【作者】 李亚迪；

【导师】 冯晓源； 何慧瑾；

【作者基本信息】 复旦大学 ， 影像医学与核医学， 2009， 博士

【摘要】 第一部分aMCI与轻度AD基于体素的全脑形态学MRI研究目的：利用磁共振3D T1WI成像研究遗忘型轻度认知障碍(amnestic-type mild cognitive impairement，aMCI)、轻度阿尔茨海默病(Alzheimer’s disease，AD)患者相对于正常老年人灰质体积改变的特点。方法：采用3.0T磁共振，对33例aMCI患者、32例轻度AD患者及31例正常老年人进行三维T1WI扫描，利用基于SPM5的DARTEL工具箱对扫描获得的结构图像进行预处理，再对aMCI组、轻度AD组和对照组的全脑灰质体积进行基于体素的统计学比较。结果：与正常老年组比较，aMCI组左侧海马、海马旁回、舌回、颞上回，双侧岛叶、颞中回等结构的灰质体积萎缩，其差异具有统计学意义(P＜0.01，FDR corrected，K≥50体素)。轻度AD组的双侧海马、海马旁回及杏仁核、双侧丘脑、双侧颞顶叶皮质等结构灰质体积萎缩，额叶与枕叶皮质也出现灰质萎缩，其差异具有统计学意义(P＜0.05，FDR corrected，K≥50体素)。结论：基于体素的形态学研究能够发现早期AD患者中大脑灰质广泛的及细微结构的萎缩，从而能够更加早期、全面，客观地反映aMCI、轻度AD患者的脑结构改变。第二部分aMCI与轻度AD基于体素的全脑DTI研究目的：应用磁共振弥散张量成像(DTI)技术，探讨aMCI、轻度AD患者的脑白质异常的变化特点。方法：采用3.0T磁共振，对33例aMCI患者、32例轻度AD患者及31例正常老年人进行DTI检查。运用SPM5分析软件，采用基于体素的分析方法，比较全脑FA值的差异及变化。结果：与正常老年组比较，aMCI组的双侧额颞叶，左侧枕下回旁白质、扣带前部、舌回旁白质、梭状回旁白质、顶下小叶，右侧脑室三角区外上方白质的FA值减低，其差异具有统计学意义(P＜0.001，uncorrected，K≥20体素)。轻度AD组双侧额颞枕叶、海马旁白质、扣带前部、胼胝体、侧脑室三角区旁旁白质、顶下小叶，右侧楔前叶与楔叶、距状裂邻近白质、左侧颞干FA值减低，其差异具有统计学意义(P＜0.001，uncorrected，K≥20体素)。与本课题第一部分出现灰质萎缩的部位进行比较，我们发现aMCI患者组脑白质FA值减低的脑区与灰质萎缩脑区大部分不一致，仅在左侧颞中叶有部位相近的差异区域；轻度AD患者组脑白质FA值减低的脑区与灰质萎缩脑区部分一致。结论：基于体素的全脑FA图分析，能够比较客观、全面地发现aMCI、轻度AD患者脑白质异常的区域。脑白质改变的模式与灰质萎缩的模式不尽相同，可能是多种因素导致的结果。第三部分aMCI与轻度AD的静息态fMRI研究目的：应用成组独立成分分析技术(GroupICA)，探讨aMCI与轻度AD患者静息态脑功能的变化特点。方法：采用3.0T磁共振，对33例aMCI、32例轻度AD患者和31例正常老年人进行静息态fMRI扫描。运用REST、GIFT、SPM5分析软件，进行基于体素的全脑分析，比较静息态功能网络的变化。结果：种子相关分析及成组独立成分分析技术发现的正常老年组的默认模式网络构成脑区类似，包括：后扣带回(posterior cingulate cortex，PCC)、楔叶及楔前叶、顶下小叶、前额叶内侧区(Dorsal-medial prefrontal cortex，DMPFC)、前额叶背外侧区(Dorsal-lateral prefrontal cortex，DLPFC)、部分颞枕叶(P＜0.05，FWER，K≥10体素)，此外成组独立成分分析结果还包括双侧丘脑、海马、海马旁回、岛叶。aMCI患者默认模式网络中大部分脑区都出现了激活程度的减低，而在双侧楔前叶、楔叶、扣带回中部、壳核、纹状体、右侧顶下小叶出现激活程度的增加(P＜0.05，FWER，K≥10体素)；aMCI患者组视觉网络中BA(18、19)区在也出现了激活程度的增加(P＜0.001，uncorrected，K≥10 voxels)；轻度AD患者两个偏侧化的网络中右侧额中回(BA9)、右侧顶下小叶(BA40)、左侧顶下小叶(BA40)与顶上小叶(BA7)、左侧额叶(BA 6、9、8)的激活减弱及右侧额中回(BA 10、46)激活增加(P＜0.05，FWER，K≥10体素)。默认模式网络激活程度减低的脑区中除了左侧海马等少许脑区，其余的静息态网络脑区功能的异常在形态学异常之前出现；两个偏侧化的网络中激活减弱的脑区未发现明显的灰质萎缩。DTI发现的结构连接的破坏为功能连接的异常提供了解剖学上的证据。结论：通过成组ICA可以在aMCI、轻度AD患者脑结构出现异常之前发现静息态脑功能网络特征性得改变。联合MRI结构与功能分析能够提供更多的信息，是将来研究的方向。更多还原

