相位同步性分析方法在癫痫研究中的应用

同步振荡是神经系统中普遍存在的一种现象,是理解大脑信息处理机制的重要切入点,不同频段的同步现象能够编码不同的刺激信息.癫痫是由大脑整体或局部整合的不足或异常造成的,同步振荡现象在癫痫的发生过程中扮演了重要的角色.本文介绍了应用希尔伯特变换和小波变换提取脑电的瞬时相位方法,介绍了脑电相位同步性的相干分析、相位同步指数和香侬熵法,以及这些方法在癫痫同步放电研究中的应用,为提取癫痫患者大脑功能的信息和更深入地探索癫痫发作的机理提供了理论依据.     

脑电同步性分析在癫痫机制研究中的应用

脑电相位同步性是研究癫痫超同步放电机制的方向之一。介绍了应用Hilbert变换提取脑电的瞬时相位的方法,分析脑电相位同步性的互相关法、互信息法和同步指数法,以及脑电信号的小波变换,综述了以上方法在癫痫发作的超同步放电机制研究中的应用。     

脑电同步性分析在癫痫机制研究中的应用

脑电相位同步性是研究癫痫超同步放电机制的方向之一。介绍了应用Hilbert变换提取脑电的瞬时相位的方法，分析脑电相位同步性的互相关法、互信息法和同步指数法，以及脑电信号的小波变换，综述了以上方法在癫痫发作的超同步放电机制研究中的应用。     

想象左右手运动的脑电特征提取及分类研究

针对想象运动的脑机接口(BCI)系统存在分类准确率低、抗干扰能力差等不足,提出一种将离散小波变换(DWT)和BP神经网络相结合的脑电识别方法(DWT-BP法).通过计算想象左、右手运动的C3、C4的平均功率,合理确定时间窗设置,对时间窗内的平均功率信号进行离散小波变换,并选取尺度6上的逼近系数A6的组合信号作为脑电信号特征,以BP神经网络为分类器实现对脑电观测数据的分析.实验结果表明,DWT-BP方法能够较准确地提取脑电信号的本质特征,具有较好的抗干扰能力和分类性能,以及识别运动想象脑电信号的有效性,同时为实现运动想象在线BCI系统打下基础.     

一种基于皮层成像的自动眼电伪迹去除方法

脑电信号十分微弱,并且特别容易受到眼电的干扰.这些干扰给阅读和分析脑电信号带来了很大的困难,因此自动消除眼电对脑电的干扰一直是研究人员重视的问题.本研究提出一种基于皮层成像的自动眼电伪迹去除方法,对于已经完成滤波的脑电数据段,通过设立阈值的方法识别伪迹,利用基于相关系数的眼电伪迹识别算法标记眼电伪迹数据段,然后通过结合脑电信号时空信息的、基于皮层成像技术的眼电伪迹处理方法(CAST),处理已经标记好的眼电伪迹数据段,并通过真实的事件相关电位数据验证了方法的有效性.验证结果表明,此方法能够实现眼电伪迹的自动识别和去除,去除伪迹后的信号与原始无眼电伪迹的标准信号之间的相关系数为0.953 7±0.042 3.     

胃动力信息检测系统的设计与实现

将胃电与胃机械活动相关的胃阻抗信息作为诊断依据,建立新的、更准确的胃动力评价方法.基于阻抗法设计了获取胃动力信息的系统,通过小波变换技术分离与胃电同步的胃蠕动信息.结果表明系统通过了调试,实现设计功能,采集到了阻抗和同步胃电信号.小波变换提取的信号排除了干扰,获得了胃蠕动信息.基于阻抗技术的方法可检测胃蠕动状态,联合同步胃电分析技术为临床检测提供了新方法.     

基于小波变换和独立分量分析相结合的癫痫脑电分析

临床上分析癫痫脑电信号非常重要。由于临床记录的癫痫脑电信号中含有大量的伪迹干扰,特别是肌电伪迹,所采集的脑电信号无法正确反映大脑的生理及病理状况。本研究利用小波变换的多分辨率特性和独立分量分析(ICA)的盲源分离特性,把用连续小波变换分解的脑电子带信号作为ICA输入,经ICA分离后,有效地消除了癫痫脑电中的肌电伪迹,并分离出了癫痫样特征波,效果理想。     

意识任务识别中事件相关脑电的时频域信息熵表达

事件相关脑电的量化表征对于研究意识任务识别和认识大脑思维机制具有重要意义.本研究对左右手想象意识任务的脑电信号进行小波包分析,提取出时频域信息熵用来表征事件相关脑电的变化;进而,分析了时频域信息熵特征的事件相关去同步/同步的变化时程,应用互信息评价时频域信息熵对事件相关脑电的表征能力,将相同步理论应用于导联间的脑电信号分析;设计时变线性分类器实现左右手想象运动意识任务识别,获得了满意的结果,最小分类错误率为9%.结果表明,时频域信息熵与频带能量具有一致的变化时程;时频域信息熵具有比频带能量更好的分离性,是事件相关去同步,同步的一个敏感的量化参数;时频域信息熵结合相同步相干性指数.能够提供更多反映大脑意识任务的状态信息.     

