基于运动想象脑电特征分析的...激时长对皮层活跃度影响研究

版权申诉

matlab

毕业论文

24 浏览量 2023-10-27 20:30:35 上传评论收藏 3.8MB PDF 举报

资源推荐

资源详情

资源评论

杭州电子科技大学硕士学位论文

摘要

经颅直流电刺激(Transcranial Direct Current Stimulation, tDCS)和运动想象

(Motor Imagery, MI)联合疗法可以调控脑皮层活跃度，进而促进神经元重塑，加

速运动功能康复。然而在这种联合疗法中，tDCS 刺激时长参数设置的标准尚未

明确。针对该问题，本文首先设计实验方案，再搭建实验所需环境，并采集脑

电信号(Electroencephalogram, EEG)。通过小样本实验数据验证事件相关去同步

(Event-related Desynchronization, ERD)、样本熵(Sample Entropy, SampEn)、功率

谱密度(Power Spectral Density, PSD)以及瞬时能量(Instantaneous Energy, IE)四个

脑电特征在本次实验中作为脑皮层活跃度的测度的可行性。接着，本文通过自

研的基于 Web 的特征提取工具对本次实验采集到的 600 份数据进行批量处理。

最后结合配对样本 t 检验和单因素方差分析研究三种电流强度下的 tDCS 刺激时

长对脑运动皮层活跃度的影响。本文主要的研究内容如下：

(1) 构建实验环境。本文设计 3 种电流强度下(1、1.5 和 2 mA)的刺激时长实

验，每个实验包括 4 种不同时长(10、15、20 和 25 min)的阳极刺激，通过采集刺

激前后 C3 区的脑电信号分析不同刺激时长对运动皮层活跃度的影响。本文使用

脑电信号采集仪采集经颅直流电刺激前后的脑电信号，并通过经颅直流电刺激

仪对 C3 区施加阳极电刺激。通过构建的实验环境获取到 EEG 数据，再利用

EEGLab 和 Matlab 对实验数据进行预处理。最后得到的 EEG 信号可以为特征提

取提供数据基础。

(2) 特征选择研究。本文选择四个特征作为 tDCS 和 MI 联合实验中量化皮层

活跃度的测度，并基于 50 份小样本实验数据验证两个特征值在本次联合实验中

量化皮层活跃度的有效性。通过 Matlab 提取刺激前后 EEG 信号对应的特征值后

进行配对 t 检验分析。结果发现，刺激后的四个特征值均显著大于刺激前

(P<0.05)，满足阳极刺激增强皮层活跃度的规律，证实所选四个特征在本次联合

实验中作为测度的有效性。

(3) 批量特征提取。本文对本次实验所需的四种 EEG 信号特征提取功能研发

了基于 Web 应用的特征提取工具。将该工具对小样本实验数据提取到的特征值

与 Matlab 提取到的特征值进行比较，结果完全一致，证实本次工具的特征提取

功能有效且结果可信。最后，本文利用该工具对全部的实验数据进行特征提取，

为后续的统计分析做好准备。

(4) 阳极 tDCS 刺激时长对脑运动皮层活跃度的影响研究。对刺激时长实验

杭州电子科技大学硕士学位论文

III

ABSTRACT

The combination of Transcranial Direct Current Stimulation (tDCS) and Motor

Imagery (MI) can modulate the activity of cerebral cortex, thus promoting neuronal

remodeling and accelerating the recovery of motor function. However, the criteria for

setting tDCS stimulation duration parameters in this combination therapy are not yet

clear. To address this problem, this paper first designs the experimental protocol, then

builds the required environment for the experiment and acquires EEG signals

(Electroencephalogram, EEG). The feasibility of the four EEG features, event-related

desynchronization (ERD), sample entropy (SampEn), power spectral density (PSD) and

instantaneous energy (IE), as measures of cortical activity in this experiment was

verified using small sample data. Then, the 600 data collected in this experiment were

batch processed by a self-developed web-based feature extraction tool. Finally, paired-

sample t-test and one-way ANOVA were combined to investigate the effect of tDCS

stimulation duration on brain motor cortex activity at three current intensities. The main

research components of this paper are as follows:

