White Matter Microstructural Abnormalities In Type 2 Diabetes Mellitus: A Diffusional Kurtosis Imaging Analysis

Objective:To investigate WM microstructural alterations using the diffusion kurtosis imaging(DKI)method in type 2 diabetes mellitus(T2DM)patients without cognitive impairments and detect associations between DKI metrics and clinical/cognitive measurements,and further explore the neuroimaging markers for predicting the early cognitive decline in T2 DM patients without T2DM-related cognitive impairments.Subject and Method:Fifty-eight T2 DM patients without T2DM-related cognitive impairments and 58 healthy controls(HCs),whose age,gender,and years of education were well matched with the T2 DM patients,were enrolled.MRI scanning,cognitive assessments and laboratory examination were performed on all participants.The preprocessing of the DKI data was performed using Statistical Parametric Mapping software(SPM8;http://www.fil.ion.ucl.ac.uk/spm/software/spm8)and FMRIB’s diffusion toolbox(FSL 4.0,http://www.fmrib.ox.ac.uk/fsl).The preprocessing included eddy current-induced distortion and motion artifacts,skull-stripping,calculation of kurtosis and diffusion metrics,spatial normalization and smoothed with a 6 mm of full width at half maximum.To detect the intergroup differences in primary DKI metrics(FA,MD and MK),the voxel-wise general linear mode(GLM)was conducted within group(T2DM vs controls)as the main effect and age,gender and years of education as the nuisance regressors(P < 0.05,AlphaSim corrected).In order to further inspect the reason for the primarily altered DKI parameters(FA,MD and MK),ROI-based intergroup comparisons were further performed using a GLM with age,gender and years of education as the nuisance regressors to reveal the intergroup differences in the AD and RD values(the reason for the altered FA and MD values)and AK and RK values(the reason for the altered MK values).Furthermore,partial correlations between the significantly altered DKI metrics and the clinical/cognitive data and differences of the demographic,clinical,cognitive data between the two groups were analyzed by using SPSS 19.0.The significant level was set as P < 0.05.Results:T2DM patients had increased fasting blood glucose(P < 0.001)and HbA1c(P < 0.001)levels and reduced high-density lipoprotein levels(P < 0.05)compared with the HCs.No significant intergroup differences were observed in terms of age,gender,years of education and cognitive performance(P > 0.05).In the voxel-based analysis,T2 DM patients exhibited significantly reduced FA values in the right prefrontal WM(rPF_WM)and the splenium of corpus callosum(CC),significantly increased MD values in the bilateral prefrontal WM,the right superior temporal WM(rST_WM),the left external capsule(lEC),the splenium of the CC and the pons,significantly decreased MK values in the splenium of the CC and the pons compared with HCs(P < 0.05,AlphaSim correction).ROI-based intergroup comparisons showed that the RD values within the rPF_WM significantly increased in T2 DM patients compared with controls(P < 0.05),which leads to the decreased FA in T2 DM patients.However,there was no significant intergroup difference in the AD values within the rPF_WM(P > 0.05).The AD and RD values within the splenium of the CC significantly increased in T2 DM patients compared with controls(P < 0.05),which leads to the decreased FA in T2 DM patients.The intergroup differences in MD values were driven by increased AD and RD within corresponding ROIs in T2 DM patients(P < 0.05).Besides,the AK and RK values within the splenium of the CC and the pons were significantly decreased in T2 DM patients compared with controls(P < 0.05),which drive the decreased MK within corresponding ROIs in T2 DM patientsAdditionally,by using the partial correlation analyses,we found a negative correlation between the decreased FA values of the rPF_WM and the reaction time of the attention network test(RT of ANT)(r =-0.280,P = 0.047)and a positive correlation between the RD values of rPF_WM and RT of ANT(r = 0.322,P = 0.021).in T2 DM patients.Moreover,positive correlations were demonstrated between the disease duration and MD values(r = 0.273,P = 0.044),AD values(r = 0.269,P = 0.047)and RD values(r = 0.270,P = 0.046)of the superior rPF_WM in T2 DM patients.Conclusion:With both conventional diffusion and additional kurtosis metrics,DKI can provide additional information of WM microstructural abnormalities in T2 DM patients.And this is the first study using DKI to analyze WM microstructural alterations in T2 DM patients and further detect associations between DKI metrics and clinical/cognitive measurements.Our results suggested that,in T2 DM patients,WM microstructural abnormalities occur before cognitive decline and may be used as neuroimaging marker for predicting the early cognitive impairments.