Participants

This study prospectively collected clinical and imaging data of pSS patients at the Rheumatology and Immunology Department of the First People’s Hospital of Hangzhou from March 2021 to August 2023. Approval was obtained from the Research Ethics Committee (2021-023-01 and ZN-20230331-0053-01), adhering to the Helsinki Declaration, with all participants providing written informed consent.

Inclusion criteria for pSS patients were as follows: (1) Patients diagnosed with pSS according to the European-American consensus criteria; (2) Absence of psychiatric and psychological disorders prior to pSS diagnosis; (3) Absence of other connective tissue diseases such as systemic lupus erythematosus, antiphospholipid syndrome, systemic sclerosis, and rheumatoid arthritis; (4) Age between 20 and 75 years; (5) Right-handedness; (6) Absence of motor, auditory, and visual system disorders, no language barriers, and normal corrected vision; (7) Voluntary participation in the study.

Inclusion criteria for healthy controls (HCs) were: (1) Age between 20 and 75 years; (2) Right-handedness; (3) Good physical condition, no history of tumors or psychiatric disorders; (4) Absence of motor, auditory, and visual system disorders, no language barriers, and normal corrected vision; (5) Voluntary participation in the study.

Exclusion criteria were: (1) Severe hypertension, diabetes, and other diseases affecting brain function; (2) Previous history of cerebral organic lesions, head trauma, or invasive surgery; (3) Contraindications to MRI scans; (4) History of alcohol dependence or substance abuse.

After quality control assessment, 68 pSS patients and 69 HCs were included in the final analysis following the exclusion of invalid data (Fig. 1).

Fig. 1

Exclusion of invalid data. FD, Framewise displacement

Disease duration, disease activity assessment, and neuropsychological assessments

The time of the first diagnosis of pSS was defined as the onset of the disease. Therefore, the disease duration was calculated as the period from disease onset to inclusion in the study.

The extent of inflammatory activity in systemic tissues and organs caused by pSS is referred to as disease activity. In this study, disease activity was assessed using the EULAR Sjögren’s Syndrome Disease Activity Index (ESSDAI), developed by the European League Against Rheumatism in 2009. The ESSDAI evaluates disease activity across twelve domains: systemic symptoms, lymph nodes, joints, glands, skin, lungs, kidneys, muscles, central nervous system, peripheral nervous system, hematological system, and serological markers. The total score reflects the overall disease activity of the patient. Higher scores in a specific domain or overall indicate greater disease activity in that domain or the entire system.

All participants underwent a series of neuropsychological tests within 2 h after the MRI examination. Cognitive function was evaluated using the Mini-Mental State Examination (MMSE). Anxiety and depression states were assessed using the Self-Rating Anxiety Scale (SAS) and the Self-rating Depression Scale (SDS), respectively. Attention and information processing speed assessed using with the Digit Symbol Test (DST), whereas psychomotor ability was evaluated with the Number Connection Test-Type A (NCT-A). All assessments were conducted by the same physician to ensure consistency.

MRI parameters

Participants underwent MRI data acquisition using a 3.0 T MRI scanner (Siemens, MAGNETOM Verio, Germany) with an 8-channel phased-array head coil. They were instructed to remain awake with their eyes closed and to wear foam pads and earplugs to minimize head motion and scanner noise interference.

Initial scans included T1-weighted imaging (T1WI), T2-weighted imaging (T2WI), T2 fluid attenuated inversion recovery (T2 FLAIR), and diffusion-weighted imaging (DWI) to rule out intracranial pathologies, with the following parameters: T1WI: repetition time (TR)/echo time (TE) = 500ms/8.5ms, matrix = 120 × 256, field of view (FOV) = 230 mm×230 mm, slice thickness = 5 mm, slice gap = 1.5 mm, number of slices = 20. T2WI: TR/TE = 5000ms/117ms, matrix = 248 × 320, FOV = 220 mm×220 mm, slice thickness = 5 mm, slice gap = 1 mm, number of slices = 20. T2 FLAIR: TR/TE = 8000ms/94ms, matrix = 186 × 256, FOV = 230 mm×230 mm, slice thickness = 5 mm, slice gap = 1 mm, number of slices = 20. DWI: TR/TE = 5100ms/100ms, matrix = 192 × 192, FOV = 230 mm×230 mm, slice thickness = 5 mm, slice gap = 0.4 mm, number of slices = 20, b-value = 1000s/mm².

Anatomy images were obtained using a three-dimensional T1-weighted imaging (3D-T1WI) sequence with the following parameters: TR = 1900ms, TE = 2.52ms, inversion time = 900ms, flip angle = 9°, FOV = 256 mm×256 mm, slice thickness/slice gap = 1/0 mm, matrix = 256 × 256, with a total of 176 sagittal slices.