【Abstract】 PARTⅠVoxel-Based MRI Analysis of Whole Brain Gray Matter in Patients withaMCI and mild Alzheimer’s DiseaseObjective To study the pattern of volume changes of the whole braingray matter in patients with amnestic mild cognitive impairment(aMCI) and mildAlzheimer’s disease (AD) by voxel-based morphometry (VBM). Material andMethods 33 patients with aMCI, 32 patients with mild AD and 31 normal agingvolunteers as control subjects were enrolled in the study. Gray matter volumedifferences of the whole brain were assessed using SPM5-based DARTEL toolbox tomake voxel-based morphometric comparison between patients and the control group.Results The volume of the left hippocampus,parahippocampal gyrus,lingualgyrus,superior temporal gyrus, bilateral insulae, middle temporal gyri,etc in the aMCIpatient group was significantly smaller than that in the control group (P＜0.05, FDRcorrected,K≥50 voxels). The volume of the bilateral hippocampi, parahippocampalgyri, amygdalae,thalami,temporo-occipital gyri,inferior parietal lobules,superiorparietal lobules,precuneuses,middle frontal gyri,and left cuneus,fusiform gyrus, rightposterior cngulate gyrus,inferior frontal gyrus,etc in the mild AD patients group wassignificantly smaller than that in the control group(P＜0.05,FDR corrected P＜0.05,FDR corrected,K≥50 voxels). Conclusion VBM can reveal widespreadvolumetric reduction of gray matter’s fine structure in early-stage Alzheimer’s diseasewith the advantage of objectivity. PARTⅡWhite Matter Abnormalities in Patients with aMCI and mildAlzheimer’s Disease investigated with Voxel-Vased Diffusion Tensor ImagingObjective To detect white matter abnormalities in patients with amnesticmild cognitive impairment(aMCI) and mild Alzheimer’s disease (AD) by voxel-basedDTI analysis. Material and Methods 33 patients with aMCI,32 patients with mild ADand 31 normal aging volunteers as control subjects were enrolled in the study.Fractional anisotropy (FA) maps were preprocessed using SPM5 to make voxel-wisecomparison of anisotropy in whole brain between patients and the control group.Results Significant reductions in FA values were found in the white matter of inferioroccipital gyrus,anterior part of cingulum,lingual gyrus,fusiform gyrus,inferior parietallobule in the left side and the white matter adjacent to the triangular part of the rightlateral ventricle and the white matter in bilateral frontal and temporal lobes in patientswith aMCI(P＜0.001,uncorrected, K≥20 voxels). In patients with mild AD, significantreductions in FA values were found in the corpus callosum,bilateral anterior part ofcingulum and the white matter of bilateral frontal lobes,temporal lobes,occipitallobes,inferior parietal lobules,right precuneus and cuneus and the white matter adjacentto bilateral hippocampi, the triangular part of the bilateral lateral ventricle and the rightcalcarine sulcus(P＜0.001,uncorrected, K≥20 voxels). Most of the areas with reducedFA values were not consistent with those with reduced gray matter volume in patientswith aMCI.While some of the areas with reduced FA values were in concordance withthose with reduced gray matter volume in patients with mild AD.ConclusionVoxel-wise comparison of anisotropy in whole brain between patients and the controlgroup can reveal widespread reduction of white matter anisotropy in aMCI and mildAlzheimer’s disease with the advantage of objectivity.The fact that the pattern ofanisotropic changes of white matter was not in concordance with that of volumetricreduction of gray matter reveals that white matter anisotropic changes may be outcomefrom a couple of factors rather than wallerian degeneration solely. PARTⅢResting-State Functional MRI Analysis in aMCI and mild Alzheimer’sDiseaseObjective To investigate the brain activity in resting-state of patientswith aMCI and mild Alzheimer’s disease by GroupICA. Material and Methods 33patients with aMCI,32 patients with mild AD and 31 normal aging volunteers ascontrol subjects were scanned with 3.0T MRI.fMRI data were preprocessed with suchsoftwares as SPM5,REST,GIFT and then compared between patients and the controlgroup. Results GroupICA detected 8 resting-state networks which act for differentfunctions.The spatial map of Default mode network(DMN) of aging volunteersrevealed by ROI-based correlation analysis was similar with that found by GroupICAwhich includes bilateral PCC,precuneus,cuneus,inferior parietal lobules,dorsomedialprefrontal cortices(DMPFC),dorsolateral prefrontal cortices(DLPFC),part of temporaland occipital Iobes(P＜0.05, FWER, K≥10 voxels).In addition, the DMN revealed byGroupICA also includes bilateral thalami,hippocampi,parahippocampi,insulae,whichindicate that ICA can reveal more areas than ROI-based correlationanalysis(P＜0.05,FWER, K≥10 voxels). GroupICA found that most areas of DMN haddecreased BOLD signal while increased BOLD signal were found in such areas asbilateral precuneus,cuneus,the middle part of cingulate gyri, putamina, striatumes andfight inferior parietal lobule(P＜0.05,FWER, K≥10 voxels).Increased activity werealso revealed in visual network(BA 18, 19) in patients with aMCI (P＜0.001,uncorrected,K≥10 voxels).These areas may provide compensation for the malfunction of memorysystem.GroupICA found areas with decreased activity in 2 lateralized networks withincreased activity in fight middle frontal gyrus (BA10,46) in patients with mildAD(P＜0.05,FWER, K≥10voxels).Both of the 2 lateralized networks play an importantrole in memory and attention function.Among all the areas of DMN found byGroupICA,only few areas such as left hippocampus presented gray matter atrophy aswell as decreased resting-state activity in the aMCI group.All the other areasmanifested decreased resting-state activity before the gray matter atrophy waspresented.The result of partⅡprovided the evidence that the damage of structuralconnectivity might be the physical foundation in the impairment of functionalconnectivity.Conclusion GroupICA can detect the abnormalities of resting-state brain activity before the structural altemation can be demonstrated.The combination ofstructural and functional analysis of the brain can reveal much more information thaneach of them solely, which is an important direction for future research.更多还原