毕业论文
题 目： 基于独立分量算法的脑电信号分析
学生姓名： 学生学号： 系 别： 专 业： 届 别： 指导教师：
淮南师范学院2011届本科毕业论文
1
目 录
摘 要.............................................................................................................................. 3 1 前言 ........................................................................................................................... 5 2 脑电信号的研究概况和独立分量分析简介 ........................................................... 5 2.1 脑电信号研究方法概述 ..................................................................................... 5 2.2独立分量分析简介 .............................................................................................. 8 3 脑电信号的分类和获取方法 ................................................................................... 9 3.1自发脑电信号的分类和获取方式 .................................................................... 10 3.2 自发脑电信号的获取设备 ............................................................................... 11 4 独立分量分析的基本原理和典型算法 ................................................................. 13 4.1独立分量分析(ICA) ............................................................................................ 13 4.1.1 ICA的定义及线性模型 .............................................................................. 14 4.1.2 ICA判据 ................................................................................................... 16 4.1.3高斯性负熵判据 ......................................................................................... 17 4.1.4互信息判据 ................................................................................................. 18 4.2 ICA典型算法 ................................................................................................... 19 4.3 Infomax及扩展Infomax算法 ....................................................................... 19 4.4 FastICA算法 ....................................................................................................... 20 5 脑电信号伪迹消除 ................................................................................................. 21 5.1工频干扰的消除 ................................................................................................ 21 5.1.1无参考源的工频干扰消除 ......................................................................... 22 5.1.2 基于参考源的工频干扰消除 .................................................................... 23 5.2 基于ICA的工频干扰消除 .............................................................................. 23 5.3眼电伪迹的消除 ................................................................................................ 28 5．4 本章小结 ........................................................................................................ 28
2
基于独立分量算法的脑电信号分析
6 基于KICA方法的去噪及其实验 ............................................................................ 29 6.1 FICA算法去噪的具体步骤 ................................................................................ 29 6.2 仿真实验 ........................................................................................................... 30 6.3 实验结果分析 ................................................................................................... 30 7 总结和展望 ............................................................................................................. 33 致 谢............................................................................................................................ 35 参考文献...................................................................................................................... 35
淮南师范学院2011届本科毕业论文
3
基于独立分量算法的脑电信号分析
学生：王思汗 指导教师：李营 淮南师范学院电气信息工程系
摘 要：脑电是大脑神经元突触后电位的综合，是大脑电活动产生的电场经容积导体(由皮层、颅骨、脑膜及头皮构成)传导后在头皮上的电位分布，分为自发脑电(electro encephalon graph，EEG)和诱发电位(evoked potential，EP)两种。脑电在临床诊断、军事医学、航天医学、生理学和生物学研究中都具有重要的意义，所以脑电信号的提取一直是神经科学领域的重要课题。
独立分量分析(ICA)方法是最近几年发展起来的一种新的统计方法。ICA方法是基于信号高阶统计特性的分析方法，经ICA方法分解出的各信号分量之间是相互独立的。正是因为这一特点，使ICA在信号处理领域受到了广泛的关注。ICA方法的发展十分迅速，国内外的众多研究人员都致力于研究新的算法，应用于脑电信号的噪声分离之中。
本文利用FastICA算法的特点，尝试将其应用到思维作业脑电信号中，发现其也能得到较好的结果，证明其在脑电信号分析方面的有效性。 关键词：EEG信号；独立分量分析：FastICA算法
Algorithm based on independent component analysis
of EEG
Student:Wang Sihan Supervisor: Li Ying
Huainan normal college electrical information engineering
4
基于独立分量算法的脑电信号分析
Abstract：EEG is brain neurons postsynaptic potential comprehensive, is the brain electrical activity in the electric field generated by volume conductor (by cortex, skull, dural and scalp constitute) conduction in scalp after the potential distribution, divided into spontaneous electro electroencephalography (EEG) and encephalon graph, evoked potentials（EP) two kinds. EEG in clinical diagnosis, military medical, aerospace medicine and physiology and central biology study has an important meaning, so the extraction of EEG is always an important issue in the field of neuroscience.
Independent component analysis (ICA) method is developed in recent years, a new statistical method. ICA method is based on the signal analyzing the characteristic of higher order statistics method, a method by ICA each signal decomposition between weight were independent of each other. It is for this characteristic, make ICA in signal processing field is extensive attention. The main method is developing very rapidly, and domestic and international numerous researchers are working to research the new algorithm, applied to the noise separation of EEG.
Based on the characteristics of FastICA algorithm, try its application to thinking in homework eeg signals, finds its also can get good results in, to prove its effectiveness of EEG analysis.
Key words：EEG signal；ICA algorithm with reference signals；FastICA algorithm