基于多模态神经影像研究人脑结构连接和功能连接的关系，可为理解人脑的结构-功能关系如何成为认知和行为的基础提供重要依据。但现有研究主要集中在结构连接-功能连接的对应关系与连接属性的差异性对比，忽略了结构连接的交互模式与功能连接的关系，及结构连接如何在短时间尺度上交互产生动态变化的功能连接等问题。图神经网络可以基于图数据构建深度神经网络框架，学习节点和连接的多层表征模式，在提取表征过程中获得基于连接的信息传递方式，从而获得不同图数据间复杂高阶的映射关系。本项目拟采用图神经网络的方法，从结构连接生成多个时间点的功能信号以及功能连接，通过模型解析，获得结构连接交互模式，推演网络信息传递的模式，以阐释结构连接和功能连接的关系，探索大脑结构对功能的影响，并进一步研究轻度认知障碍患者在经过认知功能训练前后，脑结构和功能连接关系的变化，为揭示轻度认知障碍疾病的认知康复机制提供新的线索，具有重要的学术价值。

The relationship of human brain structure and function is important for understanding its key role in cognition and behavior by using multi-modality imaging technology. Previous studies mainly focused on the differences of the corresponding connections between brain regions and topological properties by comparing the structural connectivity and functional connectivity, and ignoring the relationship between the structural interaction mode and the functional connection, as well as how the structural connection interacts dynamically to generate the dynamic connection on the short time scale. Graph neural network can learn the multilayer contributes of nodes and connections, and obtain the complex higher-order relationships between two graph data by recognizing the information transfer mode between nodes based on structural connections in the process of feature extraction. This project plans to use the graph neural network to generate functional signals at multiple time points and functional connections from structural connections. The mode of structural connection interaction is obtained through the analysis of the model. Then, the information transmission mode of the network is deduced to explain the relationship between structural and functional connections and explore how the static brain structure organizations are the basis of flexible brain function. Changes of the relationship between brain structural and functional connections of patients with mild cognitive impairment before and after cognitive training will also be studied, which may provide new clues for revealing the rehabilitation mechanism of mild cognitive impairment.