基于ECoG的想象与执行手部空间运动和不同握力负荷作用神经编解码机制比较研究

Brain science research and artificial intelligence innovation are urgent to explore cerebral cortex neural encoding and decoding mechanism of imagining and executing limb space movement and grip load. Electrocorticography (ECoG) has the unique advantages of high spatial resolution，signal-to-noise ratio, and directly observation of functional cortex neuron electrical activity and cognitive changes, which are especially suitable for revealing the brain encoding and decoding mechanism. But ECoG has not yet been fully used due to the difficulties of the implantation surgery and maintenance of good performance. Additionally, few studies were reported on the neural encoding and decoding mechanism of imaging and executing hand space movement and grip load. This project intends to raise patients with epilepsy who have been implanted cortical electrodes to join experiments of imagination and the execution the hand space movement, and design new ECoG acquisition paradigm for imaging and executing hand space motion under different grip load; then，analyze the characteristics of neuronal activation and functional network in the brain area during hand motor imagery and execution, and reveal neural encoding mechanism in motor cortex and compare the differences; finally decode imagination and execution model of hand space motion and grip load by using sparse Bayesian method. This project is expected to provide a scientific basis for revealing neural encoding and decoding mechanism of brain motor function, and a new idea for the design of new ECoG brain computer interface and bionic intelligent robot.

脑科学研究、人工智能创新皆急需探索想象与执行肢体空间运动和握力负荷的皮层神经编解码机制。皮层脑电（ECoG）有高时空分辨率、优质信噪比，可直接观察皮层功能区神经元集群电活动和认知思维变化规律等独特优势而尤适合揭示脑神经编解码机制，但因植入手术及维护良好性能困难，尚未充分应用。而想象与执行手部空间运动和握力负荷的神经编解码机制亦鲜见研究。项目拟募集已植入皮层电极的癫痫患者参加手部运动想象与执行实验，设计想象与执行手部不同握力负荷下空间运动的ECoG采集新范式；分析手部空间运动想象与执行时相关脑区神经元群激活模式与功能网络特征，揭示其运动皮层神经编码机制并比较差异；最后采用稀疏贝叶斯方法解码想象与执行手部空间运动和握力负荷作用模式。项目可望为揭示脑运动功能神经编解码机制提供科学依据，为新型ECoG脑机接口和仿生智能机器人设计提供新思路。

脑科学研究、人工智能创新皆急需探索运动想象和运动执行的编解码机制。本项目设计了多种手部运动执行与运动想象的实验范式，精细化的实验范式为进一步探索运动想象的神经机制提供基础；与项目组研究团队配合，根据设计的实验范式采集了ECoG、SEEG、EEG等脑信号数据；对比分析了手部运动想象和运动执行的脑网络特征，发现大脑相关运动区之间的脑因果连接关系和连接强度是与被试者的任务手和利手相关，并且手指运动执行的脑网络图比手指运动想象具有更多的有效连接；在脑信息处理的预处理、信号特征提取，数据增强，模式分类等各阶段提出了多种信号处理新方法，提升了分类解码的准确率；研究了在线神经反馈方法，构建了一种基于实时神经反馈的脑电分析处理系统，发现神经反馈方法可被试为主动的精神或运动想象中起到调节的作用和有效性；构建了基于脑机接口的脑控应用系统，实现了脑控四旋翼飞机应用。项目为揭示脑运动功能神经编解码机制提供科学依据，为新型运动想象脑机接口和脑控智能机器人设计提供了技术基础。