基于复合肢体想象动作的脑电特征识别技术研究
合集下载
- 1、下载文档前请自行甄别文档内容的完整性，平台不提供额外的编辑、内容补充、找答案等附加服务。
- 2、"仅部分预览"的文档,不可在线预览部分如存在完整性等问题,可反馈申请退款(可完整预览的文档不适用该条件!)。
- 3、如文档侵犯您的权益，请联系客服反馈,我们会尽快为您处理(人工客服工作时间：9:00-18:30)。
（保密的学位论文在解密后适用本授权说明）
学位论文作者签名：
签字日期：
年月日
导师签名： 签字日期： 年 月 日
摘要
近年来，脊髓损伤导致瘫痪的患者日渐增多，如何有效恢复和重建其肢 体运动功能成为康复工程领域一个亟待解决的难题。基于想象动作范式的脑 -机接口能够建立一种接近于原有损伤信道的人工神经通路，其内源性诱发响 应作业模式可直接输出使用者的主观运动意识。相对于简单肢体想象动作， 复合肢体想象动作反映了大脑各功能区相互整合和协同关系，更适于使用者 的日常生活习惯，同时又可满足控制信息的大指令集输出，对脊髓损伤后肢 体功能康复具有重要意义。
为进一步探索手足复合肢体想象动作范式下的诱发脑电特征，本文设计 了七类想象动作任务模式，包括左手、右手、双足、双手、左手右足、右手 左足和静息状态。为考察训练过程对任务作业效果的影响，研究中分别设置 了两个实验阶段：第一阶段 10 名受试者均未经过想象动作训练；第二阶段 10 名受试者均进行了一定时间的想象训练，其中 6 人参加过第一阶段实验。 本文首先利用时频图谱、脑地形图和功率谱密度曲线定性分析了各想象动作 模式诱发的事件相关去同步（ERD）特征的时频、能量特性以及脑区分布规 律。针对关键导联的 ERD 值分析结果表明复合手足协同想象动作诱发的 ERD 特征要强于简单的手部或足部想象动作，并且经过训练之后各个想象动作诱 发的 ERD 特征都得到了明显的增强。Fisher 可分性分析结果证实训练之后动 作模式间的差异性也得到了提高。文中尝试了 CSP、GECSP、sTRCSP 三种共 空间模式算法，采用支持向量机实现了对七类手足想象动作任务模式的分类 识别，分类结果表明，与其余两种共空间模式算法相比，sTRCSP 具有一定 的优越性，并且训练之后的脑电数据具有更高的分类正确率，其两分类最高 正确率可达 99%，七分类最高正确率可达 84%。本文所设计多分类手足复合 肢体想象动作范式及其脑电信号特征识别方法有望为想象动作型脑机接口 指令集的有效拓展，进而实现脊髓损伤者“思想变成行动”的理想康复目标提 供技术上的支持和帮助。
关键词：脊髓损伤，脑-机接口，复合想象动作，事件相关去同步，共空间模
式，支持向量机
ABSTRACT
In recent years, the number of paralyzed patients caused by spinal cord injury has increased. How to recover and reconstruct the limb movement function effectively has become a problem urgently to be solved in the field of rehabilitation engineering. The brain-computer interface based on motor imagery can develop an artificial neural pathway close to the original damaged neural channel, which operation mode based on endogenous induced response can output the user's subjective movement-related consciousness directly. With respect to motor imagery of simple limb movement, motor imagery of compound limb movement reflects mutual integration and synergies between different functional areas of the brain, which is more close to the user's customs in daily life and, at the same time, also can satisfy the requirements of multiple instructions output of control information. So it has great significance for the limb function rehabilitation after spinal cord injury.
