Article Text
PDF
Abstract
Objective This study explored the dynamic functional connective (DFC) alterations in patients with rheumatoid arthritis (RA) and investigated the correlation between the neuropsychiatric symptoms, peripheral inflammation and DFC alterations.
Method Using resting-state functional MRI, we investigated the DFC based on spatial independent component analysis and sliding window method for 30 patients with RA and 30 healthy controls (HCs). The Spearman correlation was calculated between aberrant DFC alterations, Montreal Cognitive Assessment (MoCA), Hospital Anxiety and Depression Scale (HAD), C reactive protein (CRP) and erythrocyte sedimentation rate (ESR). Diagnostic efficacy of indicators was assessed using receiver operating characteristic analysis (ROC).
Results Three dynamic functional states were identified. Compared with HC, patients with RA showed reduced FC variabilities between sensorimotor network (SMN) and insula, SMN and orbitofrontal cortex, which were the crucial regions of sensory processing network. The above FC variabilities were correlated with the MoCA, HAD, CRP and ESR in patients with RA. Additionally, the CRP and ESR were negatively correlated to MoCA and positively related to HAD in patients with RA. The ROC analysis results showed that MoCA, HAD and FC variabilities of the sensory processing network could distinguish patients with RA from HC and also identify patients with RA with high ESR.
Conclusion Our findings demonstrated that abnormal DFC patterns in sensory processing networks in patients with RA were closely associated with peripheral inflammation and neuropsychiatric symptoms. This indicates that the dynamic temporal characteristics of the brain functional network may be potential neuroimaging biomarkers for revealing the pathological mechanism of RA.
- rheumatoid arthritis
- inflammation
- psychology
- magnetic resonance imaging
Data availability statement
Data are available on reasonable request.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Rheumatoid arthritis (RA) has cognitive impairment and changes in brain function, with a significantly increased risk of neuropsychiatric comorbidities.
WHAT THIS STUDY ADDS
This study revealed the abnormal dynamic temporal properties of the sensory processing network in patients with RA and its association with peripheral inflammation and neuropsychiatric disorders.
Altered dynamic functional network connectivity could significantly distinguish patients with RA from health subjects, which suggests the dynamic temporal characteristics of the brain functional network may be potential neuroimaging biomarkers for revealing the pathological mechanism of RA.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
This provides a potential neuroimaging technique for studying the pathological mechanisms of RA. It may be possible to visualise the neuroinflammation in patients with RA by analysing the temporal characteristics of the dynamic function connection of the brain networks.
Introduction
Rheumatoid arthritis (RA) is a systemic autoimmune disorder, distinguished by the presence of inflammatory arthritis and additional manifestations beyond the joints.1 It is regarded as the most prevalent chronic inflammatory arthritis, with a prevalence estimated to range from 0.5% to 2% in the population, and women are disproportionately affected, with a 2–4 times higher incidence than men.2 Multiple research studies have indicated that individuals with RA experience cognitive decline and alterations in brain activity.3–5 A previous study has shown that patients with RA display altered patterns of brain connectivity.6 However, the investigation of functional connectivity (FC) in patients with RA predominantly used static methodologies, which assumed the constancy of brain connectivity patterns over time. Recent research has unveiled that brain activity is highly dynamic, and FC consistently fluctuates during scanning.7 Dynamic FC (DFC) captures temporal dynamic patterns and time-varying characteristics of neurological diseases and offers a more comprehensive understanding compared with traditional static FC.8 It has been suggested by some scholars that the detection of abnormal DFC could provide valuable understanding of the underlying pathophysiological mechanisms of Alzheimer’s Disease (AD).9 However, only static brain network functional connections have been reported in RA brain network studies, and the dynamic brain function changes in patients with RA remain unclear.
Additionally, individuals with RA are at a significantly increased risk for developing neuropsychiatric comorbidities. It is worth mentioning that the occurrence of major depressive disorder among individuals with RA has been approximated at 16.8%, exceeding that of the overall populace.10 In addition, patients with RA demonstrate higher levels of anxiety and cognitive decline in comparison to healthy individuals.11 12 It is believed that these neurological symptoms arise from neuroinflammation, which is triggered by peripheral inflammation.13 Infiltration of immune cells from the blood or inflammatory activation of endothelial cells can facilitate the entry of peripheral inflammatory mediators into the central nervous system (CNS) via the blood–brain barrier (BBB) or choroid plexus.14 15 Previous experimental studies involving small cohorts of healthy individuals exposed to inflammatory substances have revealed temporary changes in brain FC.16 Our previous research has also shown that altered brain functional signal in patients with RA with visuospatial cognitive impairment correlates with inflammation levels.17 The importance of recognising and measuring neuropsychiatric comorbidities in patients with RA goes beyond simply addressing the mental health disorder. It also includes reducing the negative effects that depression has on the outcomes of RA, such as functional progression and response to treatment.
