基于小波变换和Teager能量算子的癫痫脑电自动分类

提出一种利用小波变换和能量算子对EEG进行预处理提取癫痫特征信号,进行近似熵估计,对脑电信号进行分类的新方法。首先利用小波分析将EEG信号进行4层分解分成多个子频带,对频率接近棘波的第1,2层小波系数计算非线性能量算子,再对能量算子进行近似熵估计,最后用SVM对EEG信号进行分类。结果表明,该方法对癫痫发作期EEG和正常的EEG分类效果比较理想。     

基于样本熵的睡眠呼吸暂停综合征脑电研究

目的:探讨睡眠呼吸暂停综合征(SAS)患者脑电的动力学特性,为SAS诊治提供依据。方法:基于脑电的非平稳和非线性特性,采用样本熵(Samp En)对6名SAS患者和6名健康人的睡眠脑电进行分析,研究SAS组和对照组在清醒、浅睡、深睡和快速眼动期(REM)的脑电变化及差异特性。结果:SAS患者和健康者睡眠脑电的样本熵变化有相同规律,即随着睡眠加深,其样本熵值均逐渐减小,但到REM期时,样本熵值又上升至觉醒期水平;与此同时,SAS组的样本熵值在各个睡眠阶段均低于健康组,两组间存在显著差异(P0.01);ROC曲线下面积达到0.858。结论:SAS病理状态对大脑神经活动影响显著,SAS组脑电样本熵值与对照组的显著差异为SAS研究及诊断提供新的方向和依据。     

基于BP网络的睡眠分阶方法及睡眠质量评估研究

我们利用不同睡眠期脑电复杂性特征与睡眠深度的关系及多道睡眠图功率谱特征,基于3层BP网络进行了睡眠自动分阶的研究,并提出了能部分反映睡眠质量的睡眠时间、浅睡时间、深睡时间、REM时间、觉睡比、醒转次数等参数。通过6例全睡眠监护实验说明,该方法可为睡眠质量的评价提供途径。     

75例颞叶癫癎患者动态EEG与过度换气EEG癎样放电规律探讨

目的：了解睡眠EEG与过度换气EEG的痢样放电规律。方法：评估75例颞叶癫痫患者动态EEG与过度换气EEG的痫样放电特点。结果：过度换气EEG痫样放电检出率明显低于浅睡期EEG，差异具有极显著意义（P〈0．01）；但与清醒期和深睡期EEG痫样放电检出率比较差异无显著意义（P〉0．05）。结论：颞叶癫痫患者浅睡期EEG痫样放电率明显高于过度换气EEG，对颞叶癫痫患者进行睡眠EEG检测，有助于提高痫样放电的检出率。     

低频光诱发脑电在浅睡期的样本熵复杂度分析

睡眠障碍患者通常表现为从浅睡期进入深睡期存在困难,分析浅睡期脑电波的变化对研究睡眠效率和睡眠质量至关重要。通过分析低频光刺激下睡眠过程中脑电波的复杂度值变化,研究人在浅睡期脑电波对光刺激的响应,进而探讨外部光刺激对睡眠过程中脑电波的影响。使用美国neuroscan型脑电图仪,采集10例志愿者的光刺激睡眠和正常睡眠的脑电数据。首先,利用时频分析,对睡眠过程中的脑电信号进行分期,获得浅睡期脑电信号;然后,使用小波包分解,获得该期脑电波的各频段分量（δ波、θ波、α波和纺锤波）;接着,采用样本熵算法,分别计算浅睡期脑电信号的复杂度以及各频段脑电波的复杂度;最后,对志愿者在光刺激（5 Hz）和正常睡眠下浅睡期脑电复杂度进行比较,研究光刺激对脑电复杂度的响应情况。结果显示：在低频光刺激下,浅睡期脑电波复杂度的均值为0514 15,明显低于正常睡眠复杂度的均值0589 23,在中央区和顶区有显著性差异（P<005）。研究表明,5 Hz光刺激可诱发浅睡期θ波的同步响应,增强脑电波的节律性,有助于更好地进入深度睡眠。     

