针对医护病人的脑机接口诱发范式优化及其应用研究

Brain computer interface (BCI) technology has gradually developed from the primary stage in laboratory to the application stage in hospitals and rehabilitation centers. The usability of the brain control system would be affected not only by the application environment, but also by the physiology limitation of patients, which was a “big challenge” for brain control system. In this study, the optimization method for stimulus-based brain control system would be presented: 1) New stimulus will be designed to decrease the interference between the adjacent stimuli based on human cognition, and new adaptive strategy will be designed for brain control system to fit the different application environments. 2) Different space distributions accompany with different audio stimuli will be used for audio-based BCI. New music stimuli paradigm will be presented to make the audio-based BCI friendly. The sensitive position of different patients’ bodies will be surveyed for tactile-based BCI. New stimulus presentation pattern based on massage will be used to make the users feel more comfortable. 3) Generic model based on online training method will be used to improve the online performance on patients. Optimized anti-noise and online adaptive algorithm will be used in the brain control system for the application on patients, especial for cerebral Palsy patients. In this study, the performance of stimulus-based brain control system will be improved not only through the new paradigms and the optimized algorithms, but also through the practical experiments on patients, which would improve the practical usability of stimulus-based brain control system.

脑机接口技术已逐渐从面向实验室的初级阶段发展到面向医院康复中心的实际应用阶段。基于脑机接口技术的控制系统（脑控系统），其使用不仅受环境因素的干扰，而且受病人特殊生理因素的限制，这为脑控系统的设计与开发带来了巨大的挑战。本项目针对环境因素和医护病人生理因素提出诱发脑机接口的优化方法：1）基于人类认知规律优化刺激源，减少刺激源间的相互干扰，并设计具有环境适应性的视觉诱发脑控系统。2）提出空间和声音刺激的多样化设计，利用音乐等刺激呈现方式提高听觉脑机接口的友好性。探索触觉脑机接口的刺激形式和位置，利用按摩等刺激呈现形式提高此类范式的舒适性。3）研究基于特定群体的泛化分类模型，提高系统的在线训练能力；研究应用于医护病人的抗干扰和在线修正方法。本项目不仅通过范式优化和算法设计来满足实际应用需要，而且通过临床实践改进系统对于病人的应用性能，从而推进诱发脑控系统的实用化进程。

脑-机接口在脑卒中康复和生活辅助上显示出了巨大的应用潜力。为了进一步推进脑机接口的临床应用，本项目结合环境因素和医护病人生理特殊性提出脑机接口系统设计和优化方法，以解决现在脑机接口临床应用遇到的问题。首先通过刺激呈现优化、多感官刺激、刺激参数调整等方式来提高脑-机接口范式的性能，提高诱发脑电位的幅值和识别准确率；然后，设计运动意图解析算法，通过搭建公用模型和自适应在线修正模型提升脑机接口系统的稳定性；最后，提出优化导联组合和时间窗口等方法提高脑机接口的在线性能。此外，项目针对脑卒中、肌萎缩侧索硬化症人和昏迷病人展开应用研究，开发了基于脑机接口的脑卒中康复系统，信息交互系统和昏迷病人评估系统，近250余名病人使用了以上系统并取得了良好的使用效果。