have shown that temporal lobe involvement is associated with unpleasant emotions, and the temporal lobe is mainly responsible for pain perception and emotional processing (50,51). In our study, patients with mild to moderate CLBP exhibited reduced temporal variability of regional activity in the left inferior temporal gyrus. Research further suggests that increasing cerebral blood flow in this region may improve pain-related emotional responses, potentially helping prevent depression in these patients (52). Both the parahippocampal gyrus and the inferior temporal gyrus are crucial for the DMN. Chronic pain can induce neuroplastic changes in the brain, affecting DMN stability (53). This instability is associated with symptoms like depression, anxiety, sleep issues, and poor decision-making in CLBP patients (53), which explains the higher anxiety scores in our study. Abnormal brain region in seed-based FC Our brain is a composite network of anatomically separate yet functionally connected regions. Prior studies indicate that CLBP-related pain and psychological symptoms stem from brain-network alterations and disrupted functional connections, not just local activity changes (54). Seed-based FC analysis aids in understanding these abnormal cross-regional brain communications (55). The results showed that, compared with the control group, patients with mild to moderate CLBP exhibited decreased functional connectivity between the third cerebellar vermis and the right hippocampus, and increased functional connectivity between the left inferior temporal gyrus and the left pallidus. The FC values of the left inferior temporal gyrus and left pallidus were positively correlated with SF-MPQ sensory, affective, and total scores; a positive correlation with the VAS score was also observed but was not significant after Bonferroni correction. The hippocampus is primarily responsible for memory formation and is also involved in processing pain stimuli and nociception (56). The cerebellum and hippocampus have polysynaptic connections (57). The cerebellar cortex’s inhibition widely and selectively modulates the hippocampus. In chronic pain, this modulation is closely tied to pain perception and emotional regulation. In a temporomandibular disorder study, patients had increased cerebellar vermis-hippocampal functional connectivity, linked to orofacial pain and depression severity (40). This relationship, mediated by pain, was found via mediation analysis. Our study found reduced functional connectivity between the third cerebellar vermis and right hippocampus in mild to moderate CLBP patients, unlike the temporomandibular disorder study. This difference may stem from our focus on patients with milder pain and less severe depression, affecting the cerebellar vermis-hippocampus connectivity strength. The cerebellum connects to the limbic system via brain-cerebellar pathways (cortico-ponto-cerebellar and cerebello-thalamo-cortical loops), playing a role in emotion regulation. In addition to the cerebello-hippocampal pathway, our seed-based FC analysis revealed another distinct dimension of whole-brain network alterations involving the left inferior temporal gyrus and the left pallidus. Although these two regions are not grouped within the same functional network, they may interact through structurally connected white-matter tracts and their respective neural circuits (58). The pallidus, a key part of the basal ganglia, is involved in pain sensory discrimination and emotional and cognitive aspects. It is also a main node in the cortical-basal ganglia-cortical circuit, and may relate to pain’s sensory, motor, emotional, memory, and cognitive aspects. Liu et al. observed increased functional connectivity of the pallidus in trigeminal neuralgia patients, proposing this as a potential adaptive response to prolonged pain episodes (59). Complementing these findings, a rodent study demonstrated that sustained nociceptive stimulation activates the ventral pallidum, which induces hyperalgesia, anxiety, and depression-like behaviors via ventral pallidum-derived cholinergic circuits (60). While cross-species differences must be considered, this mechanistic evidence aligns with our findings and provides valuable insights that may help contextualize our experimental outcomes. Our study further revealed that functional connectivity between the left inferior temporal gyrus and left pallidum exhibited significant correlations with SF-MPQ sensory, affective, and total scores (the VAS correlation was not significant after Bonferroni correction). Notably, stronger functional connectivity was associated with higher SF-MPQ scores and amplified negative pain-related experiences, suggesting a neurofunctional substrate for pain chronification. In summary, our study has identified distinct localized and long-range functional connectivity alterations within specific cerebello-hippocampal and cortico-basal ganglia networks in patients with mild to moderate CLBP. Through correlation analysis with clinical parameters, we consider these specific changes mainly related to two aspects: (I) in patients with mild to moderate CLBP, altered central pain processing may be present, manifesting as increased subjective pain sensitivity and altered intrinsic neural activity within pain-related regions; (II) patients with mild to moderate CLBP exhibit emotional disorders, with abnormalities in brain regions and networks involved in emotional processing. These two aspects may contribute to the persistence of the condition and the increased susceptibility to depressive symptoms in patients with mild to moderate CLBP. However, most patients are unaware of the impact of mild to moderate pain on their central nervous system. During the recruitment of HCs, we found that many patients with mild to moderate CLBP regarded themselves as completely healthy. Our findings may enhance patient awareness regarding the significance of mild to moderate pain, facilitating timely medical consultation to mitigate disease progression. Consequently, this proactive approach may help reduce recurrence rates, symptom duration, and the associated risk of depressive symptoms. Furthermore, we advocate for increased attention from academic and clinical communities toward this under-recognized patient population, emphasizing the critical role of early intervention and a multidimensional understanding of how mild to moderate pain impacts overall health. Limitations Our study has the following limitations. First, it only covers mild to moderate CLBP. Functional connectivity changes may vary with pain intensity. We compared only mild to moderate CLBP patients with HCs. Future studies should include severe CLBP patients and conduct subgroup analyses. Second, our cross-sectional study may not explain the different brain-activity patterns during various pain stages, especially when acute or subacute low back pain turns chronic. This also accounts for inconsistent results in previous studies. Future research should involve longitudinal surveys to examine functional connectivity changes during disease progression. Third, clinical medication usage was not strictly controlled or excluded before the MRI scans. Although our patients with mild to moderate pain primarily relied on intermittent, over-the-counter analgesics rather than regular heavy sedation, the potential confounding effects of such drugs on resting-state BOLD signals cannot be completely ruled out. Fourth, this study relied exclusively on resting-state baseline metrics and lacked task-based functional measures or quantitative sensory testing (QST). Consequently, we could not directly capture or evaluate evoked real-world pain experiences or precise sensory thresholds, which limits the clinical interpretation of our findings. This critical dimension will be prioritized in future research. Finally, as a single-center study, our results need more verification. Multi-center and trans-disciplinary cooperation should be done in future research to increase the sample size and improve result accuracy. Conclusions This study shows that patients with mild to moderate CLBP exhibit significant alterations in local brain function across several regions, along with decreased temporal variability in ReHo metrics. Seed-based FC further reveals that these central nervous system changes are both region-specific and involve broader cortical networks. These functional alterations are closely associated with the clinical characteristics of mild to moderate CLBP and provide valuable insights into the underlying neural substrates of the condition. Supplementary The article’s supplementary files as