面向运动想象康复训练的脑机交互系统研发
发布时间：2019-03-21 07:22
【摘要】：运动想象(Motor Imagery,MI)训练是一种新型康复训练方法。本文借助脑机交互系统,通过神经反馈的方式,对其增强MI康复训练效果进行探索。本文首先提出一种MI康复训练脑机交互系统框架,再就MI脑电信号(Electroencephal ogra-m,EEG)的眼电伪迹(Ocular Artifact,OA)去除算法、特征提取算法以及分类算法的编程实现进行研究,并构建相应功能模块,组成在线MI康复训练脑机交互系统,并就有无神经反馈的情况下,MI训练的效果作对比研究,对所研发系统的有效性进行验证。本文的主要研究内容可分为以下5个方面:(1)本文介绍了系统的基本概念、系统的组成以及国内外的研究现状,并分析目前该类系统研究中的关键技术难题。同时,了解人脑的结构与EEG产生的机理以及MI过程中EEG具有的事件相关去同步/同步(Event-Related Desynchr-onization/Synchronization,ERD/ERS)现象,以此作研究的理论支撑。(2)提出系统的总体架构以及各模块应具备的功能,并设计EEG采集方案,介绍采集所需的实验设备和实验对象,并提出实验中需要注意的要点,最后记录实验中具体采集情况。(3)提出一种自动去除OA的方法:首先将水平和垂直眼电(ElectroOculogram,EOG)信号按一定比例混叠成一导新的信号,与EEG一起通过改进独立分量分析(Improved Independent Component Analysis,IICA)算法获取各导信号的独立分量,再利用相关系数自动识别并去除混叠信号独立分量,最后通过ICA逆变换获取纯净EEG。(4)EEG的特征提取与分类研究分二个方面展开:先由小波变换获取EEG的小波能量,再计算相对小波能量作为特征;再构建Logistic分类器对特征进行分类。(5)完成EEG在线分析处理功能,与神经反馈功能,实现系统整体构建。最终,该系统既能分析已保存的EEG,又能在线实时处理EEG,并将处理结果转换成控制信号,完成虚拟人体模型的控制,反馈用户MI状态。在线实验结果表明该系统能辅助受试者更有效地进行MI,从而提升康复训练效果。
[Abstract]:Motor imagination (Motor Imagery,MI) training is a new method of rehabilitation training. In this paper, with the help of brain-computer interaction system, through the way of neural feedback, we explore how to enhance the effect of MI rehabilitation training. In this paper, we first propose a framework of brain-computer interaction system for MI rehabilitation training, and then study the MI EEG signal (Electroencephal ogra-m,EEG) eye artifact (Ocular Artifact,OA) removal algorithm, feature extraction algorithm and the programming implementation of classification algorithm. The corresponding functional modules are constructed to form a brain-computer interactive system for online MI rehabilitation training. The effects of MI training are compared with or without neural feedback, and the validity of the developed system is verified. The main contents of this paper can be divided into the following five aspects: (1) this paper introduces the basic concept of the system, the composition of the system and the research status at home and abroad, and analyzes the key technical problems in the current research of this kind of system. At the same time, we understand the structure of the human brain and the mechanism of EEG generation and the event-related desynchronization / synchronization (Event-Related Desynchr-onization/Synchronization,ERD/ERS) phenomenon of EEG in the process of MI. (2) the overall architecture of the system and the functions of each module are proposed, and the EEG acquisition scheme is designed, the experimental equipment and objects required for the collection are introduced, and the main points needing attention in the experiment are put forward, and the main points for attention in the experiment are put forward, and the main points to be paid attention to in the experiment are put forward. Finally, the specific data collected in the experiment are recorded. (3) an automatic method of removing OA is proposed: firstly, the horizontal and vertical ElectroOculogram,EOG signals are mixed into a new signal in a certain proportion. Together with EEG, an improved Independent component Analysis (Improved Independent Component Analysis,IICA) algorithm is used to obtain the independent components of each derived signal, and then the correlation coefficient is used to automatically identify and remove the independent components of the aliasing signals. Finally, the feature extraction and classification of pure EEG. (4) EEG obtained by inverse ICA transform are divided into two aspects: firstly, the wavelet energy of EEG is obtained by wavelet transform, and then the relative wavelet energy is calculated as the feature; Then the Logistic classifier is constructed to classify the features. (5) the function of on-line analysis and processing of EEG and the function of neural feedback are completed to realize the overall construction of the system. Finally, the system can not only analyze the saved EEG, but also process the EEG, in real time on line and convert the processing results into control signals. The virtual human model can be controlled and the user's MI status can be fed back. The results of on-line experiments show that the system can help the subjects to carry out MI, more effectively and improve the effect of rehabilitation training.
