基于视听觉诱导运动神经通道重建的脑控康复机器人关键技术研究及系统集成验证

Brain injury is the leading cause of human movement dysfunction, how to raise the level of brain rehabilitation has become one of the main concerns of modern society. Considering the problems that current rehabilitation devices cannot effectively reconstruct the neural pathways of connecting sense, central and motor nervous systems, this proposal relies on the frontier visual and auditory perception and brain-computer interfaces (BCIs) techniques and proposes a novel research of key techniques of brain-controlled rehabilitation robot for vision and audition evoked motor nerve pathways reconstruction and its systematic integration verification. Firstly, based on the theory of mirror neuron system (MNS) in the brain, this proposal studies the mechanism of collaborative induce of visual and motor intents by biological motion stimulation and the corresponding BCI techniques for rehabilitation applications. Secondly, relying on the stochastic resonance (SR) mechanism embedded in the brain, this proposal studies multi-source noise-induced enhancement of visual and auditory evoked signals. On the basis of above studies, this proposal explores dynamic identification of limb movement intents to carry out cooperatively active and passive rehabilitation techniques for physically disabled patients, and implements brain-controlled preview navigation techniques to aid the movement of patients. This proposal also studies movement decoding from EEG and EMG signals to implement synchronous driving and rehabilitation applications of EMG based orthotics for patients with residual EMG signals. Furthermore, this proposal accomplishes brain condition assessment and rehabilitation program optimization on the fusion of EEG, EMG and fMRI information from patients, which is based on the time-frequency features of EEG and EMG signals and spatial features of fMRI signals evoked quantitatively by visual and auditory stimulations. Finally, this proposal integrates and verifies the brain-controlled rehabilitation robot for vision and audition evoked motor nerve pathways reconstruction, in the hope of leading innovative development of rehabilitation engineering and brain-controlled robot techniques.

脑损伤是人类运动功能障碍的首要诱因，提高脑康复水平已成现代社会关注焦点。课题针对现有康复装置无法有效重建意念—中枢—运动神经通道的难题，依托视听觉认知、脑机接口前沿技术，进行视听觉诱导运动神经通道重建的脑控康复机器人关键技术研究与系统集成验证。首先，基于大脑镜像神经系统原理，研究生物运动的视觉-运动意念协同诱发机理及康复应用脑机接口；同时，依托脑神经系统随机共振机制，实现视听觉诱发信号的多源噪声增强；在此基础上，针对肢体失能患者探索脑运动意图动态辨识的主被动协同康复技术，并实现辅助运动的脑控预瞄导航；针对肌电残留患者实施脑肌运动解码的肌电辅具同步驱动与康复技术；进而基于视听觉定量诱发的脑肌时频响应特征和功能核磁的空间特征，实现脑肌、核磁信息融合的脑康复状态评估与康复方案优化。最后，进行基于视听觉诱导运动神经通道重建的脑控康复机器人集成与验证，引领康复工程与脑控机器人技术的创新发展。

本项目针对意念-中枢反应-运动神经通道重建的技术难题，重点进行视听觉诱导运动神经通道重建的脑控康复机器人关键技术研究与集成系统的性能验证。在视听觉意念诱发方面，创新了具有高交互性和传输率的稳态视觉运动和色彩诱发电位脑机接口范式；依托镜像神经元理论，提出了基于生物运动的视觉-运动意念协同诱发机理及基于视觉诱发的脑控主被动协同康复训练新方法，使得重症中风患者脑机协同训练成为可能；在微弱信号增强与识别方面，揭示了视觉诱发信号的视听耦合噪声增强机制与微弱脑电信号的自适应同频消噪和特征提取技术，探索了大脑神经系统随机共振机制和噪声强化规律；在脑肌电意念解码方面，提出了基于脑功能网络的人体运动脑电信息和基于自主肌电模式信息的连续动作解码技术，实现了脑肌运动解码的肌电辅具同步驱动与康复运动；在脑机交互控制方面，构建了基于脑肌电感知与控制的康复机器人生机交互模型，实现主被动协同康复训练的人体机能补偿控制方法，支持了运动神经通道的康复重建；在脑肌康复状态评估方面，提出了融合脑肌与体感信息的中风康复状态综合评估策略，并给出中风损伤脑电复杂度、脑肌神经传递熵和人体运动姿态近似度等评价指标，为脑中风康复客观、量化评估提高新的评估方法；在脑机协同康复机器人集成与验证方面，基于稳态视觉运动诱发电位，研制了脑控智能轮椅和支持国标汉字库（三千五百字）的脑控中文拼写系统，在渐冻人患者试用中取得了良好效果；设计制作了低成本轻量化康复机械手系统，成功地在两例前臂截肢患者上进行了肌电驱动上肢康复运动辅具原型系统的验证；基于生物运动视觉-运动协同诱发机制，开发了脑控主被动协同下肢康复训练机器人系统，经过20名患者的医院临床对比实验，与传统康复训练手段相比主要康复评价指标提高20%以上，验证了中风脑控康复技术的先进性，为脑中枢功能障碍患者运动神经通道重建与功能康复训练探索了一条新的途径。