(1) Setting up the experimental environment. We designed experiments with three

current intensities (1, 1.5 and 2 mA), each including four different anodal stimulation

durations (10, 15, 20 and 25 min), and analyzed the effects of different stimulation

durations on motor cortex activity by collecting EEG signals from C3 area before and

after stimulation. In this paper, the EEG signals before and after transcranial direct

current stimulation were acquired using an EEG signal acquisition device, and anodal

electrical stimulation was applied to the C3 area by transcranial direct current stimulator.

After the data were acquired, the experimental data were pre-processed by EEGLab and

Matlab. The final pure EEG signal obtained can provide the data basis for feature

extraction.

(2) Feature selection study. In this paper, four features were selected as the

measures for quantifying cortical activity in the joint experiment of tDCS and MI, and

the validity of two feature values for quantifying cortical activity in this joint

experiment was verified based on 50 small samples of experimental data. Paired t-test

analysis was performed after extracting the corresponding eigenvalues of EEG signals

before and after stimulation by Matlab. It was found that the four feature values after

stimulation were significantly larger than those before stimulation (P<0.05), which

杭州电子科技大学硕士学位论文

satisfied the law of cortical activity enhancement by anodal stimulation and confirmed

the validity of the selected four features as a measure in this joint experiment.

(3) Batch feature extraction. A web application-based feature extraction tool was

developed for the four EEG signal feature extraction functions required for this

experiment. The feature values extracted by this tool for small sample experimental

data are compared with those extracted by Matlab, and the results are in complete

agreement, confirming that the feature extraction function of this tool is effective and

the results are credible. Finally, this paper uses the tool to extract features from all the

experimental data to prepare for the subsequent statistical analysis.

(4) Study of the effect of anodal tDCS stimulation duration on brain motor cortex

activity. The EEG data from the stimulation duration experiments were subjected to

feature extraction, and the amount of change of each feature before and after stimulation

was calculated, and then the data on the amount of feature change were statistically

analyzed by one-way ANOVA, and it was concluded that the effect of stimulation

duration on cortical activity at different current intensities was consistent and non-linear.

Among them, there was a positive correlation between the stimulation duration

parameter and the enhancement effect of anodic tDCS on cortical activity in the range

of 10 to 20 min, while the enhancement effect of 25 min stimulation on cortical activity

diminished and appeared to inhibit cortical activity. Therefore, in the range of 10 to 20

min, the anodal tDCS stimulation duration parameter of 20 min is the optimal choice.

In addition, there is a positive correlation between current intensity and the

enhancement effect of anodal tDCS on cortical activity in the current range of 1 to 2

mA. Furthermore, in the stimulation parameter space consisting of a current range of 1

to 2 mA and a duration range of 10 to 20 min, anodal stimulation at 2 mA and 20 min

has the best effect on the enhancement of cortical activity.

With the above studies, the results of this paper are expected to provide meaningful

references for the stimulation duration settings of motor function rehabilitation therapy

using the combined tDCS and MI techniques, while the developed Web application-

based feature extraction tool can be helpful for subsequent studies.

