基于多模态磁共振成像的认知老化的人脑连接组学研究

Cognitive decline is commonly observed with advanced aging in both healthy elderly and patients with mild cognitive impairment (MCI). However, the underlying neural mechanisms of normal and abnormal cognitive aging remained unclear. In this longitudinally-designed project, we will utilize multimodal MRI data to explore the brain mechanisms of cognitive aging in both healthy elderly and MCI patients from the perspective of human brain connectome. Both high-resolution individual structural and functional brain networks will be constructed for each participant, and the topological properties of the brain networks will be quantified with the graph theoretical analyses. First, we aim to compare the age-related trajectory of brain structural and functional connectome between healthy elderly and MCI patients, and further examine the relationship with the longitudinal cognitive decline with aging. Second, we will combine the brain connectome model with machine-learning approaches, to select the brain network-based features, and develop the effective classifier for the individual classification of MCI patients and individual prediction of longitudinal cognitive decline. Finally, we will develop the brain connectome-based prediction platform for normal and abnormal cognitive aging. This project is important for the understanding of brain mechanisms of cognitive aging in both normal and abnormal populations. The project applicant has worked on the methodology and applications of human brain connectome for 9 years, and has published 22 SCI papers as the first or corresponding author, accumulating plenty of research experience for this project.

不论对于正常老化，还是轻度认知障碍，都伴随着多种高级认知能力随年龄的逐渐衰退，然而正常和异常认知老化的神经机制尚不明确。本项目拟采用多模态磁共振成像技术对正常和异常老年群体进行纵向追踪研究，从人脑连接组学的角度，通过构建个体水平高分辨率的脑结构和功能网络，结合现代图论的定量分析，刻画正常和异常老化过程脑网络的拓扑属性随年龄的变化轨迹，并进一步考察与认知衰退之间的关系。此外，我们将人脑连接组学与机器学习方法相结合，进行基于多模态脑网络的特征提取和分类器设计，训练出能有效预测异常认知老化的分类器，进行个体判别和早期预测。此外，开发相应的基于人脑连接组的认知老化的早期预测系统，有望用于对异常认知老化和神经退行性疾病的早期辅助诊断。该研究对于揭示正常和异常认知老化的神经机制具有重要科学意义。本项目申请人从事人脑连接组学研究已有9年，以第一或通讯作者发表SCI论文22篇，为该项目积累了丰富经验。

本项目采用多模态磁共振成像技术对认知正常和异常老年群体进行纵向追踪研究，从人脑连接组学角度，通过构建个体水平多个空间尺度下的大脑结构和功能网络，结合现代图论的定量分析方法，系统刻画了正常和异常老化过程脑网络的拓扑属性随年龄的变化轨迹，并进一步揭示了与认知衰退，和分子水平葡萄糖代谢之间的关系。此外，我们比较正常与异常老化群体脑网络不同变化模式，以提取对异常认知老化早期敏感的脑网络标记物。进一步，我们将人脑连接组学与机器学习方法相结合，进行基于多模态脑网络的特征提取和分类器设计，训练出能有效预测阿尔茨海默病（AD）早期及高危人群的分类判别模型，进行AD患者个体判别和早期预测。最后，基于大样本多中心神经影像数据，开发了基于人脑连接组的多领域认知功能（执行功能，注意，记忆和语言）和大脑年龄的个体预测模型，有望用于对异常认知老化和神经退行性疾病的早期辅助诊断。该研究对于揭示正常和异常认知老化的神经机制具有重要科学意义。通过该项目实施，申请人围绕认知老化人脑连接组学方向发表SCI论文8篇（IF>5的6篇），培养年轻教师1名，合格毕业博士研究生2名，硕士研究生3名。项目组成员参加该领域的国内外重要学术会议5次，并做口头报告4次。