基于ERD和ERP混合特征的非视觉依赖脑-机接口关键技术研究

Brain-computer interfaces (BCIs) based on event-related desynchronization (ERD) and tactile/auditory evoked event-related potential (ERP) are available for application scenarios where users have restricted vision or the BCI operating environment requires extra visual interaction, which have significant potential in the fields of control and interaction as they require no user visual interaction to operate (so called vision independent BCIs). This project aims to develop a vision independent BCI using the hybrid features of ERD and tactile-auditory evoked ERP, and the direction-congruent bimodal paradigm. The proposed hybrid BCI could provide more abundant, significant and higher signal-to-noise rate multimodal cognitive EEG features. To further enhance the accuracy and speed of the BCI system, this project also includes studying the neural mechanism of feature integration inducing by the affects and interferences among features, and developing a new information fusion algorithm for fast decoding the hybrid features. The proposed vision independent BCI could provide a new theory and method for state-of-the-art control and interaction technologies in the fields of rehabilitation, weapon systems and manned space engineering.

基于事件相关去同步（ERD）和触/听觉响应诱发事件相关电位（ERP）的脑-机接口（BCI）的交互控制不依赖于人的视觉通道，适用于操控者视觉受限以及系统操控需要额外的视觉交互等应用场景（属于非视觉依赖BCI），在控制与交互领域具有强大应用潜力。项目采用ERD和触听觉ERP脑电混合特征，基于目标方向一致性设计方法构建非视觉依赖混合BCI范式，以获得有用信息量大、特征显著、信噪比高的多模态认知脑电，研究由各种模态特征之间的相互影响和干涉形成的特征整合神经机制，提出高效解码混合特征的信息融合新算法，提高该BCI系统的准确率和速度，为康复助残、武器装备和载人航天等领域的新型控制与交互技术研究提供新理论、新方法。

本项目研究基于ERD与ERP高性能混合BCI技术，以提高非视觉依赖BCI的速度和准确率为方向牵引，计划实现应用于视觉受限或视觉通道被占用的复杂环境和任务下的先进脑机交互技术，提升非视觉依赖BCI系统的整体性能。围绕预期研究目标，项目组开展了针对性的研究，揭示了混合BCI范式中ERD和ERP特征的演进规律和信息整合神经机制，建立了ERD和ERP的混合特征模型；设计出高效的ERD和ERP混合特征信息的特征提取与识别分类新算法，提出了适用于基于ERD和ERP的非视觉依赖BCI的信息融合策略；构建了多目标非视觉依赖BCI在线操控系统，本项目的研究成果在康复助残领域具有重要的实用价值。