Keywords: Transcranial direct current stimulation duration, Motor imagery,

Electroencephalogram, Feature extraction, Motor cortex activity

第 1 章绪论............................................................................................... 1

1.1 研究背景和意义 ............................................................................... 1

1.2 经颅直流电刺激概述 ....................................................................... 2

1.2.1 即刻效应 ..................................................................................... 3

1.2.2 后效应 ......................................................................................... 3

1.3 运动想象与 t DCS 联合方法概述 ................................................... 4

1.3.1 刺激时长概述 ............................................................................. 5

1.3.2 研究方法概述 ............................................................................. 7

1.4 主要研究内容与本文组织结构 ...................................................... 7

第 2 章 tDCS 与 MI 联合的实验环境构建 ........................................ 10

2.1 脑电信号概述 ................................................................................. 10

2.1.1 大脑结构及功能分区 ............................................................... 10

2.1.2 脑电信号分类与选取 ............................................................... 11

2.2 实验范式设计及实施 ..................................................................... 12

2.2.1 实验对象 ................................................................................... 12

2.2.2 实验范式设计 ........................................................................... 13

2.2.3 实验设备及参数设置 ............................................................... 13

2.2.4 实验实施 ................................................................................... 14

2.3 脑电信号预处理 ............................................................................. 15

2.4 本章小结 ......................................................................................... 20

第 3 章 EEG 特征提取 ............................................................................ 21

3.1 特征选取 ......................................................................................... 21

3.1.1 ERD/ERS ................................................................................... 21

3.1.2 功率谱密度 ............................................................................... 22

3.1.3 样本熵 ....................................................................................... 23

3.1.4 平均瞬时能量 ........................................................................... 24

3.2 基于 Matlab 验证所选特征作为测度的有效性 ......................... 25

3.2.1 Matlab 实现流程 ...................................................................... 26

3.2.2 数据结果分析 ........................................................................... 26

3.3 基于 Web 应用的特征提取系统需求分析 ................................... 28

3.4 系统设计 ......................................................................................... 29

3.4.1 前端设计 ................................................................................... 29

3.4.2 后端设计 ................................................................................... 30

3.4.3 前后端数据交互设计 ............................................................... 32

剩余83页未读，继续阅读

评论收藏

内容反馈

版权申诉

2201_75761617

粉丝: 20
资源: 7339

基于运动想象脑电特征分析的...激时长对皮层活跃度影响研究_毕业论文.pdf

最新资源

基于运动想象脑电特征分析的...激时长对皮层活跃度影响研究_毕业论文.pdf

基于深度学习框架的多类运动想象脑电分类研究.pdf

论文研究-异步BCI系统中运动想象诱发脑电特征识别.pdf

论文研究-基于能量熵的运动想象脑电信号分类.pdf

基于CSP与卷积神经网络算法的多类运动想象脑电信号分类.pdf

基于CSP与SVM算法的运动想象脑电信号分类.pdf

论文研究-基于复杂网络的脑电信号分析.pdf

免疫多域特征融合的多核学习SVM运动想象脑电信号分类.docx

论文研究-运动想象脑电信号特征提取与分类算法研究.pdf

基于ERD/ERS现象的运动想象脑电信号分类研究 (2015年)

基于微状态的静息态脑电信号分析.pdf

基于运动想象脑电控制的智能家居系统.pdf

单次运动想象脑电的特征提取和分类.

基于深度学习的精神分裂症脑电分析.pdf

论文研究-运动想象脑电信号识别研究.pdf

基于MATLAB+GUI的脑电网络分析系统_毕业论文.pdf

论文研究-基于等分符号化熵的情感脑电信号分析.pdf

基于深度学习的脑电控制系统.pdf

基于特征融合和粒子群优化算法的运动想象脑电信号识别方法.pdf

Matlab 基于支持向量机(SVM)的数据回归预测 SVM回归

Matlab 基于BP神经网络的数据分类预测 BP分类

LSTM时间序列神经网络预测MATLAB代码

ADRC控制器仿真 simulink 2017a版本

2022建模国赛代码(三天坚持不易) 包括K-meas算法、bp预测、回归预测,(python和matlab做的).zip

matlab2020b ubuntu.txt

基于蚁群算法的三维路径规划(matlab实现)

信号分选SDIF序列差直方图算法matlab仿真程序

基于智能优化算法的双层优化求解(matlab代码)

IEEE33节点配电网simulink模型

最新资源