Multimodal MRI-Based Biomarkers Of Cognitive Changes In Normal Aging And Alzheimer’s Disease

Population aging is a major problem faced by the global society.Therefore,more and more studies have focused on the mechanism of cognitive decline in normal aging and aging-related neurodegenerative diseases,such as Alzheimer’s disease(AD).Multimodal magnetic resonance imaging(MRI)can study the changes in the nervous system and function deviating from normal aging caused by neurodegenerative diseases in many aspects.For example,the brain aging prediction model using structural MRI suggests that the biological aging of individuals who later develop AD is faster than healthy older people of the same age.Therefore,using various MRI patterns to identify potential biomarkers is an essential topic for aging-related neurodegenerative diseases in the future.Using a large public data set,this dissertation studied the biomarkers of cognitive performance based on multimodal MRI in normal elderly and AD patients through indepth analysis of structural,functional,and diffusion MRI data.Chapter 2 of this work studied the changes in the brain glymphatic system activity and its association with cognitive performance and white matter hyperintensity(WMH)in normal aging.Using MRI data from 157 old adults,the data was divided into two age groups:middle-aged(4264 years)and elderly(>65 years).In this study,WMH was identified and labeled by an automatic segmentation algorithm.The diffusion tensor imaging analysis along the perivascular space(DTIALPS)index was used to quantify the function of the glymphatic system and then analyzed its correlates using analysis of variance and multiple linear regression.The results show that the older adults have more WMHs and lower ALPS-index than the middle-aged.In addition,in the middleaged group,the ALPS-index was negatively correlated with the trail-making test part B(TMT-B),while WMH was positively correlated with TMT-A and TMT-B.The existing studies on the changes in cerebellar structure and function caused by normal aging are limited by cerebellar segmentation methods.They do not analyze the functional connection(FC)between cerebellar lobules and the brain but only evaluate the changes in functional connectivity between the brain and the whole cerebellum in the elderly.Therefore,the third chapter analyzes the changes in cerebellar lobule morphology and cortical-cerebellar connectivity in the elderly and its relationship with cognitive function.This study selects the data of 264 old adults as a sample and divides them into five age groups.The results showed that the volume of cerebellar lobules decreased with age,and there were significant changes in intracerebellar FC and cortico-cerebellar FC.In addition,correlation analysis showed that FC changes between the Crus-Ⅱ lobule and the right anterior cingulate cortex(ACC)were associated with memory loss and poor performance on TMT-B in the elderly.Chapter 4 examines the potential use of longitudinal changes in regional brain atrophy and FC in the triple-network model as biomarkers of aging-related cognitive decline.This study collected the longitudinal MRI and neuropsychological data of 74 old adults,constructed the prediction model using the support vector machine(SVM)regression model,and compared the prediction abilities between local brain atrophy and FC as input through cross-validation.The results showed that after six years of follow-up,the FC of all three brain networks decreased significantly,and there were significant differences in gray matter volume and cortical thickness.The improvement of memory and the decline of executive function were also observed.Finally,when comparing the two predictive SVM models,it was found that regional brain atrophy was more effective than FC in predicting cognitive scores in the next six years.The fifth chapter of this study studied the relationship between apolipoprotein E4 genotype(APOE4),local brain atrophy,cerebral cerebellar FC,and cognitive ability in patients with Alzheimer’s disease(AD).The study used MRI and neuropsychological data from 234 old adults,including 143 APOE4 positives and 91 APOE4 negatives.Functional MRI seed-based approach was used to investigate the cerebro-cerebellar connectivity,and the amyloid PET was used to evaluate individual amyloid-β(Aβ)burden which were later used as covariates.Finally,cerebellum and cerebrum morphometries were computed using cerebellum segmentation(CERES)algorithm and Freesurfer5.3,respectively.The results showed that the FC of APOE4 carriers in the cerebral-cerebellar motor network was lower,and there were significant changes in local brain volume and cortical thickness.Finally,multiple linear regression analysis showed that AD patients with the APOE4 gene had more severe episodic memory impairment.Overall,this work uses multimodal MRI data to evaluate the mechanism of brain morphological and functional changes that may lead to cognitive impairment or delayed neurodegenerative diseases in healthy older adults.The results of this paper supplement the evidence that the changes in the brain glymphatic system activity and cerebellar structure and function play a significant role in aging and AD cognitive impairment and propose a machine learning model for predicting aging-related cognitive impairment using structural MRI biomarkers.