【学位授予单位】：杭州电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R318.0;TN911.7
本文编号：2444716
[Abstract]:Motor imagination (Motor Imagery,MI) training is a new method of rehabilitation training. In this paper, with the help of brain-computer interaction system, through the way of neural feedback, we explore how to enhance the effect of MI rehabilitation training. In this paper, we first propose a framework of brain-computer interaction system for MI rehabilitation training, and then study the MI EEG signal (Electroencephal ogra-m,EEG) eye artifact (Ocular Artifact,OA) removal algorithm, feature extraction algorithm and the programming implementation of classification algorithm. The corresponding functional modules are constructed to form a brain-computer interactive system for online MI rehabilitation training. The effects of MI training are compared with or without neural feedback, and the validity of the developed system is verified. The main contents of this paper can be divided into the following five aspects: (1) this paper introduces the basic concept of the system, the composition of the system and the research status at home and abroad, and analyzes the key technical problems in the current research of this kind of system. At the same time, we understand the structure of the human brain and the mechanism of EEG generation and the event-related desynchronization / synchronization (Event-Related Desynchr-onization/Synchronization,ERD/ERS) phenomenon of EEG in the process of MI. (2) the overall architecture of the system and the functions of each module are proposed, and the EEG acquisition scheme is designed, the experimental equipment and objects required for the collection are introduced, and the main points needing attention in the experiment are put forward, and the main points for attention in the experiment are put forward, and the main points to be paid attention to in the experiment are put forward. Finally, the specific data collected in the experiment are recorded. (3) an automatic method of removing OA is proposed: firstly, the horizontal and vertical ElectroOculogram,EOG signals are mixed into a new signal in a certain proportion. Together with EEG, an improved Independent component Analysis (Improved Independent Component Analysis,IICA) algorithm is used to obtain the independent components of each derived signal, and then the correlation coefficient is used to automatically identify and remove the independent components of the aliasing signals. Finally, the feature extraction and classification of pure EEG. (4) EEG obtained by inverse ICA transform are divided into two aspects: firstly, the wavelet energy of EEG is obtained by wavelet transform, and then the relative wavelet energy is calculated as the feature; Then the Logistic classifier is constructed to classify the features. (5) the function of on-line analysis and processing of EEG and the function of neural feedback are completed to realize the overall construction of the system. Finally, the system can not only analyze the saved EEG, but also process the EEG, in real time on line and convert the processing results into control signals. The virtual human model can be controlled and the user's MI status can be fed back. The results of on-line experiments show that the system can help the subjects to carry out MI, more effectively and improve the effect of rehabilitation training.
【学位授予单位】：杭州电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R318.0;TN911.7
【参考文献】
中国期刊全文数据库 前10条
1 刘珑;李胜;;基于快速独立分量分析的脑电波信号降噪[J];计算机测量与控制;2014年11期
2 罗志增;周镇定;周瑛;何海洋;;双树复小波特征在运动想象脑电识别中的应用[J];传感技术学报;2014年05期
3 李明爱;崔燕;杨金福;;脑电信号中眼电伪迹自动去除方法的研究[J];电子学报;2013年06期
4 刘小燮;毕胜;高小榕;杨志;闫铮;寇程;马林;高上凯;;基于运动想象的脑机交互康复训练新技术对脑卒中大脑可塑性影响[J];中国康复医学杂志;2013年02期
5 郑舟军;刘晓虹;张丽平;戎燕;龚戬芳;刘文琴;;脑卒中患者自我效能水平与其肢体功能康复进程的相关研究[J];中华护理杂志;2012年05期
6 赵大庆;王俊;;小波多重分形在脑电信号分析中的应用[J];中国生物医学工程学报;2010年05期
7 李明爱;王蕊;郝冬梅;;想象左右手运动的脑电特征提取及分类研究[J];中国生物医学工程学报;2009年02期
8 李明爱;刘净瑜;郝冬梅;;基于改进CSP算法的运动想象脑电信号识别方法[J];中国生物医学工程学报;2009年02期
9 谢松云;张振中;张伟平;赵海涛;;基于ICA的脑电信号去噪方法研究与应用[J];中国医学影像技术;2007年10期
10 杨帮华;颜国正;鄢波;;基于离散小波变换提取脑机接口中脑电特征[J];中国生物医学工程学报;2006年05期
中国硕士学位论文全文数据库 前1条
1 昌凤玲;多类运动想象脑电模式识别及其在电动轮椅控制上的应用[D];杭州电子科技大学;2014年
，本文编号：2444716
本文链接：https://www.wllwen.com/kejilunwen/xinxigongchenglunwen/2444716.html
