【作者】 门卫伟；

【导师】 范明霞；

【作者基本信息】 华东师范大学 ， 无线电物理， 2013， 博士

【摘要】 本研究的主要目的是构建一套适用于中国人脑科学研究的大样本中国人脑模版、组织概率图、解剖标准模板和皮层下神经核团模板,并使之能实际应用于中国人脑科学研究。目前常用的人脑模板是基于西方人脑数据所构建,由于中西方人种的不同,中西方人脑形态存在较大差异。为了找到构建大样本中国人脑模板的理论依据,本文对120名中国人脑数据和120名年龄与性别完全匹配的西方人脑数据进行了详细的人脑整体形态学比较,结果发现中西方人脑在体积上无统计学差异；西方人脑较中国人脑长,中国人脑较西方人脑宽、高；中西方人脑在顶视投影图上,中国人脑均比西方人脑圆。为了验证中西方人脑局部的形态学差异,本研究编写了一套高分精度胼胝体厚度分析软件对中西方人脑胼胝体厚度进行了对比分析,并将这套分析方法应用于爱因斯坦大脑的胼胝体研究。结果发现无论是中国男女性还是西方男女性胼胝体在厚度上均无统计学差异,中国人群在局部比西方人群胼胝体厚,说明中国人脑左右半球间的联系更加密切；爱因斯坦大脑胼胝体研究结果揭示爱因斯坦超凡的智力与天赋不仅与其独特的大脑沟回及组织学特点有关还与其左右大脑半球间的密切联系相关。以上结果说明中西方人脑由于人种的不同确实存在差异,由于这种形态学上的差异,使用西方人脑模板对中国人脑进行研究时,中国人脑数据在空间标准化过程中需要较大的空间变形,从而会影响数据的标准化结果以及功能定位的准确性。目前已有多家研究机构对中国人脑模板进行了研究,但这些模板研究均存在一定的问题,如被试人数有限、男女性别不对称,中国人脑标准空间选择不具有代表性,模板形式单一不能满足实际科研需要等。本文针对以上中国人脑模板研究中的局限性并结合本实验室的优势条件从1363人次健康志愿者高分辨率人脑结构像数据中筛选出200例右利手健康青年汉族志愿者脑部数据,男女各100名,年龄18-29岁,平均22.97±2.29岁。使用统计参数图软件SPM8及多种数据处理软件和自编程序进行复杂的图像处理,采用线性和非线性配准算法相结合的方式构建了一个不依赖于单个个体脑空间的左右完全对称的中国人脑标准空间。在此标准空间的基础上构建了高分辨率中国人脑T1、T2加权像脑模板及五种参数的DTI脑模板、用于人脑组织分割的中国人脑组织概率图、用于解剖定位的中国人脑解剖标记模板、以及用于皮质下神经核团定位的中国人脑皮质下神经核团模板,该套模板是一个完整的系列可以满足中国人脑科学研究的需要。为了验证该套大样本中国人脑模板在中国人群脑科学研究中组织分割、空间标准化及功能定位的准确性,本文对五位健康志愿者高分辨率结构像与功能定位任务数据分别使用西方人脑模板和本文构建的大样本中国人脑模板进行数据处理与分析。结果显示使用本文构建的大样本中国人脑模板及组织概率图可以正确分割中国人群人脑组织,分割精度要优于SPM8软件自带的组织概率图分割结果；空间标准化及功能定位结果显示使用本文构建的大样本中国人脑模板与组织概率图可以准确对中国人群人脑数据进行空间标准化,功能激活结果定位精度可以达到或超过现有的西方人脑模板精度。为了进一步验证本文构建的大样本中国人脑模板在实际中国人脑科学研究中的使用效果,本文对青少年广泛性焦虑症患者及正常对照组的大脑结构像数据进行了基于体素的形态学分析。结果显示广泛性焦虑症患者右侧尾状核头和伏隔核体积均小于正常对照组(p<0.001未校正),此结果与广泛性焦虑症的临床表征具有相关性。广泛性焦虑症组大脑灰质体积异常降低区域均显示于中国人脑模板灰质区域,从而证实了本文构建的大样本中国人脑模板具有实用性并且分析结果可靠。本文还使用该套大样本中国人脑模板与组织概率图对目前几种常用的功能磁共振数据处理方法进行了比较和分析。证明了常规的数据采集与分析方法中存在激活结果不可靠脑区,这些脑区主要存在于左右侧额极、颞下回和颅底直回等与人脑表面较近的空气腔附近。另外还给出了能在一定程度上提高功能磁共振成像激活结果可靠性的方法,即通过场图校正技术对功能像变形进行校正以及使用联合分割方法对平均功能像进行分割并进行空间标准化的方法。更多还原