独创性声明
本人声明所呈交的学位论文是本人在导师指导下进行的研究工作和取得的 研究成果，除了文中特别加以标注和致谢之处外，论文中不包含其他人已经发表
或撰写过的研究成果，也不包含为获得 天津大学 或其他教育机构的学位或证
书而使用过的材料。与我一同工作的同志对本研究所做的任何贡献均已在论文中 作了明确的说明并表示了谢意。
基于复合肢体想象动作的脑电特征 识别技术研究
Research on EEG Feature Recognition Technique Based on Compound-limbs
Motor Imagery
学科专业：生物医学工程 研 究 生：奕伟波 指导教师：明 东 教授
天津大学精密仪器与光电子工程学院 二零一二年十二月
In order to explore the EEG feature induced by motor imagery of compound limb movement combining hands with feet, seven kinds of motor imagery task mode were designed in this study, including left hand, right hand, both feet, both hands, left hand combined with right foot, right hand combined with left foot, and rest state. To investigate the effect of train to task operation, two experimental stages have been designed: 10 subjects in the first stage untrained for motor imagery, and 10 subjects in the second stage, including six from the first stage, trained for motor imagery a certain time. Firstly, time-frequency spectrum, brain topographic map and power spectral density curve were used to analyze the time-frequency, energy feature and region distribution of event related desynchronization (ERD) induced by each motor imagery pattern qualitatively. The results of ERD value analysis on key channels showed that the ERD feature induced by compound motor imagery of hand movement combined foot movement was stronger than that induced by simple motor imagery of hand or foot movement, and the ERD feature of every motor imagery mode was enhanced obviously after training. The results of Fisher separability analysis showed that the difference between every two modes was amplified after training. Three kinds of common spatial pattern algorithm including CSP，GECSP and sTRCSP were used for feature extraction. The classification of seven kinds of motor imagery task mode involving hands and feet was done by support vector machine (SVM). The
classification results showed that sTRCSP preformed better than the other two algorithms, and higher classification accuracy was obtained from the EEG signal after training than that before training. The highest accuracy of binary classification could reach 99%, and the highest accuracy of seven-class recognition could reach 84%. Multi-class compound motor imagery patterns combining hands with feet and its feature extraction methods from EEG signal are expected to provide techniqual support and helps to effectively expand the instructions of brain-computer interface based on motor imagery and further achieve the ideal goal of rehabilitation to "let thoughts into action" for the patients suffered from spinal cord injury.
目录第一章 绪论 ....源自.........................................................................................................1
1.1 研究背景 ......................................................................................................1 1.1.1 脊髓损伤现状........................................................................................1 1.1.2 脊髓损伤后的康复措施........................................................................4 1.1.3 脑-机接口 ..............................................................................................5 1.1.4 复合肢体想象动作电位研究意义及其发展现状................................8
学位论文作者签名：
签字日期：
年月日
学位论文版权使用授权书
本学位论文作者完全了解 天津大学 有关保留、使用学位论文的规定。 特授权 天津大学 可以将学位论文的全部或部分内容编入有关数据库进行检
索，并采用影印、缩印或扫描等复制手段保存、汇编以供查阅和借阅。同意学校 向国家有关部门或机构送交论文的复印件和磁盘。
KEY WORDS：spinal cord injury, brain-computer interface, compound motor
imagery, event-related desynchronization, common spatial pattern, support vector machine
学位论文作者签名：
签字日期：
年月日
导师签名： 签字日期： 年 月 日
摘要
近年来，脊髓损伤导致瘫痪的患者日渐增多，如何有效恢复和重建其肢 体运动功能成为康复工程领域一个亟待解决的难题。基于想象动作范式的脑 -机接口能够建立一种接近于原有损伤信道的人工神经通路，其内源性诱发响 应作业模式可直接输出使用者的主观运动意识。相对于简单肢体想象动作， 复合肢体想象动作反映了大脑各功能区相互整合和协同关系，更适于使用者 的日常生活习惯，同时又可满足控制信息的大指令集输出，对脊髓损伤后肢 体功能康复具有重要意义。
为进一步探索手足复合肢体想象动作范式下的诱发脑电特征，本文设计 了七类想象动作任务模式，包括左手、右手、双足、双手、左手右足、右手 左足和静息状态。为考察训练过程对任务作业效果的影响，研究中分别设置 了两个实验阶段：第一阶段 10 名受试者均未经过想象动作训练；第二阶段 10 名受试者均进行了一定时间的想象训练，其中 6 人参加过第一阶段实验。 本文首先利用时频图谱、脑地形图和功率谱密度曲线定性分析了各想象动作 模式诱发的事件相关去同步（ERD）特征的时频、能量特性以及脑区分布规 律。针对关键导联的 ERD 值分析结果表明复合手足协同想象动作诱发的 ERD 特征要强于简单的手部或足部想象动作，并且经过训练之后各个想象动作诱 发的 ERD 特征都得到了明显的增强。Fisher 可分性分析结果证实训练之后动 作模式间的差异性也得到了提高。文中尝试了 CSP、GECSP、sTRCSP 三种共 空间模式算法，采用支持向量机实现了对七类手足想象动作任务模式的分类 识别，分类结果表明，与其余两种共空间模式算法相比，sTRCSP 具有一定 的优越性，并且训练之后的脑电数据具有更高的分类正确率，其两分类最高 正确率可达 99%，七分类最高正确率可达 84%。本文所设计多分类手足复合 肢体想象动作范式及其脑电信号特征识别方法有望为想象动作型脑机接口 指令集的有效拓展，进而实现脊髓损伤者“思想变成行动”的理想康复目标提 供技术上的支持和帮助。
关键词：脊髓损伤，脑-机接口，复合想象动作，事件相关去同步，共空间模
式，支持向量机
ABSTRACT
In recent years, the number of paralyzed patients caused by spinal cord injury has increased. How to recover and reconstruct the limb movement function effectively has become a problem urgently to be solved in the field of rehabilitation engineering. The brain-computer interface based on motor imagery can develop an artificial neural pathway close to the original damaged neural channel, which operation mode based on endogenous induced response can output the user's subjective movement-related consciousness directly. With respect to motor imagery of simple limb movement, motor imagery of compound limb movement reflects mutual integration and synergies between different functional areas of the brain, which is more close to the user's customs in daily life and, at the same time, also can satisfy the requirements of multiple instructions output of control information. So it has great significance for the limb function rehabilitation after spinal cord injury.
