Application of multivariate empirical mode decomposition for seizure detection in EEG signals
Application of multivariate empirical mode decomposition for seizure detection in EEG signals
复制标题
DOI：
10.1109/iembs.2010.5626665
复制
发表时间：
2010-11
期刊：
影响因子：
--
通讯作者：
Naveed ur Rehman;Yili Xia;D. Mandic
中科院分区：
文献类型：
--
作者：
Naveed ur Rehman;Yili Xia;D. Mandic
我们提出了一种分析脑电图（EEG）信号的方法，它有可能区分发作和无癫痫的颅内EEG记录。这是通过分析多通道EEG记录中的常见频率分量，使用多元经验模式分解（MEMD）算法来实现的。信号的平均频率通过对所得的固有模式函数（IMF）应用希尔伯特-黄变换来计算。它已被证明，平均频率估计为发作和癫痫发作的EEG记录是统计上不同的，因此，可以作为一个测试统计量来区分这两类信号。对真实的脑电信号的仿真结果支持了分析，并证明了该方案的潜力。
We present a method for the analysis of electroencephalogram (EEG) signals which has the potential to distinguish between ictal and seizure-free intracranial EEG recordings. This is achieved by analyzing common frequency components in multichannel EEG recordings, using the multivariate empirical mode decomposition (MEMD) algorithm. The mean frequency of the signal is calculated by applying the Hilbert-Huang transform on the resulting intrinsic mode functions (IMFs). It has been shown that the mean frequency estimates for the ictal and seizure-free EEG recordings are statistically different, and hence, can serve as a test statistic to distinguish between the two classes of signals. Simulation results on real world EEG signals support the analysis and demonstrate the potential of the proposed scheme.