视觉和本体觉差异化条件下的平衡脑电传递熵网络分析方法  

Analysis Method of Brain Region Entropy of Human Balance Function Based on Differentiated from Visual Proprioception

作　　者：苏巧钻 罗志增[1] 王哲远 Su Qiaozuan;Luo Zhizeng;Wang Zheyuan(School of Automation,Hangzhou Dianzi University,Hangzhou 310018,China;Hangzhou Dianzi University ITMO Joint Institute,Hangzhou 310018,China)

机构地区：[1]杭州电子科技大学自动化学院,杭州310018 [2]杭州电子科技大学圣光机联合学院,杭州310018

出　　处：《中国生物医学工程学报》2024年第3期286-294,共9页Chinese Journal of Biomedical Engineering

基　　金：国家自然科学基金(62171171);浙江省自然科学基金(LZ23F030005)。

摘　　要：平衡是人体完成各项运动的基础,现有人体平衡能力评估方法多基于外在表现。本研究以平衡中枢神经调节这一内源性角度为切入点,研究静力平衡调节过程的感觉运动皮层整合,分析大脑皮层的激活状态,建立静力平衡脑电传递熵网络。设计了视觉输入和本体觉输入差异化条件下的实验范式,定义了平衡脑电的相位同步化准则,构建突破因果关系双节点分析模型,提出一种以平衡事件为驱动的传递熵网络分析方法。先将节点间的因果关系降级为相关关系,进而将因果关系的研究转移至功能脑区层面,以契合人体平衡协调过程脑区间信息传递的相位同步关系。在20名被试脑电数据的基础上,基于相位同步关系确定平衡事件的中枢神经调节时间段提取平衡脑电多种内源性特征,以网络聚类系数(C),最短路径(E)和最大李雅普诺夫指数(MLE)的特征组合[C,E,MLE],用支持向量机分类。与传统网络特征分类结果对比,平均分类准确率提升了14.66%。传递熵网络分析中补充最大李雅普诺夫后能更充分表达人体平衡调节的内在规律演进过程,提高了人体平衡的分类能力。Balance is the foundation of all human movements,and existing methods of assessing human balance are mostly based on external performance.In this paper,we took the endogenous perspective of balance central neuromodulation as an entry point to study the sensorimotor cortical integration in the process of static balance regulation,analyze the activation state of the cerebral cortex,and establish an entropic network of static balance EEG transmission.The experimental paradigm was designed under the conditions of differentiation of visual and proprioceptive inputs,and the phase synchronization criterion of balance EEG was defined.The phase synchronization relationship of the transmission between the brain regions of balance information was defined.Based on the EEG data of 20 subjects,the central nervous system regulation period of balance events was determined based on the phase synchronization relationship,various endogenous characteristics of balanced EEG were extracted,and the average classification accuracy was improved by 14.66% compared with the traditional network feature classification results by using a combination of network clustering coefficient(C),shortest path(E)and maximum Lyapunov index(MLE)[C,E,MLE].The new feature of maximum Lyapunov index(M)added in the analysis of transfer entropy network fully expressed the internal law evolution process of human balance adjustment and improved the classification ability of human balance.

关 键 词：人体平衡评估 平衡脑电 相位同步 传递熵 

分 类 号：R318[医药卫生—生物医学工程]