颞叶癫痼患者长程数字化视频EEG与过度换气EEG痫样放电规律探讨

目的：了解长程数字化视频EEG与过度换气EEG的癎样放电规律.方法：评估52例颞叶癫癎患者长程数字化视频EEG与过度换气EEG的癎样放电特点.结果：过度换气EEG癎样放电检出率明显低于浅睡期EEG,差异具有极显著意义（P＜0.01）;但与清醒期和深睡期EEG癎样放电检出率比较差异无显著意义（P＞0.05）.结论：颞叶癫癎患者浅睡期EEG癎样放电率明显高于过度换气EEG,对颞叶癫癎患者进行睡眠EEG检测,有助于提高癎样放电的检出率.     

基于多分辨Tsallis熵的癫痫EEG分析

利用多分辨率Tsallis熵的方法,识别癫痫脑电中的棘波,把EEG信号分解成小波级数,在各个尺度上利用多分辨率Tsallis熵来检测棘波,和申农熵相比,多分辨率Tsallis熵能给出更详细的信息。通过对6位患者的EEG信号进行处理,发现这种方法能够比较准确的检测EEG中的棘波,具有较好的临床应用价值。     

低频光刺激在浅睡期的脑电响应

目的人的睡眠是有节律的,浅睡眠和深睡眠反复交替进行,分析浅睡期脑电波的变化对研究睡眠效率和睡眠质量至关重要。本文通过分析低频光刺激下睡眠过程中脑电波的变化,研究人在浅睡期脑电波对光刺激的响应,进而探讨外部光刺激对睡眠过程中脑电波的影响。方法利用美国Neuroscan型脑电图仪采集10例志愿者的光刺激睡眠和正常睡眠的脑电数据。首先,利用时频分析对睡眠过程中的脑电信号进行分期,获得浅睡期脑电信号,然后对分期后的脑电信号做傅里叶变换,获得各频段脑电波,并求其能量。最后比较并分析了志愿者在光刺激(5 Hz)和正常睡眠下浅睡期的脑电信号能量。结果在低频光刺激下,浅睡期脑电波的波能量明显高于正常睡眠,尤其在中央区和顶区增加明显。结论在光刺激下大脑皮质以更平稳的方式进入抑制状态,有助于更好地进入深睡期。     

用于睡眠精准分期的多模态生理时频特征提取网络

睡眠分期对临床疾病诊断以及睡眠质量评估至关重要。现有睡眠分期方法大多通过单通道或单模态信号,使用单分支深层卷积网络进行特征提取,这不仅阻碍了睡眠相关多样性特征的捕获,增加了计算代价,而且对睡眠分期的准确率也有一定的影响。为解决这一问题,本文提出一种端到端的用于睡眠精准分期的多模态生理时频特征提取网络（MTFF-Net）。首先,利用短时傅里叶变换（STFT）将包含脑电（EEG）、心电（ECG）、眼电（EOG）、肌电（EMG）的多模态生理信号转换为二维时频特征图;然后,使用多尺度EEG紧凑卷积网络（MsEEGNet）与双向门控循环（Bi-GRU）网络相结合的时频特征提取网络,捕获与睡眠特征波形相关的多尺度频谱特征以及与睡眠阶段转换相关的时序特征。根据美国睡眠医学学会（AASM）EEG睡眠分期判据,该模型在科英布拉大学系统与机器人研究所第三组子睡眠数据集（ISRUC-S3）上的五分类任务中取得了84.3%的准确率,其宏观F1分数（m-F1）的值为83.1%,科恩卡帕（Cohen’s Kappa）系数为79.8%。实验结果表明,本文所提模型实现了更高的分类准确率,推进了深度学习算法在辅助临床决...     