【Abstract】 The mainly purpose of this study is to construct a set of large sample Chinese brain templates, tissue probability maps, anatomy label template and subcortical nuclei template for use of Chinese brain science. So far, the common brain template, ICBM152template, which constructed by large sample western cohort. Due to the difference of human species, the cerebral morphology will be different between Chinese and Caucasian. In order to find the theory evidence for construct the large sample Chinese brain templates,120Chinese brain MRI data and120age and sexual well matched Caucasian brain MRI data were processed and compared in brain volume and global morphology, the results shown that there are no statistical difference in brain volume between Chinese and Caucasian, the brain of Caucasian is longer than that of Chinese, Chinese’s brain is wider and higher than that of Caucasian. In the top view of the brain projections both Chinese and Caucasian, Chinese brain is more round than that of Caucasian, and the female brain is more round than that of male’s whatever Chinese or Caucasian. To detect the difference of local cerebral morphology between Chinese and Caucasian, a novel callosal analysis method with high precision was developed which based on thickness of corpus callosum. After that, this analysis method was successful used in the research of Einstein’s brain and two different age control groups. The results shown that there are no difference in statistic between female and male whatever Chinese or Caucasian, but Chinese have bigger corpus callosum than Caucasian and thicker in callosal rostrum, genu and splenium, which means that the connectivity between two hemispheres are more closer. The findings of Einstein’s corpus callosum between two different age control groups suggest that Einstein’s extraordinary cognition was related not only to his unique cortical structure and cytoarchitectonics, but also involved enhanced communication routes between at least some parts of his two cerebral hemispheres. The results above indicate that there are differences of brain morphology between Chinese and Caucasian due to the different human species, so if the western brain template was used in Chinese brain research, there need larger spatial distortion during normalization of Chinese brain data, which will be affected the spatial normalization results and the accuracy of functional localization.Some research institutes have did some studies about Chinese brain templates, however these researches of Chinese brain templates all have some limitations. This study point to those limitations and combine the advantage conditions in our laboratory,200healthy Chinese youth subjects high quality MRI data of head from1363subjects data base were random selected, all subjects are right-handed and the Han nationality,100males and100females, age from18to29years old, mean22.97±2.29years old, males and females are well matched in age. The MRI data were processed by the toolbox of SPM8, various softwares and some house made codes, the normalization algorithm of linear and nonlinear were combined and employed. By data processing, a completely symmetrical Chinese brain standard space was constructed, which was not dependent on any individual brain space. This study constructed high resolution T1and T2weighted brain templates, six different parameters DTI brain templates, Chinese brain tissue probability maps use for brain tissues segmentation, Chinese brain anatomy label template for anatomy localization and a Chinese subcortical nuclei template for nuclei localization. Those brain templates are a holonomic series which could be satisfied the need of Chinese brain researches.For test the accuracy with the constructed large sample Chinese brain templates in Chinese brain research, this study processed five healthy Chinese subjects high resolution structure images and task data of functional localization with constructed Chinese brain templates and western brain templates, respectively. The results shown that the Chinese tissue probability maps can correct segment the brain tissues of Chinese brain, the accuracy of segmentation is better than that of tissue probability maps which included in SPM toolbox; the results of spatial normalization and functional localization shown that with this set of Chinese brain templates can accurate normalize Chinese cohort brain into its standard space, the accuracy of functional activity reached or exceed that of western brain templates.In this study, fifteen adolescents with and without generalized anxiety disorder were employed respectively, the MRI structure data were processed by SPM8and the constructed Chinese brain templates for voxel-based morphology analysis, the results further authenticates that the constructed Chinese brain templates are practical and reliable. Three different fMRI data spatial normalization methods with or without the field map distortion correction were compared and analyzed by our constructed large sample Chinese brain templates and brain tissue probability maps, the results shown that there are the unreliable brain regions in the common fMRI data scanning and processing methods, mainly exist in the bilateral poles of temporal, inferior temporal gyri, straight gyri, poles of frontal, margins inferior of cerebel and calvarium, which are near the large airfilled cavities of the human brain. It can increase the reliability of activity or analysis results using "unite segment" method segments the mean functional image by distortion correction of the EPI data and the generated normalization parameter to be used for normalize the functional images into the standard brain space.更多还原