Since DFC has the ability to capture temporal fluctuations in network connections during various physiological or pathological brain states, thereby yielding additional insights into the mechanisms underlying neuroinflammation in RA. However, no study has explored the relationship among brain dynamics, peripheral inflammation and neuropsychiatric disorders in RA. Therefore, the objective of this study is to identify the significant abnormal DFC patterns in patients with RA when compared with healthy controls (HCs), and to examine the correlation between the neuropsychiatric symptoms, peripheral inflammation and DFC alterations to further understanding the neuropathology of RA.
Method
Participants
For this cross-sectional investigation, the study included two groups of subjects: patients with RA and HCs. RA were chosen as subjects from the Department of Rheumatology at the First Affiliated Hospital of Shantou University Medical College between December 2019 and April 2021. The eligibility criteria for the RA group included meeting the 2010 American College of Rheumatology and European Alliance Against Rheumatology classification criteria,18 encompassing individuals aged between 18 and 60 years, having education more than 6 years, being right-handed, having normal/corrected-to-normal vision. The study’s exclusion criteria encompassed other types of rheumatic immune diseases, organic brain disorders, mental illness, drug and alcohol dependence, MRI contraindications, and head movement of more than 2 mm or 2° during MR scan. Initially, a total of 35 patients with RA were included in the experiment, among them, 4 cases including 2 males and 2 females were excluded from this study due to head motion angle exceeding 2° during scanning, and 1 male patient was excluded for incomplete MRI data collection because he could not tolerate the noise during MRI scan. Finally, 30 female patients were included. All patients were treated with medication, including glucocorticoid (methylprednisolone/prednisone), Disease-Modifying Antirheumatic Drugs (DMARDs: hydroxychloroquine/methotrexate/sulfasalazine) and Nonsteroidal Antiinflammatory Drugs (NSAIDS). The specific medication regimen for each patient with RA is shown in online supplemental table S1. The HC group consisted of 30 female subjects recruited in the same period, whose age, sex handedness and educational level were matched with the above patients with RA.
Supplemental material
Questionnaire scale assessment and clinical data
The general cognitive status of all participants was evaluated by the Montreal Cognitive Assessment (MoCA).19 The mental state of the participants was assessed using the Hospital Anxiety and Depression Scale (HAD), and scoring for anxiety and depression was conducted independently, with a higher score indicates a more severe mood disorder than lower scores. Scores range from normal=0–7, mild=8–10, moderate=11–15 and severe=16–21 for HAD subscales.20 To evaluate the extent of inflammation in RA, measurements of C reactive protein (CRP) and erythrocyte sedimentation rate (ESR) were obtained.
Functional MRI acquisition and data preprocessing
The imaging data were obtained using a 1.5T MR scanner manufactured by GE Medical Systems in Milwaukee, Wisconsin, USA. A typical GE head coil was used, and headphones and earplugs were given to minimise disruption caused by ambient noise. High-resolution structural images were obtained using three-dimensional T1-weighted magnetisation and fast spoiled gradient recalled (FSPGR) sequences. The imaging parameters for these sequences were as follows: repetition time (TR) of 1.6 ms, echo time (TE) of 5.1 ms, flip angle (FA) of 20°, matrix size of 256×256, field of view (FOV) of 256mm×256 mm, slice thickness of 1.3 mm and interleaved acquisition of 244 slices. Additionally, gradient echo plane imaging sequences were used to acquire rest-state functional images for a duration of 6 min. The imaging parameters for these sequences were as follows: TR of 2000ms, TE of 45 ms, FA of 90°, matrix size of 64×64, FOV of 250 mm×250 mm, slice thickness of 6.0 mm and interleaved acquisition of 20 axial slices.
The resting-state-functional MRI (fMRI) data underwent preprocessing using the DPARSF V.4.0 toolbox (https://www.rfmri.org/dpabi). The preprocessing steps encompassed the elimination of the initial 10 volumes, adjustment for slice timing, correction for head motion, spatial smoothing using an FWHM of 6 mm and then normalised to the standard coordinates of the MNI152 (Montreal Neurologic Institute standard space). Participants exhibiting conspicuous head movements exceeding 2 mm or rotation surpassing 2° during scanning were excluded. A total of 60 subjects were included for subsequent analyses.
