Altered Brain Structural Topological Properties In Type 2 Diabetes Mellitus Patients: A Diffusion Tensor Imaging Analysis

Objective:Type 2 diabetes mellitus(T2DM)is one of the risk factors of cognitive dysfunction.The microstructural impairments of white matter(WM)plays a critical role in T2DM-related cognitive decline,and WM destructions have been detected in patients with T2 DM even before clinical diagnosis of cognitive dysfunction.We investigated the changes of brain structural topological properties and their correlation with behavior in T2 DM patients without complications.Materials and Methods:In this study,fifty-seven T2 DM patients(mean age: 55.98 ± 8.19 years)and 57 healthy controls(HCs,mean age: 54.46 ± 6.93 years),whose gender,age and level of education were well matched with the T2 DM patients,were recruited.Magnetic resonance imaging data was obtained from each participant using a 3.0T MR system(Discovery MR750;General Electric,Milwaukee,WI,USA),and a battery of cognitive assessments and laboratory examination were performed.The DTI data were preprocessed using the FSL software,and the steps included eddy current and motion artifact correction,brain extraction,diffusion index calculation and realignment.Diffusion tensor tractography(DTT)was performed using a PANDA toolbox based on FMRIB Software Library v5.0.In the process of DTT,fiber assignment by continuous tracking(FACT)algorithm was utilized,and the threshold of fractional anisotropy(FA)and turning angle was set at 0.2 and 45°.Using GRETNA(https://www.nitrc.org/projects/gretna)based on Matlab to construct brain WM topology network.The Anatomical Automatic Labeling(AAL)atlas was used to parcellate each brain into 90 regions defining as 90 nodes,and interconnections between brain regions were taken as the edges,which was used to construct the WM structural network and to evaluate the topological properties(both global and nodal graph parameters)of structural networks.SPSS 21.0 statistical software was used for statistical analysis,and to calculate the statistical significance in the clinical,demographic,cognitive data between the two groups.We applied general linear model to determine the significance of intergroup differences in global and nodal parameters after controlling for sex,age,and level of education.Significance level was set at two-tailed P < 0.05 for global parameters.To correct the multiple comparisons for nodal parameters,P < 0.011(1/90)was considered significant.Moreover,for the network parameters that showed significant between-group differences,we computed partial correlations between these parameters and clinical/cognitive variables for the T2 DM patients controlling for age,gender,and education levels(P < 0.05).Results:There were no significant differences in age,sex,education,body mass index(BMI),blood pressure,triglyceride(TG)and total cholesterol(TC)between the two groups(P > 0.05).As expected,fasting plasma glucose(FPG)and glycated hemoglobin(HbA1c)levels in T2 DM patients were significantly higher than those in the HCs group(P < 0.001).There were no significant differences in cognitive and neuropsychological tests between the two groups(P > 0.05).Both T2 DM patients and HCs presented small-world organization(? > 1)in WM structural networks.Normalized clustering coefficients(?)and characteristic path length(?)were increased in T2 DM patients(P < 0.05),whereas there were no significant differences in global efficiency(Eglob)and local efficiency(Eloc)(P > 0.05).Disrupted nodal parameters were mainly present in the frontal,parietal,temporal,and basal ganglia regions in T2 DM patients.The right hippocampus,right amygdala and the left pallidum exhibited decreased global efficiency;the left post central gyrus and the superior pole of the right temporal lobe(SPT.R)had decreased local efficiency,and the right inferior frontal gyrus(IFG.R)had increased nodal degree.These results were relatively stable across all five fiber number(FN)thresholds.In T2 DM patients,? was negatively correlated with the accuracy rate(ACC)of the spatial working memory(SWM)task(r =?0.299,P = 0.044),whereas ? was positively correlated with the response time(RT)of the attention network test(ANT)(r = 0.334,P = 0.023).In the control group,no significant correlations were found between global network parameters and behavioral performance.In T2 DM patients,nodal global efficiency in the right hippocampus was positively correlated with the score of backward digital span(r = 0.300,P = 0.043),and nodal local efficiency in the SPT.R was positively correlated with the ACC of the ANT(r = 0.315,P = 0.033),as well as the score of backward digital span and forward digital span(r = 0.346,P = 0.018;r = 0.333,P = 0.024).In the control group,there were no significant correlations between nodal parameters and behavioral performance.In T2 DM patients,there were no significant correlations between changes in network parameters and either disease duration or blood glucose levels.Conclusion:This study analyzed the changes of WM network in T2 DM patients without complications,and further analyzed the relationship between the changes of WM network properties related to T2 DM and clinical indicators and behavioral performance.The results verified the existence of WM structural network changes and the association between structural properties and cognitive state in T2 DM patients before the occurrence of complications.Research of structural properties may contribute to our understanding of the intrinsic links between T2 DM and cognition.