癫痫脑电信号的奇异性研究

基于小波变换模极大值在多尺度上的变化,研究了癫痫脑电的奇异性,并用Lipschitz指数来表征.提出了一种高阶统计的方法来研究癫痫脑电的高阶奇异谱特征,并和健康脑电进行比较.实验结果表明,癫痫脑电的Lipschitz指数和高阶奇异谱与健康脑电相比存在明显的差异,说明该方法对研究脑电是有效的.     

基于离散小波变换提取脑机接口中脑电特征

在脑机接口中，针对脑电特征提取利用单一种类信息、使用数据量大、分类性能较差等缺点，提出一种新颖的基于离散小波变换的方法。分析了小波变换特征提取的特点和特征表示方式，用Daubechies类db4小波函数对脑电信号进行6层分解，抽取小波变换各子带关键的部分逼近系数、小波系数、小波子带系数均值组成特征向量。以分类正确率为指标检验了提取特征的性能。实验结果表明，这种方法能够利用少量数据提取脑电信号本质特征，具有较高的分类性能，为利用脑电识别人的不同意图提供了快速而有效的手段。     

视觉听觉同时刺激模式下ERP的同步性研究

目的：研究视觉、听觉脑区在认知过程中的同步性。方法：设计了视觉、听觉同时刺激模式下的脑电实验，采用128道高密度脑电采集系统，记录了14位年龄在18～29岁之间的在校男性大学生的诱发脑电信号数据，并运用希尔伯特(Hilbert)相位同步算法对视觉、听觉区域的事件相关电位(ERP)进行同步量化。结果：在视听觉同时利用模式下，视觉脑区(枕叶)和听觉脑区(颞叶)之间的同步指数明显大于它们与其他脑区间的同步指数。结论：人在感知和认知事物时，相关的脑区间自动产生了神经活动的同步化。     

基于分段Prony建模的脑电相位同步分析方法

本研究提出采用一种分段的Prony方法(PPM),将非平稳EEG信号分解为幅度为指数上升或下降的正弦波,从而获得各频率成分的幅度、频率和初始相位,并计算两通道相位差;用香侬熵来衡量两路信号的相位同步程度.首先对该方法的频率、相位分辨率和检测同步程度的效果等进行仿真研究,结果显示了该方法具有高的频率和相位分辨率.最后将该方法用于实际的三组脑电信号,得到的结果与经典方法一致.将该方法与希尔伯特变换方法进行了比较,说明该方法具有较好的抗噪性能,可作为非平稳信号的相位同步检测的有效工具.     

Noninvasive Localization of Epileptic Sites from Stable Phase Synchronization Patterns on Different Days Derived from Short Duration Interictal Scalp dEEG

The stochastic behavior of the phase synchronization index (SI) on different days during a hospital stay of epileptic patients was studied for noninvasive localization of the epileptogenic areas from high density (256 channel) scalp EEG recordings. The study was performed on three subjects with interictal EEG data on different days. The seizure areas were localized with subdural recordings with an 8 × 8 grid electrode array. The study was performed in low gamma (30–50 Hz) band with short duration (0–180 s), seizure-free and spike-free scalp EEG data. A detrended fluctuation analysis was used to find the averaged stochastic fluctuations in the SI. The phase synchronization was computed after taking Hilbert transform of the EEG data. Contour plots were constructed with 20 s time–frames using a montage of the layout of 256 electrode positions. It was found that the stochastic behavior of the SI was higher in epileptogenic areas on different days for each subject. Also, a stable higher pattern of SI emerged after 60–100 s in the epileptogenic areas. These findings suggest that it is possible to localize the epileptogenic areas from the short duration (60–100 s), seizure-free and spike-free high density scalp EEG recordings.     

Estimation of Delay Times in Biological Systems

The problem of delay time estimation in biological systems is addressed with the focus on practical applicability of methods. Four delay time estimators are described: a cross correlation method and three increasingly sophisticated interpretations of the phase spectrum, ranging from a pointwise interpretation of the phase spectrum in terms of a delay to a Hilbert transform method. The four methods are compared through simulation studies showing that, in general, the Hilbert transform method performs best. The methods are then used to estimate delay times in three physiological systems: vestibular stimulation, cerebral autoregulation, and human orthostatic tremor. In all three cases, the Hilbert transform method yields the best results, leading in some cases to physiologically more sensible interpretations of experiments than the other methods. © 2003 Biomedical Engineering Society. PAC2003: 8710+e, 8780Tq     

Visually evoked phase synchronization changes of alpha rhythm in migraine: correlations with clinical features.