Group independent component analysis and component selection
After preprocessing, the fMRI data of all participants were decomposed using a standardised pipeline for spatial group independent component analysis (GICA) implemented in the fMRI Toolbox (GIFT). A principal component analysis was first used to reduce participant-specific data to 120 principal components. Using the expectation maximisation algorithm, the participant-reduced data were further decomposed into 100 independent components (ICs). ICASSO,the software for investigating the reliability of ICA estimates by clustering and visualization, was used to construct the aggregated spatial map using the Infomax algorithm, with 20 repetitions. Lastly, spatiotemporal regression inverse reconstruction with Fisher’s Z transformation was used to obtain the spatial maps and time courses of the subjects.21 We employed a previously established method to identify pertinent intrinsic connectivity networks from a pool of 100 ICs. This method involved locating peak activation coordinates primarily within grey matter, while minimising spatial overlap with vascular, ventricular and edge regions associated with artefacts.22 Additionally, we considered components whose time course was predominantly characterised by low-frequency fluctuations, with a power ratio of <0.10 Hz to 0.15–0.25 Hz.23 Ultimately, we have chosen 50 ICs that bear resemblance to those identified in prior decompositions employing a high model order ICA.7 22 24–26 The ICs were classified into 9 resting-state brain networks (RSNs): auditory network (AN: 5 ICs); visual network (VN: 10 ICs); sensorimotor network (SMN: 8 ICs); salient network (3: 3 ICs); default mode network (DMN; 7 ICs); left frontoparietal network (LFP: 5 ICs); right frontoparietal network (RFP: 4 ICs); dorsal attention network (DAN: 5 ICs); ventral attention network (VAN: 3 ICs), as shown in figure 1.
Spatial maps of the 50 independent components (ICs). 50 ICs were selected and divided into 9 resting state networks: auditory network (AN: 5 ICs); visual network (VN: 10 ICs); sensorimotor network (SMN: 8 ICs); salient network (SN: 3 ICs); default mode network (DMN; 7 ICs); left frontoparietal network (LFP: 5 ICs); right frontoparietal network (RFP: 4 ICs); dorsal attention network (DAN: 5 ICs); ventral attention network (VAN: 3 ICs).
In order to enhance the reduction of noise sources, supplementary postprocessing methods were implemented on the time courses of the 50 ICs. These methods encompassed detrending, involving linear, quadratic and cubic trends, regression analysis of the six parameters associated with head movement, despiking and low-pass filtering with a cut-off frequency of 0.15 Hz.
Dynamic functional network connectivity analysis
The DFC analysis was conducted using the sliding-window approach with the temporal dynamic FNC toolbox, as implemented in GIFT. In accordance with previous studies, a rectangular window measuring 44s (22 TR) was convolved with a Gaussian function (σ=3 TR) to segment the time courses by sliding across the entire scan at intervals of 1 TR.7 Each participant produced a total of 148 consecutive windows. For each window, a 50×50 regularised inverse covariance matrix was computed to illustrate the changing patterns of covariance (correlation) between components. The graphical LASSO technique was employed to apply an L1 penalty on the precision matrix, with the objective of promoting sparsity, and this procedure was repeated 100 times. After calculating the DFC, the Fisher-Z transformation was applied to all FC matrices, which were then adjusted for nuisance variables. Subsequently, K-means clustering was performed on the FC matrices of each participant to explore the frequency and structure of recurring FC patterns, known as states.27 The temporal properties of DFC states were subsequently calculated for each participant, incorporating metrics such as mean dwell time, fraction time and number of transitions. The temporal variability of FC was determined by the variance of FC across all matrices, which plays a critical role in the dynamical integration and coordination of neural systems.
Statistical analysis
The differences in dFC between the patients with RA and HSs groups were analysed using the stats module of the GIFT software package, and other data were analysed using SPSS V.26.0 software. We used two-sample t-tests to compare the differences on the age and years of education between two groups. We analysed differences in MoCA, HAD-A, HAD-D and temporal properties of DFC states between patients with RA and HCs using Mann-Whitney U-tests. Correlation analysis was conducted using Spearman’s correlation. The study assessed the predictive values of the MoCA, HAD and FC variability through receiver operating characteristic (ROC) curve analysis. This analysis was conducted to compare the RA group with the HC group, as well as to compare patients with RA with higher ESR (≥20 mm/hour) and patients with RA with lower ESR (<20 mm/hour). The area under the ROC curve (AUC) was used to measure the overall performance of the diagnostic tests, with values ranging from 0 to 1. A higher AUC value indicates better performance.28 The measurement data were presented as means±SD. Statistical significance was determined at a significance level of p<0.05, and the familywise error method was used for multiple comparison correction.