Rs-fMRI data were acquired using a gradient-echo imaging sequence with the following parameters: TR = 2000ms, TE = 30ms, slice thickness/slice gap = 3.2/0 mm, FOV = 220 mm×220 mm, flip angle = 90°, with 250 time points collected per scan.

Image preprocessing

Examine the quality of all 3D-T1WI and rs-fMRI images, excluding incomplete or artifact-affected images. Rs-fMRI data preprocessing utilized the Data Processing and Analysis of Brain Imaging (DPABI) 6.2 toolbox on MATLAB (2018b, MathWorks, Natick, MA, United States).

The following were the specific image preprocessing steps: (1) Removal of the first 10 time points of each rs-fMRI dataset to ensure MRI signals reached a stable state; (2) Temporal alignment of the remaining rs-fMRI data; (3) Correction of head motion, while excluding subjects with maximum head displacement exceeding 3 mm, rotation exceeding 3°, or frame-wise displacement exceeding 0.5; (4) Registration of the subject’s structural image to the corresponding mean functional image; (5) Segmentation of the registered structural image into gray matter, white matter, and cerebrospinal fluid to obtain registration matrices between the mean functional image and the standard space; (6) Transformation of subject data from the original space to the Montreal Neurological Institute standard space by applying the registration matrices onto the functional image, with voxel size resampled to 3 mm×3 mm×3 mm; (7) Removal of linear drift by linear regression to eliminate signal changes due to machine heating during continuous operation; (8) Regression of covariates, including Friston-24 head motion parameters, cerebrospinal fluid signal, and white matter signal; (9) Band-pass filtering of rs-fMRI data using a 0.01 ∼ 0.08 Hz bandpass filter to remove low-frequency linear drift and high-frequency physiological noise such as respiration and heartbeat, with this step specifically applied for ReHo value calculation.

Static indicators calculation

A fast Fourier transform was performed on whole-brain voxels to convert the BOLD signal into the frequency-domain power spectrum. The square root of the power spectrum was calculated at each frequency, and the average value within the range of 0.01 to 0.08 Hz was used to calculate the sfALFF metric. sReHo values were calculated using Kendall’s coefficient of concordance method, which assesses the temporal synchronization between the time series of a voxel and its 26 neighboring voxels to generate whole-brain sReHo values for each participant. To ensure comparability, sfALFF and sReHo values of individual voxels were normalized to the mean values of the entire brain. Spatial smoothing was applied using a Gaussian kernel with a full width at half maximum (FWHM) of 6 mm to minimize incomplete registration and improve the image signal-to-noise ratio, thereby improving the reliability of the results.

Dynamic indicators calculation and validation

The analysis of dfALFF and dReHo was performed using the DPABI-based dynamic analysis toolbox. Dynamic indicators were calculated using a sliding window approach, recognized for its sensitivity in detecting temporal changes and assessing whole-brain indicator variability [23]. The length of the sliding window is a critical parameter. It should be sufficiently large to enable robust analysis of the lowest frequencies of interest in the signal, yet small enough to capture transient signals [24]. For our analysis, a sliding window length of 50 TRs (100s) was used, with a moving step of 2 TRs (4s). Subsequently, the standard deviation of fALFF and ReHo for each voxel within the time window was calculated to generate the dfALFF and dReHo matrices, which characterize the dynamic variations in fALFF and ReHo. Lastly, consistent with the static indicators, spatial smoothing was performed using a Gaussian kernel with an FWHM of 6 mm.

To ensure the reliability of the results, dfALFF and dReHo results were further validated using dynamic methods with sliding window lengths of 50 TRs (100s) and moving steps of 5 TRs (10s), as well as sliding window lengths of 100 TRs (200s) with a moving step of 2 TRs (4s).

Statistical analysis

The Shapiro-Wilk test was performed to evaluate the distribution of continuous variables using SPSS 25.0 software. Normally distributed data were presented as mean ± standard deviation (\(\:\stackrel\pm\:s\)) and compared between groups using the independent samples t-test. Non-normally distributed data were presented as median (interquartile range) [M (P25, P75)] and compared between groups using the Wilcoxon rank sum test. The chi square test is used for inter group comparison of categorical variables. Statistical significance was set at P < 0.05. In DPABI software, a two independent samples t-test was used to compare pSS patients and HCs on sfALFF, sReHo, dfALFF, and dReHo indicators within the classic frequency band (0.01~0.08 Hz). Group comparisons were conducted applying Gaussian random-field theory (GRF, voxels P < 0.001, clusters P < 0.05). sfALFF, sReHo, dfALFF, and dReHo values were extracted from brain regions exhibiting significant differences between pSS and HC groups. Subsequently, partial correlation analysis was used to evaluate the relationships between disease duration, treatment, laboratory indicators, and neuropsychological scores with these significant brain regions. Gender, age, years of education, and head motion were included as covariates in the analysis. Statistical significance was assessed using a Bonferroni-corrected threshold of P < 0.008 (0.05/6).