Time-Variant, Frequency-Selective, Linear and Nonlinear Analysis of Heart Rate Variability in Children With Temporal Lobe Epilepsy
Time-Variant, Frequency-Selective, Linear and Nonlinear Analysis of Heart Rate Variability in Children With Temporal Lobe Epilepsy
复制标题
DOI：
10.1109/tbme.2014.2307481
复制
发表时间：
2014-02
影响因子：
4.6
通讯作者：
K. Schiecke;M. Wacker;D. Piper;F. Benninger;M. Feucht;H. Witte
中科院分区：
文献类型：
--
作者：
K. Schiecke;M. Wacker;D. Piper;F. Benninger;M. Feucht;H. Witte
我们研究的主要目的是证明时变、频率选择、线性和非线性分析方法的协调组合可以有益地用于癫痫患者心率变异性（HRV）的分析，以揭示关于即将发生的癫痫发作的先兆信息，并提供更多关于导致自主神经系统变化的机制的信息。探索是为了证明，结合的方法获得新的见解，在特定的短期模式，心率变异性发作前，发作和发作后时期的癫痫儿童。采用连续Morlet小波变换，通过频谱分析、锁相分析、频带功率分析和二次相位耦合分析，研究心率变异性的时频特性。这些结果是由来自信号自适应方法的时变特性完成的。高级经验模式分解被用来分离出某些HRV分量，特别是血压相关的Mayer波（10.1 Hz）和呼吸性窦性心律失常（10.3 Hz）。他们的时变非线性预测性进行了分析，使用本地估计的最大李雅普诺夫指数（点预测误差）。在癫痫发作前约80-100 s，可以观察到两种HRV分量的时间和协调。发现了更高程度的同步性，并且由此发现了HRV的更高的可预测性。所有调查的线性和非线性分析有助于这些结果的具体重要性。
The major aim of our study is to demonstrate that a concerted combination of time-variant, frequency-selective, linear and nonlinear analysis approaches can be beneficially used for the analysis of heart rate variability (HRV) in epileptic patients to reveal premonitory information regarding an imminent seizure and to provide more information on the mechanisms leading to changes of the autonomic nervous system. The quest is to demonstrate that the combined approach gains new insights into specific short-term patterns in HRV during preictal, ictal, and postictal periods in epileptic children. The continuous Morlet-wavelet transform was used to explore the time-frequency characteristics of the HRV using spectrogram, phase-locking, band-power and quadratic phase coupling analyses. These results are completed by time-variant characteristics derived from a signal-adaptive approach. Advanced empirical mode decomposition was utilized to separate out certain HRV components, in particular blood-pressure-related Mayer waves (≈0.1 Hz) and respiratory sinus arrhythmia (≈0.3 Hz). Their time-variant nonlinear predictability was analyzed using local estimations of the largest Lyapunov exponent (point prediction error). Approximately 80-100 s before the seizure onset timing and coordination of both HRV components can be observed. A higher degree of synchronization is found and with it a higher predictability of the HRV. All investigated linear and nonlinear analyses contribute with a specific importance to these results.