OBJECTIVE: This study aimed to compute phase synchronization of the alpha band from a multichannel electroencephalogram (EEG) recorded under repetitive flash stimulation from migraine patients without aura. This allowed examination of ongoing EEG activity during visual stimulation in the pain-free phase of migraine. METHODS: Flash stimuli at frequencies of 3, 6, 9, 12, 15, 18, 21, 24, and 27 Hz were delivered to 15 migraine patients without aura and 15 controls, with the EEG recorded from 18 scalp electrodes, referred to the linked earlobes. The EEG signals were filtered in the alpha (7.5-13 Hz) band. For all stimulus frequencies that we evaluated, the phase synchronization index was based on the Hilbert transformation. RESULTS: Phase synchronization separated the patients and controls for the 9, 24 and 27 Hz stimulus frequencies; hyper phase synchronization was observed in patients, whereas healthy subjects were characterized by a reduced phase synchronization. These differences were found in all regions of the scalp. CONCLUSIONS: During migraine, the brain synchronizes to the idling rhythm of the visual areas under certain photic stimulations; in normal subjects however, brain regions involved in the processing of sensory information demonstrate desynchronized activity. Hypersynchronization of the alpha rhythm may suggest a state of cortical hypoexcitability during the interictal phase of migraine. SIGNIFICANCE: The employment of non-linear EEG analysis may identify subtle functional changes in the migraine brain.     

基于静息态fMRI相位同步的ADHD儿童脑网络研究

探索相位同步和复杂网络方法在注意缺陷多动障碍(ADHD)脑网络机制研究中的应用, 选取135例ADHD患者和102例正常对照作为研究对象。以这237例被试的功能磁共振图像时间序列作为研究数据, 利用相位同步分析方法获得脑区间的连接关系, 并在此基础上构建脑网络。然后, 利用复杂网络的局部效率指标评估静息态脑功能, 并采用多元线性回归和方差分析等统计方法, 分析ADHD患者和正常对照在静息态下脑区的局部效率可能存在的差异。结果表明, ADHD患者与正常对照在年龄、性别、量表分值(注意力和自制力)、3种智商值(语言智商、操作智商和总智商)等方面均无统计学差异, 在诊断和头动参数上有显著差异(P<0.05, 校正后)。诊断方面发现, 11个局部效率正常对照组与ADHD组具有统计学差异的脑区(P<0.05), 其中主要的脑区为左侧尾状核(0.118±0.317 vs 278±0.433)、丘脑(0.345±0.425 vs 0.541±0.435)、颞横回(0.467±0.476 vs 0.654±0.444)和右侧背外侧额上回(0.536±0.401 vs 0.681±0.333)、额中回(0.505±0.377vs 0.641±0.331)、尾状核(0.144±0.329 vs 0.298±0.423)。在静息态下, ADHD患者和正常对照在左侧中央前回、尾状核、丘脑等脑区的局部效率差异可能与患者尾状核、丘脑等特定脑区的功能异常有关, 也可能与患者注意和执行有关的神经网络损伤有关。     

Assessment of Event-Related EEG Power After Single-Pulse TMS in Unresponsive Wakefulness Syndrome and Minimally Conscious State Patients

In patients without a behavioral response, non-invasive techniques and new methods of data analysis can complement existing diagnostic tools by providing a method for detecting covert signs of residual cognitive function and awareness. The aim of this study was to investigate the brain oscillatory activities synchronized by single-pulse transcranial magnetic stimulation (TMS) delivered over the primary motor area in the time–frequency domain in patients with the unresponsive wakefulness syndrome or in a minimally conscious state as compared to healthy controls. A time–frequency analysis based on the wavelet transform was used to characterize rapid modifications of oscillatory EEG rhythms induced by TMS in patients as compared to healthy controls. The pattern of EEG changes in the patients differed from that of healthy controls. In the controls there was an early synchronization of slow waves immediately followed by a desynchronization of alpha and beta frequency bands over the frontal and centro-parietal electrodes, whereas an opposite early synchronization, particularly over motor areas for alpha and beta and over the frontal and parietal electrodes for beta power, was seen in the patients. In addition, no relevant modification in slow rhythms (delta and theta) after TMS was noted in patients. The clinical impact of these findings could be relevant in neurorehabilitation settings for increasing the awareness of these patients and defining new treatment procedures.     

Phase synchronization between alpha and beta oscillations in the human electroencephalogram

Coordination of neuronal oscillations generated at different frequencies has been hypothesized to be an important feature of integrative brain functions. The present study aimed at the evaluation of the cross-frequency phase synchronization between electroencephalographic alpha and beta oscillations. The amplitude and phase information were extracted from electroencephalograms recorded in 176 healthy human subjects using an analytic signal approach based on the Hilbert transform. The results reliably demonstrated the presence of phase synchronization between alpha and beta oscillations, with a maximum in the occipito-parietal areas. The phase difference between alpha and beta oscillations showed characteristic peaks at about 2 and -1 radians, which were common for many subjects and electrodes. A specific phase difference might reflect similarity in the organization and interconnections of the networks generating alpha and beta oscillations across the entire cortex. Beta oscillations, which are phase-locked to alpha oscillations--alpha-synchronous beta oscillations--were largest in the occipito-parietal area with a second smaller maximum in the frontal area, thus demonstrating a topography, which was different from the conventional alpha and beta oscillations. The strength of the alpha-synchronous beta oscillations was not exclusively defined by the amplitude of the alpha rhythm indicating that they represent a distinct feature of the spontaneous electroencephalogram, which allows for a refined discrimination of the dynamics of beta oscillations.