Result
Demographic, neuropsychological and clinical characteristics
The demographic and clinical indicators of participants are presented in table 1. No disparities in age and educational attainment were observed between patients with RA and HC. Patients with RA exhibited higher HAD-D and HAD-A scores compared with HC individuals. Patients with RA had a lower average MoCA score compared with HC (p<0.001). The ESR and CRP values in patients were 42.63±29.81 mm/hour and 18.94±17.33 mg/L, respectively.
Demographic and clinical characteristics
Dynamic functional brain networks
The study determined that the optimal value for k was 3, as determined by the k-means algorithm. This analysis revealed the presence of three recurring DFC states across individual scans and subjects. The top 5% strongest connections for each DFC state are shown in figure 2, indicating distinct connectivity patterns. Specifically, states 1 and 2 exhibited high self-connection strength in the VN, but state 1 had sparser internetwork connections and a lower occurrence rate compared with state 2. State 3 demonstrated a strong connection between the DMN. Patients with RA showed lower fraction time and mean dwell time in state 2 compared with HC patients. State 3 exhibited the opposite trend. In addition, patients with RA had fewer transitions than HC patients (figure 3A–C).
The three states identified by the k-means clustering method and the percentage of the occurrences and total number of windows are listed above each centroid. (A) Cluster centroids for each state. (B) The strongest 5% connections of each state. AN, auditory network; DAN, dorsal attention network; DMN, default mode network; LFP, left frontoparietal network; RFP, right frontoparietal network; SN, salient network; VAN, sensorimotor network.
Differences in the temporal properties and functional connectivity variability of each state between RA and HC groups. (A–C) Temporal properties: fraction time; mean dwell time; the number of transitions between states; (D–E) Differences functional connectivity variability of each state between RA and HC groups. HC, healthy control; RA, rheumatoid arthritis. *p < 0.05; **p < 0.001.
As shown in figure 3D–E and table 2, group differences were evident in FC variability across the three identified connectivity states. In state 1, individuals with RA exhibited heightened variability in FC among the SMN-SMN, LFP-VN and LFP-SMN, while experiencing reduced FC variability between SMN-SN. In state 2, patients with RA showed increased FC variability between the VN-SMN, VN-LFP, VN-RFP, SMN-VAN, while decreased FC variability between the LFP and RFP. Additionally, in state 3, the FC variability between the LFP-VN and SMN-DAN increased in patients with RA but decreased between the VN-SMN, AN-SMN, LFP-LFP, AN-DAN, LFP-DAN and SMN-VAN.
Difference functional connectivity variabilities in states 1, 2 and 3 between RA and HC groups
Correlation and ROC result
As shown in figure 4, both the RA and HC groups exhibited a positive association between HAD-D and HAD-A, while displaying a negative correlation between HAD-D and FC variability between the postcentral gyrus (PostCG) and right orbital inferior frontal gyrus (ORBinf.R). Additionally, the MoCA score was negatively associated with both HAD-A and HAD-D scores. Among patients with RA, we observed that the FC variability between supplementary motor area (SMA) and insula, precentral gyrus (PreCG) and right orbital middle frontal gyrus (ORBmid.R), PostCG and ORBinf.R were significantly positively related to MoCA score, while negatively associated with HAD-A, HAD-D, CRP and ESR of patients with RA. No such correlations were found in HCs. The ROC results showed that MoCA, HAD-A, HAD-D, FC variability of SMA-Insula, PreCG-ORBmid.R and PostCG-ORBinf.R could significantly distinguish patients with RA from HC (Ps<0.001), and AUC values were between 0.6911 and 0.9333 (figure 5A). Furthermore, these six parameters could identify patients with RA with higher ESR, with AUC values ranging 0.7174–0.8407 (Ps<0.05), shown in figure 5B.
Results of correlation analysis among functional connective functional connectivity variability and clinical characteristics in both RA and HC group. CRP, C reactive protein; ESR, erythrocyte sedimentation rate; HAD, Hospital Anxiety and Depression; HC, healthy control; MoCA, Montreal Cognitive Assessment; RA, rheumatoid arthritis. *p < 0.05; **p < 0.001.
