Neuroimaging Research For Recognition Of Alzheimer’s Disease Spectrum Based On Anomalously Altered Cerebral Blood Flow And Brain Network

Background: Alzheimer’ s disease(AD)is a slowly progressing neurodegenerative disease characterized by progressive deterioration of cognitive function and the capacity to live independently,which constitutes 60%-70% of all dementia cases.Extracellular amyloid-β(Aβ)aggregation and intracellular neurofibrillary tangles are the pathological features of AD.Nowdays,AD is considered as a continuous disease process that includes subjective cognitive decline(SCD),mild cognitive impairment(MCI),and AD dementia.The precise mechanism underlying AD remains unclear,however,the early detection and intervention of the AD spectrum(ADS)can effectively prevent cognitive decline and delay the disease progression! Therefore,looking for objective markers for ultra-early diagnosis and early identification of AD high-risk populations,especially for the evolutionary characteristics of transformation to dementia are issues that need to be resolved urgently.The relationship between vascular factors and the onset of AD has attracted an increasing amount of attention in recent years.Multiple research have confirmed that abnormal function of vascular appears in the early stage of AD,and the decreased cerebral blood flow(CBF)may be the earliest biomarker for AD.With the rapid development of multimodal magnetic resonance imaging(MRI)technology,the arterial spin labeling(ASL)has been used for non-invasive quantitative detection of CBF,which found that there are significant regional cerebral blood flow changes in patients with MCI and AD.In addition,resting-state functional MRI has also shows that abnormal brain activity and brain network are not only widely exist in ADS,but also participate in and regulate the occurrence of AD.Importantly,the static brain network is not stable,but shows a dynamic and fluctuating pattern.Compared with static brain networks,dynamic functional connectivity(d FC)can portray more refined brain network changes and provide a new way to quantify the time-varying characteristics of brain activity.Therefore,the study on combination of cerebral blood flow and dynamic and static brain network may provides new ideas for revealing the pathogenesis of AD,and find out reliable imaging biomarkers for early recognition of high-risk populations of AD.Chapter one: Altered Regional Cerebral Blood Flow,Brain Function and individualizes identification across the Alzheimer’s disease spectrumObjective: To investigate the variation in the characteristics of regional Cerebral Blood Flow(r CBF),brain activity and intrinsic Function Connectivity(FC)across the Alzheimer’s Disease Spectrum(ADS),and to integrate abnormal functional network changes to build ADS prediction model.Methods: 20 individuals were recruited in each of the following categories: Alzheimer’s Disease(AD),Mild Cognitive Impairment(MCI),Subjective Cognitive Decline(SCD)and cognitive normal(CN).All the participants completed the 3.0T resting-state functional magnetic resonance imaging and arterial spin labelling scans in addition to neuropsychological tests.Additionally,the normalized CBF and Regional Homogeneity(Re Ho)of individual subjects were compared in the ADS.Moreover,changes in intrinsic FC were investigated across the ADS using the abnormal r CBF regions as seeds and behavioral correlations.Finally,a support vector classifier model of machine learning was used to distinguish ADS from CN.Results: Compared to the CN subjects,AD patients showed the poorest level of r CBF in the Left Precuneus(LPCUN)and Right Middle Frontal Gyrus(RMFG)among all the participants.In addition,there was a significant decrease in Reho in the right PCC.Moreover,RMFG-and LPCUN-based FC analysis revealed that the altered FCs were primarily located in the posterior brain regions.Finally,a combination of altered FC showed a better ability to differentiate ADS from HC,AD from SCD and MCI,and MCI from SCD.Conclusions: The decreased CBF destroys the integrity of brain network and aggravates the damage of cognitive function,which reflects the disorder of neurovascular function regulation in ADS.The altered FC can effectively distinguish the ADS,which plays an important role in deeply understanding the pathogenic mechanism of AD and constructing individual diagnosis and treatment model.Chapter two: Abnormal dynamic functional connection of the cerebral perfusion brain areas associated with cognition and individualizes identification in the Alzheimer’s disease spectrumObjective: Using the sliding window technique and K-Means clustering to explore the characteristics of d FC in cognitive-related cerebral perfusion areas in ADS and to verify the power of the combination of CBF and d FC on the classification of ADS.Methods: First,33 CN controls,31 MCI,30 SCD and 21 AD were evaluated by the 3.0T resting-state functional MRI and ASL scans and neuropsychological tests.Secondly,to obtain the brain perfusion areas related to cognitive function,the partial correlation was used between ASL images and MMSE scores of all subjects in ADS group.Third,using the sliding window technique,the previously obtained brain regions were used to construct d FC brain network.At the same time,K-Means clustering was used to identify the functional connectivity state(module)of repeated expression,and the state characteristics of d FC were compared between groups.In addition,the statistical method of network-based statistics(NBS)was used to compare the clustering of different states of FC between groups(P < 0.0001),and the temporal variability of FC between groups was compared.Finally,the Fisher linear stepwise discriminant analysis(LDA)was used to construct the ADS classification model for the CBF and d FC temporal variability,and the left-one cross-validation method was used to verify the robustness of the model.Results: 1)In ADS,CBF was significantly correlated with MMSE in 19 brain regions,in which the positively correlated regions were located in the left hemisphere,and the higher the CBF in the right hemisphere,the worse the cognitive performance.2)The d FC brain network was clustered into two functional connectivity states of repeated expression,and the number of state transitions decreased at first and then increased with the progression of the disease.3)Compared with CN group,the FC intensity of ADS group decreased generally in state 1,while FC increased in most brain regions in state 2.It was found that the time variability of d FC in SCD group and AD group increased and was closely related to cognitive score.4)The combination of CBF and d FC time variability showed a better efficiency of classifying ADS population,the area under the curve(AUC)was more than 80%,and the accuracy of cross-validation is more than 70%,which was better than using CBF or d FC alone.Conclusions: First,the decline of CBF in ADS is closely related to the deterioration of cognitive function.Secondly,the disorder of vascular regulation leads to the fragility and instability of dynamic functional connectivity,which aggravates the damage of cognitive function.Finally,the combination of CBF and d FC time variability shows good classification and robustness to ADS,which is a promising diagnostic marker.