阿尔茨海默病患者大脑形态学及结构协变网络的改变

Altered brain morphometry and structural covariant networks based on cortical thickness in Alzheimer's disease

目的探讨阿尔茨海默病(Alzheimer's disease,AD)患者大脑灰质体积、灰质皮层厚度及基于皮层厚度的结构协变网络(structural covariance network,SCN)的拓扑属性改变。材料与方法本研究共筛选了250例来自ADNI数据库的被试,包括AD组100人,健康对照(healthy controls,HCs)组150人。首先,利用基于体素的形态学分析方法(voxel-based morphometry,VBM)和基于表面的形态学分析方法(surface-based morphometry,SBM)分别计算每组被试的灰质体积和皮层厚度并比较其组间差异。其次,将有组间差异的脑区定义为感兴趣区(region of interest,ROI),提取每一个ROI的灰质体积和皮层厚度值,与认知量表进行偏相关分析。最后,构建基于皮层厚度的SCN并利用图论分析方法分析该网络的全局属性及局部属性的变化特征。结果第一,相较于HCs组,AD组的灰质体积和皮层厚度显著下降[体素和顶点水平总体误差(family-wise error,FWE)校正后P<0.001]。AD组灰质体积下降的脑区主要包括双侧海马、双侧眶额皮层、左侧岛叶、右侧枕下回、左侧楔前叶、左侧中央前回、左侧中央扣带回。AD组皮层厚度变薄的脑区主要包括双侧颞叶、双侧额叶、双侧顶叶、双侧扣带回、双侧梭状回、双侧岛回、双侧楔前叶等。第二,偏相关分析表明,AD组简易精神状态检查量表(Mini-Mental State Examination,MMSE)得分分别与右侧海马体积[rs=0.35,错误发现率(false discovery rate,FDR)校正后P<0.001]、左侧海马体积(r_(s)=0.38,FDR校正后P<0.001)、右侧梭状回皮层厚度(r_(s)=0.38,FDR校正后P<0.001)呈正相关;临床痴呆评定量表(Clinical Dementia Rating Sum of Boxes,CDR-SB)评分与左侧梭状回皮层厚度(r_(s)=-0.39,FDR校正后P<0.001)呈负相关。第三,脑网络分析表明,AD组SCN的全局效率(P<0.001)、局部效率(P=0.03)及小世界属性(P<0.001)高于HCs组,最短路径低于HCs组(P<0.001)。结论联合VBM、SBM的形态学分析及SCN的图论分析有助于全面理解AD患者脑网络的重组及其意义,进而为AD患者神经影像学改变提供新的见解和证据。

Objective:To investigate the alteration of cerebral grey matter volume and cortical thickness and structural covariance network(SCN)based on cortical thickness in patients with Alzheimer's disease(AD).Materials and Methods:In this study,a total of 100 patients with AD and 150 healthy controls(HCs)were included.Firstly,we conducted voxel-based morphometry(VBM)and surface-based morphometry(SBM)analysis in Computational Anatomy Toolbox 12(CAT12)to acquire grey matter volume and cortical thickness.Subsequently,partial correlation analysis was applied to explore the correlation between brain regions with statistical differences and cognitive scales.Lastly,we constructed the SCN based on cortical thickness and analyzed its alternation of topology properties by graph theory analysis.Results:Firstly,we observed the decreased grey matter volume and cortical thickness in patients with AD[P-values after family-wise error(FWE)correction,PFWE-corr<0.001].The volumetrically decreased brain regions included bilateral hippocampus,bilateral orbitofrontal cortex,left insula,right inferior occipital gyrus,left precuneus,left precentral gyrus,left middle cingulate gyrus.The cerebral regions with thinner cortical thickness in AD group included bilateral temporal lobe,frontal lobe,parietal lobe,cingulate gyrus,fusiform gyrus,insula,precuneus,et al.Secondly,partial correlation analysis in AD group showed that Mini-Mental State Examination(MMSE)scores were respectively positively correlated to the volumes of right hippocampus[rs=0.35,P-values after false discovery rate(FDR)correction,PFDR-corr<0.001],left hippocampus(r_(s)=0.38,PFDR-corr<0.001),the thickness of right fusiform gyrus(r_(s)=0.38,PFDR-corr<0.001),and the clinical dementia rating sum of boxes(CDR-SB)scores was negatively correlated to the thickness of left fusiform gyrus(r_(s)=-0.39,PFDR-corr<0.001).Lastly,in SCN analysis,we found the global efficiency(P<0.001),local efficiency(P=0.03),sigma(P<0.001)were higher in AD patients compared to HCs,while the shortest path length(P<0.001)was lower in AD patients.Conclusions:The combination of morphological analysis by VBM and SBM and SCN analysis by graph theory was helpful to comprehensively understand the reconfiguration of brain networks and its significance,and thus provided new insights and evidence for neuroimaging changes in AD patients.