跨模态双向感知神经康复系统研究

Stroke has become a serious social development and public health problem in our country. Effective rehabilitation system has great significance to improve the living standard of stroke patients and reduce the burden of family and society. Based on the plasticity, incentive mechanism, stochastic resonance and synchronization mechanism of the nervous system, this project aims at applying the brain-computer interface technology, single-modal and cross-modal stochastic resonance technology, multi-information fusion-based emotion recognition technology and synchronous analysis of EEG-EMG technology to the rehabilitation medicine and developing a cross-modal and bi-directional perception-based neural interface, which is expected to provide tens millions of stroke patients with better neurological rehabilitation system. This project mainly focuses on (1) Pattern recognition algorithms of motor imagery EEG. (2) The mechanism of stochastic resonance to enhance the perception and control ability of the nervous system (3) Emotion recognition method based on information fusion of multiple physiological characteristics. (4) Characteristics and synchronous mechanism of EEG and EMG in rehabilitation exercises. (5) Constructing and making experiments on the cross-modal and bi-directional perception-based rehabilitation system. The cross-modal and bi-directional perception-based neural interface studied on this project will greatly promote the development of neurological rehabilitation system.

中风已成为我国严重的社会发展和公共卫生问题。研究有效的康复手段，对提高中风病人的生活水平和减轻家庭及社会的负担具有重要的意义。本项目面向中风患者，从大脑神经系统的可塑性、激励机制、随机共振以及同步机制出发，旨在将脑机接口、单模态和跨模态的随机共振、多信息融合的情绪识别、脑肌电同步分析等技术引入康复医学，建立基于跨模态和双向感知的神经接口，从而有望为数千万中风患者提供更为理想的神经康复系统。该项目主要研究(1)运动想象脑电模式识别算法。(2)随机共振增强神经系统感知和控制能力的机制。(3)基于多生理特征信息融合的情绪识别方法。(4)康复运动过程中的脑肌电特征及同步关系。(5)基于跨模态和双向感知的神经康复系统的构建和实验。本项目研究的基于跨模态和双向感知的神经接口将有力地促进神经康复系统的发展。

中风是目前威胁人类生命的严重疾病之一。研究有效的康复手段，使患者一定程度上甚至全部恢复正常功能，提高患者的生活质量，对减轻家庭和社会的负担具有重大的意义。目前康复机器人的研究尚存在以下问题： (1) 缺少支配肢体的大脑运动神经系统的直接参与。（2）情绪状态对康复效果的影响缺乏深入认识。项目从神经系统的可塑性、激励机制、随机共振以及同步机制出发，构建了基于神经信息和Barrett WAM的人-康复机器人交互系统，进行康复训练实验。一方面人的主动连续运动意图从脑电信号中识别出来，辅助患者进行“人在回路”的主动康复训练。同时通过基于生理信号的情绪识别系统感知患者康复训练过程中的情绪变化，自适应地调整康复机器人的训练参数和训练模式。另一方面采用力触觉反馈刺激强化感觉运动皮层区mu节律和beta节律的抑制，增强神经系统对康复机器人的控制。本项目的研究有望为数百万患者提供更为理想的双向感知康复系统。