青年与老年昼夜心率信号多尺度化的基本尺度熵分析

目的:本研究采用多尺度化的基本尺度熵方法分析青年组与老年组,白天和黑夜的心率变异性信号。 方法:从PhysioBank数据库中,分别提取出青年人和老年人的心率变异性信号,利用多尺度化的基本尺度熵方法对数据进行比较分析。 结果:通过计算心率变异性信号的熵值,发现在白天清醒状态下的熵值比睡眠状态下的熵值高,但是昼夜间变化相似;老年组的多尺度化的基本尺度熵相对于青年组的熵值有所偏离,但大致走势一样,可见老年组的生理状态偏离了青年组的最佳生理状态。 结论:通过熵值的分析,揭示出清醒状态下,心脏系统的自适应性和稳定性较睡眠状态要强;青年与老年的熵值具有昼夜节律相关性;多尺度化的基本尺度熵方法可以用于区别青年与老年,白天与黑夜的心率变异性信号。     

利用小波级数检测睡眠EEG中的K-复合波

本文提出了一种利用小波级数检测睡眠脑电中Ｋ－复合波的方法。这种方法使用Ｄａｕｂｅｃｈｉｅｓ正交小波基,把ＥＥＧ信号分解成４个尺度上的小波级数,利用最大尺度的信号逼近检测Ｋ－复合波中的大慢波脉冲,然后,利用信号细节检测紧跟大慢波脉冲之后的梭形波。     

Recurrence quantification analysis of sleep electoencephalogram in sleep apnea syndrome in humans

The aim of this study is to elucidate whether the results of recurrence quantification analysis (RQA) of sleep EEGs in sleep apnea syndrome are valuable for analyzing sleep EEGs in sleep apnea syndrome. We investigated the ability of RQA to discriminate sleep stages and to characterize the different behaviors of sleep EEGs in sleep apnea syndrome. RQA was applied to EEG signals during sleep stages 1, 2, slow wave sleep (SWS), REM and the stage 'awake.' The sleep EEG signals were obtained from the MIT-BIH polysomnographic database. To examine the differences in the RQA measures for all sleep stages, one-way analysis of variance (ANOVA) and post hoc analysis were performed. From the results, all sleep stages could be distinctly discriminated by means of the RQA measure of %RATIO. We observed that stage 1 and REM had fewer recurrences, and that stage 2 was more autocorrelated than the other stages. The different dynamic behaviors of wakefulness and sleep EEG were also observed. Of significant interest was the observation that RQA was able to distinguish stage 1 from REM. In conclusion, we suggest that the information obtained from RQA of sleep EEGs in sleep apnea syndrome is valuable for its analysis, and that RQA constitutes a useful tool for analyzing sleep EEGs in subjects with sleep apnea syndrome.     

小波变换在睡眠呼吸暂停脑电分析中的应用

目的：利用小波转换方法对阻塞性睡眠呼吸暂停低通气综合征（OSAHS）患者脑电信号进行处理分析。方法：对人选142例鼾症患者分为对照组、轻度组、中度组和重度组。利用Morlet小波函数对脑电信号进行变换处理得到平均能量尺度图,通过对单位时间能量的计算得到脑电能量变异的两个指标divl和div2,并分别提取两个特征参数P1、P50和P10（Pdiv2≥1．0）、P20（Pdiv2≥1.0）。检验各个指标的正态分布性和各指标在各样本组间的差异性,以及各指标与临床指标的相关性。结果：发现变异指标的四个参数各组之间的检验结果均表现出相似的规律：重度组与对照、轻、中度组之间的差异均有显著意义（P〈0．05）;其它各组之间的差异无显著意义,但有显著性趋势（0．05〈P〈0．1）;各参数与多项睡眠图（PSG）检测指标的多元回归分析结果显示：与参数P1、P50、P10（Pdivz≥1.0）、P20（Pdiv2≥1.0）相关性最强的一个指标分别为呼吸暂停总时间／睡眠总时间（TAT／TST）、呼吸暂停指数、指端血氧饱和度（SpO2）％90％时间／TST、呼吸暂停指数（标准化偏回归系数分别为-0．369、-0．720、0．317、-0．602,P均〈0．05）。结论：脑电能量变异性的定量指标div1和div2的四个参数P1、P50和P10（Pdiv2≥1.0）、P20（Pdiv2≥1.0）表现出随病情加重而变化的趋势,并且对重度OSAHS有一定的区分意义;呼吸暂停是影响脑电能量变化的主要因素。     

