面向帕金森病的多模态在线预警方法研究

Early diagnosis of Parkinson’s Disease (PD) has very important practical significance for timely slowing, preventing and controlling the disease so that it is the ideal way to cure PD. For intelligent early diagnosis of PD, the project is based on the medical knowledge and clinical experience on PD and plans to start research by importing online incremental machine learning algorithm on the platform of multimodal wearable equipment like electroencephalogram (EEG) device, electromyography (EMG) device and intelligent bracelet. And the project will focus on several medical tasks, such as multimodal physiological parameters sensing and medical feature extraction for PD early diagnosis and building and online updating the PD early diagnosis model, etc. At the end, the project will realize an automatic PD early diagnosis system for the elder. The ICT (Institute of Computing Technology, Chinese Academy of Science) and the XWH(Xuanwu Hospital Capital Medical University) will take part in the project and play their respective advantages of the IntelliSense, the new interactive, PD medical ontology and cognitive modeling, etc. And the project will also deeply explore the online additional relationship model between the multimodal physiological parameters and PD’s medical features in order to make the proposed model get clinically validated and positively applied.

帕金森病（Parkinson’s Disease，PD）的早期预警对于及时延缓、阻止或控制病情的发展，具有非常重要的现实意义，是治疗PD的理想方法。本项目面向PD的智能化早期预警，拟基于PD的相关医学知识和临床诊断经验，利用脑电仪、肌电仪、智能手环等多模态可穿戴感知设备，引入在线增量式机器学习算法和自然人机交互技术展开研究。项目将重点研究面向PD预警的多模态生理参数感知和医学特征提取、PD预警模型的构建与在线更新等内容，最后实现一套面向老年人群的PD早期自动预警系统。项目将联合中科院计算所和首都医科大学宣武医院，发挥各自在智能感知、新型交互、PD医学知识本体、认知建模等方面的优势，深度探索多模态生理参数与PD医学特征之间的在线增量式关系模型，力争在PD早期自动预警上得到临床验证和积极应用。

帕金森病（Parkinson’s Disease，PD）的早期预警对于及时延缓、阻止或控制病情的发展，具有非常重要的现实意义，是治疗PD的理想方法。本项目面向帕金森病的多模态在线预警问题开展研究，研究实现了面向PD预警的多模态异构数据协同感知方法、PD预警模型构建方法和模型的在线更新学习方法，研发了一套PD早期自动预警原型系统并在首都医科大学宣武医院和中国帕金森联盟的200多家成员医院临床应用，帮助提升帕金森病平均诊断率约10%。相关成果在IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING(IEEE TKDE)、IEEE International Conference on Bioinformatics and Biomedicine(IEEE BIBM)等国内外顶级期刊会议上发表论文9篇，申请国家发明专利4项、软件著作权1项，培养2名博士和3名硕士，获得2017年CCF科学技术奖技术发明一等奖、第十三届全国人机交互学术会议（CHCI 2017）最佳论文提名奖等奖励。