| ObjectiveBased on previous studies on the resting-state functional magnetic resonance imaging meta-analysis of CNLBP,we constructed a low back pain-related brain function network with31 predefined brain regions,and compared the topological properties of this functional network between CNLBP patients and healthy subjects by graph-theoretic analysis,and analyzed its correlation with low back pain symptoms as well as neuropsychological performance,aiming to further explore the central pathological mechanism of CNLBP and provide new ideas for its clinical diagnosis and treatment.MethodsPotential subjects were recruited in the surrounding community of the Ping Shan campus of Fujian University of Traditional Chinese Medicine.31 eligible patients with CNLBP and31 age-and sex-matched healthy subjects were screened for enrollment according to the diagnostic criteria,inclusion criteria and exclusion criteria.Subjects’ pain symptoms were evaluated by short form Mc Gill pain questionnaire(SF-MPQ),pain sensitivity questionnaire(PSQ),and oswestry dysfunction index(ODI),and their neuropsychological symptoms were evaluated via pain catastrophizing scale(PCS),fear-avoidance beliefs questionnaire(FABQ),beck depression inventory-Ⅱ(BDI-II),and beck anxiety inventory(BAI).Structural brain images and resting-state functional images were collected from all subjects at the Imaging Department,Rehabilitation Hospital,Fujian University of Traditional Chinese Medicine.The GRETNA2.0.0A toolbox was used to perform operations such as pre-processing of resting-state functional image data,functional network construction,and topological properties analysis.Based on previous studies on the resting-state functional magnetic resonance imaging meta-analysis of CNLBP,31 brain regions closely related to low back pain symptoms with non-overlapping spatial coordinates were extracted as nodes,and the functional connections between brain regions were used as edges to establish a low back pain-related brain function network.Using the selection method based on the sparsity threshold,the sparsity range was set from 0.10 to 0.50 with a step size of 0.01 and compared with 100 pre-set random networks to finally obtain the specific topological attribute parameter values and the corresponding area under the curve(AUC)of the low back pain-related brain function network at 41 sparsity levels.Topological properties are divided into global properties and nodal properties.Global properties include small-worldness,normalized clustering coefficient,normalized characteristic path length,clustering coefficient,characteristic path length,global efficiency,local efficiency and hub disruption index.Nodal properties include hub,degree centrality,betweenness centrality,nodal efficiency,and nodal clustering coefficient.The demographic and clinical data of the two groups were analyzed by SPSS 24.0 statistical software,and the independent-samples t-test was used to compare the demographic characteristics and the clinical data of the two groups for continuous variables following a normal distribution,and the Mann-Whitney U test was used for the nonparametric continuous variables.Group differences in global and nodal properties(AUC values for network topologic parameters)were examined by a two-sample t-test with Bonferroni corrections for multiple comparisons using GRETNA2.0.0A software.Pearson or Spearman was applied to correlate statistically different topological attribute indices with pain symptoms and neuropsychological symptoms in patients with CNLBP,and P<0.05 was considered statistically significant.Results1.Demographic and clinical data: The differences in age,gender,and body mass index were not statistically significant between the two groups(P>0.05).CNLBP patients had a statistically significantly higher PCS,BDI-II,and BAI scores compared with healthy subjects(P<0.05).2.Global topological properties: The low back pain-related brain function network in two groups were all characteristic of small-world properties.Compared to healthy subjects,CNLBP patients had significantly lower AUC values of σ,γ,Eglob and significantly higher AUC value of Lp within the threshold range(P<0.05),whereas the AUC values of λ,Cp,and Eloc were not statistically different between the two groups(P>0.05).The analysis of HDI showed that the κ values for Bc of CNLBP patients were negative and significantly lower than those of healthy subjects(P<0.05).3.Nodal topological properties: The results showed that the distribution of hubs in CNLBP patients and healthy subjects were the same,include the right amygdala,left Putamen and the midcingulate cortex.CNLBP patients had significantly lower AUC value of Bc in the left primary somatosensory cortex,right anterior insular cortex,and right dorsolateral prefrontal cortex,significantly lower AUC value of Ne in left primary somatosensory cortex,right posterior insular cortex,and right dorsolateral prefrontal cortex,significantly higher AUC value of Ncp in the left thalamus and right dorsolateral prefrontal cortex and a significant higher AUC value of Dc in the left amygdala(Bonferroni correction,P<0.05).4.Correlation analysis: With no significant differences between groups in gender,age,and body mass index,correlation analysis showed that the scores of ODI(r=0.558,P=0.001)and SF-MPQ(r=0.662,P<0.001)in CNLBP patients were positively correlated with BAI scores;AUC value of Eglob were positively correlated with BDI-II scores(r=0.391,P=0.033);AUC value of Ncp in left thalamus were negatively correlated with BAI(r=-0.394,P=0.031)and BDI-II scores(r=-0.509,P=0.004);AUC value of Ne in right posterior insular cortex were negatively correlated with the duration of low back pain(r=-0.386,P=0.035)and PSQ scores(r=-0.426,P=0.019).No significant correlation was seen between the remaining indicators and clinical data.ConclusionIn this study,the topological properties of low back pain-related brain function network were compared between CNLBP patients and healthy subjects by graphical analysis based on resting-state functional magnetic resonance.The results showed impaired small-world properties and reduced hub disruption index for betweeness centrality in CNLBP patients,suggesting that the information transfer between brain regions at the global level was reduced and the brain functional network was reorganized.In addition,patients with CNLBP showed significant increases or decreases in the topological properties of nodes in the thalamus,dorsolateral prefrontal cortex,amygdala,primary somatosensory cortex,and insular cortex,suggesting the damaging and compensatory changes in the corresponding brain regions.The results provided effective complements to the neural mechanisms for the complex multidimensional pain symptoms of CNLBP patients. |
PDF Full Text Request