EEG signals classification of epileptic patients via feature selection and voting criteria in intelligent method

Epileptic disease can be diagnosed by using intelligent methods on the Electroencephalograph (EEG) signals. In this paper, wavelet packet transform (WPT) was used in each of the frequency bands and wavelet coefficients were obtained, then the energy and entropy function was done on the wavelet coefficients and used as initial feature vectors. In the next step, eight and 15 features from 30 initial energy and entropy features were selected as the final features because their receiver operating characteristic (ROC) curve areas were higher than others. There were seven classifier inputs. These seven classifiers consisted of four artificial neural networks (ANN) with different structures, support vector machines (SVM), K-nearest neighbours (KNN) and a hybrid network. Each classifier was trained by 0.5, 0.8 and 0.9 EEG signals. After the training process, a fusion network based on a voting criteria was used to make the algorithm robust against the possible changes in each classifier and increase the classification accuracy. Finally, the algorithm was tested by other EEG signals. As a result, normal and epileptic classes were detected with total classification accuracy of 99–100%.     

基于脑电信号的睡眠分期算法研究

睡眠质量与人类健康息息相关,准确的睡眠质量监测对于帮助人们改善睡眠质量能够起到有效的监督作用。以MIT-BIH多导睡眠数据库slp01、slp02和slp04等3个样本的脑电信号为分析对象,采用sym7小波对其进行7层分解以去除高频细节信号,得到较为纯净的脑电信号。然后通过非线性符号动力学分析,去趋势波动分析以及频谱分析,分别提取符号熵指数,去趋势波动指数以及δ频带能量比等3个参数,对每个样本采用Kennard-Stone方法按照4〖DK〗∶1的比例建立校正集样本和预测集样本,并结合最小二乘支持向量机分类器进行样本训练拟合与分类识别。结果表明,3个特征参数与睡眠状态具有高度相关性,相关系数绝对值均高于0.83,并且确定了符号熵参数的嵌入维数为4,延迟常数为1,去趋势波动指数的分段区间为30～500,平均的睡眠分期正确率可达92.87%,比基于复杂度、近似熵等算法的分类正确率提高约5%。     

Removal of ocular artifacts from the REM sleep EEG.

The present report concerns the first study in which electrooculographic (EOG) contamination of electroencephalographic (EEG) recordings in rapid eye movement (REM) sleep is systematically investigated. Contamination of REM sleep EEG recordings in six subjects was evaluated in the frequency domain. REM-active and REM-quiet series were obtained for each subject. Transfer coefficients and power spectra of EOG and EEG indicated that (a) increases in transfer coefficients beyond 4.5 Hz are brought about by residual EEG in the EOG, and (b) EOG-EEG contamination in the delta band is most pronounced in frontal, intermediate in central and negligible in occipital leads. It was found that correction of the REM-active series resulted in significant (c) reductions in power, (d) increases in interhemispheric coherences and (e) reductions in degree of lateral asymmetry. These effects were largest for frontal leads, but still marked for central ones. The results are discussed in the light of previous findings concerning models of hemispheric functioning during REM sleep.     

A nonlinear approach to brain function: deterministic chaos and sleep EEG.

In order to perform a nonlinear dimensional analysis of the sleep electroencephalogram (EEG), we applied an algorithm proposed by Grassberger and Procaccia to calculate the correlation dimension D2 of different sleep stages under Lorazepam medication versus placebo. This correlation dimension characterizes the dynamics of the sleep EEG and it estimates the degrees of freedom of the signal under study. We demonstrate that slow-wave sleep depicts a much smaller dimensionality than light or rapid eye movement (REM) sleep, and that Lorazepam does not alter the EEG's dimensionality except in stage II and REM.