Multimodal MRI Based Neurovascular Uncoupling In Type 2 Diabetic Patients With Mild Cognitive Impairment

Aims:Type 2 diabetes mellitus(T2DM)is an important risk factor for mild cognitive impairment(MCI)and dementia.However,the detailed mechanism of pathology is still ambiguous,which increases difficulty of diagnosing and treating MCI,thus accelerates the deterioration of cognitive function.Previous studies have shown that neurovascular uncoupling may be the neural mechanism of brain dysfunction in diabetic patients with MCI,but specific clinical indicators are still lacking.Magnetic resonance imaging(MRI)technology can reflect the neurovascular characteristics of diabetic patients by exploring the mismatch between cerebral microvascular blood flow and neuronal activity.Our preliminary research showed that the mainly impaired cognitive dimensions of diabetic patients were executive ability and memory.Therefore,we intend to integrate psychological behavior,clinical biochemistry and other indicators,and use multi-mode MRI technology to explore specific imaging indicators of diabetic neurovascular uncoupling.This index will be helpful in screening diabetic patients with MCI,and has important theoretical and clinical significance in preventing and delaying the development of dementia.Methods:This study included 31 T2 DM patients and 30 healthy controls with no difference in age,gender,BMI and education degree.GE Discovery MR 750(3.0T)MRI scanner was used to collect Blood oxygenation level dependent(BOLD)images,arterial spin labeling(ASL)and high resolution T1 image.All subjects underwent mini-mental state examination(MMSE),Montreal cognitive assessment(Mo CA),self-rating anxiety scale(SAS),and self-rating depression scale(SDS),California verbal learning test(CVLT),Stroop color word test(STROOP).The BOLD signal is used for calculating the amplitude of low frequency fluctuation(ALFF),fractional amplitude of low frequency fluctuation(f ALFF)positive degree centrality(DCP)and Negative degree centrality(DCN)images.These maps could reflect the activity intensity of brain neurons from four different aspects.Cerebral blood flow(CBF),which can be calculated based on ASL signal,is used to reflect cerebral perfusion.Pearson correlation coefficients between the perfusion map and the 4 kinds of neuronal activity maps were calculated respectively at both levels of whole brain and brain region.Consequently,4 kinds of "neurovascular coupling indicators" were obtained,namely ALFF-CBF,f ALFF-CBF,DCP-CBF and DCN-CBF.Two sample t test was used as an inter-group comparison method to find the difference between diabetic group and healthy control group.Results:1.Through the analysis of the basic information of the subjects in the two groups,we found that there was no significant difference in basic and clinical information,except for type 2 diabetes related information(fasting blood glucose,2h postprandial blood glucose,glycosylated hemoglobin,and urinary microprotein;P = 0.004,P < 0.001,P < 0.001,P = 0.044,respectively).2.By evaluating the brain function of subjects in the two groups,we found that the scores of multiple subitems in the CVLT scale of diabetic patients were lower than those of normal controls(test 4,test 5,test 1-5,short-term free recall,short-term cue recall,and P values were 0.004,0.006,0.048,0.014,and 0.021,respectively).STROOP scale scale multiple and a lower score lower(total number,omission,right words,color consistent,correct word color consistent response time,word,color not consistent,correct word color inconsistent reaction time,the number of relevant correct pronunciation,the pronunciation related reaction time,unrelated to the correct number,pronunciation has nothing to do with the reaction time,P values were 0.001,0.006,in turn,0.037,0.026,0.011,0.006,0.026,0.002,0.001,0.002).This shows that the diabetic brain memory and executive function is reduced in type 2 diabetic patients.3.At the whole brain level,there was no significant difference in the four "neurovascular coupling biomarkers" between the two groups,but at the brain area level,we found significant differences in 12 brain areas,namely ALFF-CBF: left hippocampus,left fusiform gyrus,right fusiform gyrus and right putamen;f ALFF-CBF: right middle frontal gyrus,left middle frontal orbital gyrus,left amygdala,left superior temporal gyrus and right middle temporal gyrus;DCP-CBF: right medial cingulate gyrus and collateral cingulate gyrus;DCN-CBF: right medial frontal orbit and right plocygoid nucleus.4.In correlation analysis of brain function,clinical information and imaging information,we found that postprandial blood glucose was negatively correlated with “Trial 5” and “Trial 1-5” in CVLT scale(Spearman's rho =-0.566,P = 0.005;Spearman's rho =-0.485,P = 0.019).HBA1 c was positively correlated with “missing number” in STROOP scale and negatively correlated with “Trial 5” in CVLT scale(Spearman's rho = 0.455,P = 0.01);rs =-0.385,P = 0.033).The total duration of the disease was negatively correlated with the SDS score(Spearman's rho =-0.471,P = 0.036),which was negatively correlated with the “Incongruent Correct” in the STROOP scale(Spearman's rho =-0.53,P = 0.016).5.On the left side of the hippocampus,ALFF-CBF was positively correlated with 5 subitems of the STROOP scale,namely “Correct”(Spearman's rho = 0.473,P = 0.007),“Incongruent Correct”(Spearman's rho = 0.495,P = 0.005),“Incongruent Correct Reaction Time”(Spearman's rho = 0.440,P = 0.013),“Pronunciation Relevant Correct”(Spearman's rho = 0.461,P = 0.009),and “Pronunciation Relevant Reaction Time”(Spearman's rho = 0.427,P = 0.017).In addition,ALFF-CBF on the right fusiform region is positively correlated with SDS scores(Spearman's rho = 0.394,P = 0.028),but negatively correlated with three STROOP subterms,namely “Incongruent Correct”(Spearman's rho =-0.382,P = 0.034),“Pronunciation Relevant Correct”(Spearman's rho =-0.386,P = 0.032),and,and “Pronunciation Relevant Reaction Time”(Spearman's rho =-0.370,P = 0.041).Finally,f ALFF-CBF was positively correlated with three STROOP subterms in the left amygdala,namely,“Pronunciation Relevant Reaction Time”(Spearman's rho = 0.357,P = 0.049),“Incongruent Correct”(Spearman's rho = 0.429,P = 0.016),and “Incongruent Correct Reaction Time”(Spearman's rho = 0.395,P = 0.028)..Conclusions:Type 2 diabetic patients have reduced brain memory and executive function compared with normal controls.In images of patients with diabetes,brain regions associated with memory,movement,hearing and emotion appear "neurovascular uncoupling".In patients with diabetes,brain function is negatively correlated with the severity of the disease and positively correlated with the degree of neurovascular image coupling in most brain regions.The technique could help diagnose cognitive impairment caused by diabetes and monitor outcomes.