严重脑损伤患者意识水平的大数据多模态评估及预后预测建模研究

Along with the development of neurological intensive technologies and ideas, more and more severe brain injury patients will survive in the serious disturbance of consciousness state (DOC), which is a great challenge and burden. Accurate and objective index has an important significance in prejudgment whether the patient will recover consciousness. Due to the limitation of the understanding for consciousness mechanism and clinical observation, precise prejudgment is still unable to be achieved. Recent years, advancement in neuroimaging and electrophysiological technique provide an objective means for consciousness research. But most existing research were single model, without high predictability, sensitivity and specificity. They have small sample size, low reliability and lack of accurate prediction model. Our previous functional magnetic imaging study have found the brain functional connectivity, white matter characteristics, amplitude of the low-frequency and auditory task related brain activity were related with the level of consciousness and prognosis. This project aims to be multimodal monitoring in big data DOC patients and long-term follow up. To evaluate the consciousness level by combining the conscious level with the grey matter volume, white matter characteristics, functional connectivity strength within brain, Glutamate/Glutamine and GABA level, auditory network activities and EEG. Support vector machine will be used to make up prognosis prediction modeling and achieve an objective means to predict the prognosis for patients with DOC.

随着神经重症技术和理念的进展，越来越多严重脑损伤患者得以存活而处于严重意识障碍状态（DOC),是巨大挑战和负担。预判患者是否能恢复意识的准确客观指标具有重要意义,由于脑意识机制理解和临床观察预判的局限性而目前仍无法实现。近些年来神经影像和电生理技术的进展为意识研究提供了客观手段。但现有研究大多单一模态，预测灵敏度和特异性不高；样本量小可靠性低；缺乏准确预测模型。本课题组已有的脑损伤患者磁共振研究发现，脑区的功能连接、传导束特性、低频脑活动波幅以及听觉任务相关的脑活动与意识水平和预后存在相关。本课题拟大数据病例多模态监测长期追踪，从脑灰质体积、传导束特性、脑区功能连接强度、递质Glutamate/Glutamine和GABA水平、听觉网络活动和脑电结合意识行为学评分来客观评估意识水平，并利用与预后的相关性进行支持向量机SVM预后预测建模，从而力求为DOC患者提供准确预测意识预后的客观手段。

随着脑损伤理念和神经重症技术的进步，越来越多重型脑损伤昏迷患者存活，转归为严重意识障碍如植物状态和微意识状态。这些患者的长期救治给社会和家庭带来极为沉重的精神打击和巨大的经济负担。课题负责人在课题完成期间，建立了不同水平意识障碍患者的功能磁共振、PET-CT、高密度脑电的数字化大样本的数据库；建立了结合神经外科、康复医学科、影像科、中西医结合科的意识障碍多学科临床和研究团队，对意识障碍的诊疗进行深入研究，开展了意识的灰质皮层神经网络机制探索、意识的白质传导束和神经递质机制探索、建立意识障碍患者是否能苏醒的预测模式、建立基于意识神经网络为基础的意识障碍患者催醒治疗新方法。发表SCI论文4篇，发布于Human brain mapping、Neuroscience Bulletin、World neurosurgery、Scientific Reports、中华急诊医学杂志等杂志；正在申请国家发明专利2项，申请成功国家实用新型专利1项。