独创性声明
本人声明所呈交的学位论文是本人在导师指导下进行的研究工作和取得的 研究成果，除了文中特别加以标注和致谢之处外，论文中不包含其他人已经发表
或撰写过的研究成果，也不包含为获得 天津大学 或其他教育机构的学位或证
书而使用过的材料。与我一同工作的同志对本研究所做的任何贡献均已在论文中 作了明确的说明并表示了谢意。
基于复合肢体想象动作的脑电特征 识别技术研究
Research on EEG Feature Recognition Technique Based on Compound-limbs
Motor Imagery
学科专业：生物医学工程 研 究 生：奕伟波 指导教师：明 东 教授
天津大学精密仪器与光电子工程学院 二零一二年十二月
In order to explore the EEG feature induced by motor imagery of compound limb movement combining hands with feet, seven kinds of motor imagery task mode were designed in this study, including left hand, right hand, both feet, both hands, left hand combined with right foot, right hand combined with left foot, and rest state. To investigate the effect of train to task operation, two experimental stages have been designed: 10 subjects in the first stage untrained for motor imagery, and 10 subjects in the second stage, including six from the first stage, trained for motor imagery a certain time. Firstly, time-frequency spectrum, brain topographic map and power spectral density curve were used to analyze the time-frequency, energy feature and region distribution of event related desynchronization (ERD) induced by each motor imagery pattern qualitatively. The results of ERD value analysis on key channels showed that the ERD feature induced by compound motor imagery of hand movement combined foot movement was stronger than that induced by simple motor imagery of hand or foot movement, and the ERD feature of every motor imagery mode was enhanced obviously after training. The results of Fisher separability analysis showed that the difference between every two modes was amplified after training. Three kinds of common spatial pattern algorithm including CSP，GECSP and sTRCSP were used for feature extraction. The classification of seven kinds of motor imagery task mode involving hands and feet was done by support vector machine (SVM). The
classification results showed that sTRCSP preformed better than the other two algorithms, and higher classification accuracy was obtained from the EEG signal after training than that before training. The highest accuracy of binary classification could reach 99%, and the highest accuracy of seven-class recognition could reach 84%. Multi-class compound motor imagery patterns combining hands with feet and its feature extraction methods from EEG signal are expected to provide techniqual support and helps to effectively expand the instructions of brain-computer interface based on motor imagery and further achieve the ideal goal of rehabilitation to "let thoughts into action" for the patients suffered from spinal cord injury.
目录第一章 绪论 ....源自.........................................................................................................1
1.1 研究背景 ......................................................................................................1 1.1.1 脊髓损伤现状........................................................................................1 1.1.2 脊髓损伤后的康复措施........................................................................4 1.1.3 脑-机接口 ..............................................................................................5 1.1.4 复合肢体想象动作电位研究意义及其发展现状................................8
学位论文作者签名：
签字日期：
年月日
学位论文版权使用授权书
本学位论文作者完全了解 天津大学 有关保留、使用学位论文的规定。 特授权 天津大学 可以将学位论文的全部或部分内容编入有关数据库进行检
索，并采用影印、缩印或扫描等复制手段保存、汇编以供查阅和借阅。同意学校 向国家有关部门或机构送交论文的复印件和磁盘。
KEY WORDS：spinal cord injury, brain-computer interface, compound motor
imagery, event-related desynchronization, common spatial pattern, support vector machine