基于多模态磁共振的脑白质病相关认知功能障碍的动态脑网络机制研究

White matter lesions (WML) are common in older people and have been demonstrated to be associated with cognitive decline. However, the mechanism of cognitive impairment associated with WML remains largely unknown. There are still lack of effective early diagnosis and treatment in clinic, and ideal animal models that can fully simulating the disease. So far, the technique of brain network analysis based on multi-modal MRI has been verified as breakthroughs in the mechanisms of many neurological and psychiatric diseases, and may provide a clue to explore the mechanisms of WML and associated cognitive impairment. However, until now, little research has been performed for the disease. In previous study, we found significant abnormalities of cortical-subcortical and cingulated-related network in cognitive impairment associated with WML, and indicated that functional connectivity could be used as a neuroimaging marker for discriminating the patients of WML with cognitive impairment from WML without cognitive impairment. Therefore, we plan to conduct a large sample, cross-sectional and longitudinal combined study, using multi-modal MRI and fusion analysis techniques to explore the characteristic pattern of brain network architecture in cognitive impairment associated with WML. Furthermore, we will establish models for classification and early prediction of cognitive impairment associated with WML, by machine learning techniques. In addition, we intend to conduct an experimental treatment with repetitive transcranial magnetic stimulation (rTMS) in patients of cognitive impairment associated with WML targeting key damage areas for the disease, and observe the changes of brain networks before and after rTMS intervention. The study will reveal the dynamic brain network mechanisms of cognitive impairment associated with WML from multiple perspectives, and may provide a novel diagnostic and therapeutic direction for cognitive impairment associated with WML.

脑白质病（WML）在老年人群中极为常见，研究证实其和认知功能障碍密切相关。但WML相关认知障碍的具体机制尚不明确，亦缺乏有效的早期诊疗手段及能完整模拟本病的理想动物模型。基于多模态磁共振的脑网络分析已在多种神经精神疾病的机制研究中取得突破性的成果，可能会为研究WML相关认知障碍提供良好的切入点，但目前相关研究仍很有限。前期工作中我们发现，WML伴认知障碍患者的皮层-皮层下及扣带相关脑网络呈现明显异常，并可利用功能连接有效区分WML伴及不伴认知障碍患者。基于此，本项目拟采用多模态磁共振及融合分析技术，在较大样本中横向与纵向结合研究WML相关认知障碍的脑网络特异性异常及演变模式；进一步地，利用机器学习技术进行本病的早期识别和预测，并针对受损关键脑区开展rTMS试验性治疗及观察干预前后脑网络的动态变化，多角度深入探索疾病发生发展过程中的动态脑网络机制，为WML相关认知功能障碍的诊疗开辟新的方向。

脑白质病（white matter lesions, WML）是脑小血管病的典型影像学表现之一，在老年人群中极为常见，与认知功能障碍密切相关。利用神经影像和计算机技术开展脑网络分析，从脑网络、脑连接的角度探索各脑区间结构和功能的相互依赖关系及随疾病进展发生的变化，可能将为深入研究WML 致认知损害的潜在机制提供新的切入点，但相关的研究非常有限。在既往的工作中我们发现，WML伴认知障碍患者的扣带回和皮层-皮层下相关脑网络出现了异常，并与认知障碍的程度相关。本研究在前期基础上进一步扩大样本量，对WML患者进行了跟踪随访，建立了一套完整的WML患者临床、多模态影像和生物样本数据库；横断面和纵向研究相结合，明确了WML相关认知障碍的脑网络、脑连接、脑血流、脑微结构的特征性异常及演变模式，探索了WML相关认知障碍背后的动态脑网络机制和脑网络破坏关键脑区；在此基础上结合机器学习技术，利用脑网络和脑连接指标实现了对WML相关认知障碍的准确分类和早期预测，并在WML相关认知障碍患者中开展了针对脑网络的关键受损脑区的rTMS试验性治疗。进一步地，我们还开发了一套可以广泛应用于临床和科研的WML自动识别和分割系统，不仅能对WML还能针对其他多种伴随的颅内病变进行精准的定量分析，为WML相关认知障碍的机制研究和临床诊疗开辟新的思路。