基于混合脑机接口的多维多功能控制的关键技术研究

In recent years, the hybrid brain computer interface combined with a variety of EEG patterns (such as motor imagery, visual evoked P300) has got more and more attention for its advantages of high accuracy and information transfer rate. However, due to the lack of an effective and practical theory and techniques for multi-dimensional and multi-functional hybrid brain-computer interface, this results in a lot of limitations for its performance and application. To this end, the project will explore the brain-computer interface system model based on feedback training for multi-dimensional and multi-functional control. We also conduct research on some of the key technologies to improve the performance and usability of the system. Firstly, in order to improve the coordination, accuracy and conversion speed during execution the paradigm with multiple brain signals integration, we will design the interface with feedback training for three or more brain signal paradigm fusion according to their characteristics. Secondly, in order to reduce the false positive rate, we will study the EEG signal processing algorithms for multiple EEG patterns fusion. In addition, by introducing asynchronous control mode, we will perform fast idle state detection to improve the usability of the system. These technologies will be used in our BCI system. Through a combination of theoretical studies and experimental research, we will design and develop a new multi-dimensional and multi-functional hybrid brain-computer interface system.

近年来，结合多种脑电模式（如运动想象,视觉诱发P300）的混合脑机接口以其较高的准确率与信息传输率等优点受到研究者越来越多的关注。但由于缺乏一套有效实用的，针对多维多功能混合脑机接口的理论和技术，造成了混合脑机接口在性能和应用上很大的局限性。为此，本项目将通过探索基于反馈训练和多维多功能控制的脑机接口系统模型，同时针对其中的一些关键技术进行研究以提高系统的性能及实用性。首先，为了提高系统在融合多种脑信号范式时执行的协调性、准确性及转换速度，拟根据各脑信号范式的特点，设计带反馈功能的界面以进行三种或更多种脑信号范式的融合；其次，为了降低系统控制的假阳性率，拟对多种脑电模式进行融合的信号分析算法研究；另外，通过引入异步控制模式，实现快速的空闲状态检测，提高系统的实用性。这些技术应用于我们的脑机接口系统，通过理论研究与实验研究相结合，设计与开发基于脑电的新型多维多功能混合脑机接口系统。

在该项目执行期间，本课题组基本按照项目计划来执行，各成员按原分工进行。在项目执行期间，获得了一些成果，具体如下：1）在混合脑电模式分析中，我们提出了一种有效融合多种脑电模式的信号分析方法；2）为了充分吸引中风病人进行高容量、且重复性的训练，进而对其中风后的手功能进行改善，我们提出了基于肌电信号的人机界面；3）在意识障碍病人的康复方面，我们也探索了意识受损和所涉及的特定网络相互作用的脑机制。这些成果对开发有效的辅助康复方法提供了重要基础。