烟雾病的脑血流动力学异常被认为是其认知障碍的始动因素，并与脑网络的稳定性有关，但具体机制不明。疾病的复杂性导致难以构建合理且稳定的研究模型，使之成为烟雾病认知障碍机制研究的难点与热点。我们在前期研究中基于复杂网络理论，已证实烟雾病认知损害在网络空间上的改变。本课题在此基础上，增加了网络时间属性的分析，更符合疾病真实状态，同时筛选网络中功能改变较大的模块作为易损靶点，结合靶点局部的灌注／代谢负荷以及认知评分，初步构建疾病认知障碍的研究模型。鉴于烟雾病脑血流动力学的复杂性及个体差异，本课题利用医院已有的烟雾病数据库，基于机器学习的思想和智能算法对模型进行训练优化和验证，提高所建模型的科学性与可靠性。最后将该疾病模型与术中导航技术结合，评价模型在认知损害外科治疗中的应用价值。这一疾病模型的建立为烟雾病脑网络稳定性研究提供切入点，也为认知功能的个体化精准评估和治疗提供理论基础和方法学支持。

Cerebral hemodynamic disturbance is deemed to be the cause of cognitive impairment in adult moyamoya disease (MMD). The stability of brain functional interactions is involved, but the detailed mechanism is unclear. Reasonable and stable models are difficult to construct because of disease complexity, making it a difficult and hot spot in the field. Based on graphical theory, our previous work indicates that adult MMD with cognitive impairment exhibits a specific spatial pattern of functional network. To follow real brain interactions, the present study will cover both network spatial and temporal dimensions, and result in several vulnerable modules as targets. Gathering local perfusion/metabolism values of targets and neuropsychological test scores, a function model can be constructed. Considering hemodynamic complexity and individualization, we will take advantage of machine learning and intelligent algorithm to train and test the model, and to improve its scientificity and reliability based on MMD database in our hospital. Finally, we fuse the optimized model with intraoperative navigation system to examine its applied value in surgical treatment of cognitive impairment. The model may offer us an opportunity to both understand the impact of MMD on network stability, and provide theoretical and methodological basis for individualized precise evaluation and treatment of cognitive impairment.