Receiver operating characteristic (ROC) analysis. (A) The area under the ROC curve (AUC) values of the six parameters in diagnosing patients with RA. (B) The AUC values of six parameters in identifying patients with RA with high ESR. ESR, erythrocyte sedimentation rate; HAD, Hospital Anxiety and Depression; MoCA, Montreal Cognitive Assessment.
Discussion
This study investigated the DFC alterations in patients with RA and assessed their relationship with the neuropsychiatric symptoms and peripheral inflammation. The reduced FC variabilities between SMN and insula, SMN and orbitofrontal cortex (OFC), which were the crucial regions of sensory processing network, were correlated with the neuropsychiatric symptoms and peripheral inflammation in patients with RA. The ROC analysis results showed that MoCA, HAD and FC variabilities of the sensory processing network could distinguish patients with RA from HC and also identify patients with RA with high ESR.
DFC changes have been linked to particular cognitive states, mental disorders and neurological conditions.29–31 Using dynamic connectivity measurements could potentially offer a more precise evaluation of brain connectivity, enabling us to define chronomic-based biomarkers with greater accuracy and capture disease information more effectively as time progresses.32 This study demonstrated main damaged FC variability between SMN, insula and OFC in patients with RA, these brain regions are the crucial areas of the sensory processing network. The sensory processing network comprises functional architectures encompassing exteroceptive, interoceptive and motor networks, in which the lateral/orbital cortex mainly processes external sensory information, and the ventromedial cortex mainly processes internal sensory information.33 In terms of interoception, the insular cortex assumes a crucial role in directing attention towards interoceptive stimuli. The insula, a cortical region, has functional links to the basal ganglia, amygdala and various parts of the cerebral cortex. These connections are responsible for the basic involvement of the insula in emotional processing, sensorimotor and cognitive functions.34 The SMN is responsible for the intricate planning of contralateral extremity movements within the motor regions.35 The aberrant brain regions within the RA brain network serve as significant nodes within the sensory processing network, indicating the presence of abnormal sensory processing function in patients with RA. Sensory processing is a multifaceted nervous system process that involves the reception, integration and response to stimuli from the body and the surrounding environment, and therefore, plays a crucial role in survival and neurological diseases.36 In order to comprehend and react to the ever-changing surroundings, the brain employs the interpretation of sensory signals, taking into account factors such as locomotion, previous experiences and emotional states.36 Consequently, sensory processing is not just a simple perception of internal or external stimuli but also includes cognitive manifestations that facilitate the generation of an appropriate response to the perception of this stimulus. Sensory abnormalities, a prevalent manifestation observed in neurological disorders such as brain injury, developmental disorders and psychiatric disorders.37 Multiple studies have shown that RA has harmful effects on the brain function.4 6 However, there is a lack of research on the dynamic brain network in patients with RA. The focus of this study was to investigate the dynamic networks of the brain and identify the altered and temporal characteristics of the functional networks of the brain in patients with RA. To provide some basis for brain function changes in patients with RA. Additionally, the dysfunction of FC variability in sensory procession network was linked to peripheral inflammation and mood disorders in patients with RA. The increasing burden of chronic inflammatory diseases on global health systems is a result of the central role inflammation plays in most chronic pathologies. RA is distinguished by persistently elevated levels of CRP and ESR, both widely employed as biomarkers for inflammation. Presently, it is recognised that the irreversible dissociation of CRP into monomers is concomitant with its biological activation, thereby stimulating inflammatory processes that may potentially contribute to neurodegeneration.38 Our study observed a negative association between FC variability in the SMN, the insula and the OFC with inflammatory markers (CRP/ESR) in patients with RA. Furthermore, a comparison between patients with RA and HC revealed 17 abnormal FC in the RA group, along with significant differences in temporal properties, such as a lower number of state transitions. These alterations in FC and temporal properties imply the presence of cognitive flexibility impairment in patients with RA, which aligns with the cognitive dysfunction observed in this investigation. Impairment of cognitive flexibility has been documented in various chronic inflammatory conditions, including irritable bowel syndrome.39 40 This further proves the pathogenesis of neuroinflammatory in RA. There are many ways that RA can affect the nervous system, including peripheral vasculitis, joint degeneration and drug-induced neuropathy.41 Moreover, there exists substantiating evidence linking systemic inflammation-induced microvascular cerebral damage to the onset of dementia. AD is another pathology linked to RA, and several studies have highlighted the significance of persistent inflammation in the formation of brain damage in small blood vessels, activation of microglia and the occurrence of dementia.42 43 According to a meta-analysis, CRP can cross the blood-brain barrier or be discharged from injured micro-vessels to enter the brain.44 These findings indicate that inflammation-induced brain damage can be assessed by DFC abnormalities. Furthermore, a separate investigation has uncovered atypical linkage between the insula and various other regions of the brain, which were observed to have a correlation with the ESR in individuals with fibromyalgia who also have RA.45 This is consistent with the results of sensory motor network impairment in patients with RA in this study.
