【摘要】 对于癫痫患者,癫痫易发作期(seizure vulnerable period,SVP)的检测有助于预测癫痫发作并采取相应的治疗和抑制措施。本文采用量化脑电分析方法研究癫痫易发作期的临床诊断价值。对经过带通滤波预处理的量化脑电信号分析得到反映大脑状态的线性特征量(过零率和谱距)和非线性特征量(最大Lyapunov指数、复杂度、相位同步和互信息),以及比较不同导联之间的特征差异,并将它们应用于慢性颞叶癫痫大鼠模型长时程连续脑电数据和癫痫病人临床脑电数据的分析。癫痫易发作期与正常状态期间对比,分析单个导联特征量变化趋势得到最大Lyapunov指数、复杂度和过零率的值下降,而相位同步、互信息和谱距的值上升;选择两个合适的导联,分析不同导联之间的特征差异是最大Lyapunov指数、复杂度、相位同步、互信息和过零率差异减小或趋于一致,而谱距的差异增大。结果表明量化脑电特征量可以反映癫痫易发作期相对正常状态期间的变化,并且有些特征变化在每次集中发作的第一次发作之前较长的一段时间内就开始了,即进入癫痫易发作期的时刻相对第一次发作时刻在时间上具有一个提前量,这个时间提前量为临床处置提供了前提。更多还原

【Abstract】 The detection of the seizure vulnerable period(SVP) can be helpful to predict and interrupt an imminent seizure of epilepsy patients.Quantitative Electroencephalogram(qEEG) methods play a significant role in EEG based clinical diagnosis and study.By qEEG methods,linear and nonlinear features were calculated from filted EEG data,and then difference of a common feature between channels was also analyzed.In comparison with the normal period,there was a visible change of the features during the SVP of epileptic patients and Temporal lobe epilepsy rat model.The maximum Lyapunov exponent,complexity and zero passed rates decreased but the phrase synchronization,mutual information and spectral distance increased.The difference of features as the maximum Lyapunov exponent,complexity,phrase synchronization,mutual information and zero passed rates reduced,but the difference of spectral distance became larger.It was also observed that some changes occur a certain long time before the first seizure,and they can help doctors and patients to adopt treatment.更多还原