An FMRI Study On Working Memory-related Brain Networks In Type 2 Diabetics

Objective:Aimed to detect alteration patterns of activation and functional connectivity within working memory(WM) related networks in type 2 diabetes mellitus(T2DM) patients without T2DM-related complications, investigate the relationship between the alteration and behavioral performance, and explore the imaging markers for predicting early cognitive decline in T2 DM patients without T2DM-related complications.Subject and Method:Twenty-one T2 DM patients without T2DM-related complication(14 males, 7 females, average age 54.52 ± 8.78 years) and 21 healthy controls(HC)(9 males, 12 females, average age 51.14 ± 6.09 years) were enrolled. Functional MRI data were acquired during performing a block-designed 0-back and 1-back task by using a 3.0-Tesla MR scanner, and task performance(accuracy and response time) were collected.All the functional images were preprocessed using Statistical Parametric Mapping software(SPM8; http://www.fil.ion.ucl.ac.uk/spm/software/spm8). The preprocessing included slice timing, realignment, spatial normalization to the MNI, and smoothed with a 6mm full width at half maximum. Group spatial ICA was applied on smoothed fMRI data using the GIFT software based on Matlab, the independent component(IC) time-courses and spatial maps of 22 ICs for each participant were acquired, and the spatial maps were centered to standard functional connectivity degree(z score). After discarding 9 ICs of noise or of no association to the aim of this study, general linear model(GLM) in SPM8 was used to regress the time courses of the remaining 13 ICs against the design matrix for the WM task, and 13 association coefficients(β-values) of each participant representing activity levels were obtained. The averaged β-values of 13 ICs were further tested by using one-sample t-tests(P < 0.05, Bonfferoni correction) to identify seven WM task-related components.The differences in β-values of seven task-related components between T2 DM and HC groups were tested by using general linear model controlling the age, gender and education(P < 0.05, Bonfferoni corrected). The inter-group differences in regional functional connectivity within seven networks were examined by using two-sample t-tests embedded in SPM8 after controlling the age, gender and education(P < 0.05, AlphaSim corrected). Differences of the demographic, clinical, cognitive data and task performance variables between the two groups were analyzed by using SPSS 21.0(SPSS, Inc, Chiago.IL). Significance level was set at P < 0.05.Results:Compared to HC group, T2 DM group has higher education level(P = 0.03) and lower accuracy rate in the WM task(P = 0.01). No significant difference was found in clinical laboratory examinations, cognitive tests or task performance between T2 DM and HC groups.component 11(C11) and component 19(C19) showed significantly higher β-values in T2 DM group than in HC group(C11, P = 0.002; C19, P = 0.003, Bonferroni corrected) after controlling for the age, gender and education. In addition, the β-values of C11 and C19 negatively correlated to the response time of WM task in T2 DM group(P < 0.05). Voxel-wise comparison of functional connectivity showed lower regional functional connectivity in T2 DM group than in HC group in three regions located in bilateral ventral lateral prefrontal cortex(vlPFC) in the C19(P < 0.05, AlphaSim corrected) after controlling for the age, gender and education. There were no significant correlation between the decreased functional connectivity of three regions and the WM task performance in T2 DM group.Conclusion:This is the first study to analyze both brain activation and functional connectivity within the WM task-related networks extracted with ICA in T2 DM patients. Our results suggested that WM task-related activation and functional connectivity changes exist even in the early stage of T2 DM patients. the activation and functional connectivity alterations within WM task-related networks consisted of frontal and parietal regions may reflect an inefficient compensational mechanism at the early stage of T2 DM patients without complication, and the diminished functional connectivity within WM task-related networks may contribute to the mechanism of the inefficient compensation. In addition, this study demonstrated that fMRI may provide potential imaging-based markers of slightly cognitive decline in T2 DM patients.