Furthermore, mood disorders, particularly anxiety and depression, are frequently observed in RA. There is a 13%–20% prevalence of depression in patients with RA while anxiety affects 21%–70% of individuals with RA.46 According to Whitehouse et al, anxiety correlates with cognitive impairment in persons with multiple sclerosis, inflammatory bowel disease and RA.47 Our results showed the anxiety was associated with the cognition (MoCA), inflammatory factors (ESR/CRP) and the FC variability in RA. A study by Michelsen et al found that depression and anxiety decreased joint remission chances.48 The complete understanding of the biological mechanisms that link RA and mood disorders is lacking. Flodin et al discovered that patients with RA exhibited heightened brain connectivity primarily in the SMN.49 Another study reported that patients with RA displayed significantly reduced activation in regions associated with somatosensory processing when subjected to painful joint stimulation.3 Anxiety is a phenomenon that manifests in the body and has been observed to induce changes in various sensorimotor functions. Individuals with anxiety disorders frequently exhibit symptoms of increased sensitivity to external stimuli, and there seems to be a notable convergence in the epidemiology of anxiety and sensory processing sensitivity.50 In a recent study conducted by Brown et al, the researchers examined the findings of five distinct studies, with a particular emphasis on the sensory processing capabilities of individuals diagnosed with a range of mental illnesses, which suggest that individuals with mental health conditions demonstrate discernible sensory processing patterns in contrast to those without such conditions.51 Therefore, it can be posited that sensory processing difficulties are a non-specific transdiagnostic phenotype frequently linked to various psychiatric disorders. Further investigation into the importance and role of sensory processing challenges in patients with RA with cognitive dysfunction holds promise for improving long-term prognosis and treatment outcomes.
Nevertheless, it is imperative to recognise the constraints inherent in this study. First, the study population is relatively small and the explanation of the results is limited. Second, the study only includes female participants, and the sex difference of DFC should be researched in a further study. Third, other confounding factors such as sleep disturbance and the drugs medication of patients with RA that could affect FC of brain networks were not considerate. Future inquiries should strive to rectify these limitations by augmenting the sample size and accounting for influential factors.
Conclusion
This study revealed the abnormal dynamic temporal properties of the sensory processing network in patients with RA and its association with peripheral inflammation and neuropsychiatric disorders, which suggested that the sensory processing dynamic brain network can be regarded as an imaging representation of neuroinflammation in patients with RA. This provides a potential neuroimaging technique for studying the pathological mechanisms of RA.
Supplemental material
Data availability statement
Data are available on reasonable request.
Ethics statements
Patient consent for publication
Ethics approval
This study was approved by the Ethics Committee of the First Affiliated Hospital of Shantou University Medical College (B-2021-237). All participants gave written informed consent. We included data from December 2019 to April 2021.
Acknowledgments
The authors would like to thank all the participants and the assistance from department of rheumatology, the First Affiliated Hospital of Shantou University Medical College.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
YZ, ZH and SM are joint first authors.
YZ, ZH and SM contributed equally.
Contributors YZ and LX: design, conceptualisation, methodology and drafting for manuscript. JH: data evaluation. JP, S-XH and Z-DH: recruiting participants and follow-up. RG and JY: data processing and data analysis. ZH and ZZ: neuropsychological assessment. ZL: experiment preparation and data acquisition. LX and SM: supervision and funding acquisition. LX: responsible for the overall content as the guarantor.
Funding This study was funded by the grants from the National Natural Science Foundation of China (Grants Nos. 82004468, 82274657, 81774395); China Postdoctoral Science Foundation (Grant Nos. 2019M663021); Natural Science Foundation of Guangdong Province (Grant No. 2019A1515011744); Science and Technology Planning Project of Guangdong Province (Grant No. 2017A020215060); the Medical Science and Technology Research Foundation of Guangdong Province of China (Grant No. B2020138); Shantou Technology Bureau Science Foundation of China (Grant No. (2019) 106); Grant for Key Disciplinary Project of Clinical Medicine under the Guangdong High-level University Development Program (Grant No. 002-18120302).
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.