基于形态成分分析的癫痫脑电棘波检测

Detection of Epileptic Spike Wave in EEG Signals Based on Morphological Component Analysis

本文提出一种基于形态成分分析(MCA)的癫痫脑电棘波检测方法,选用离散余弦变换(WT)作为MCA的字典来提取脑电图(EEG)的背景信号,选用db4小波变换(WT)作为MCA的字典来提取棘波信号。结果表明这种方法能够方便而有效地对EEG信号中的棘波进行检测,检测率达89.09%,正确率达90.71%。作为一种特征提取/分离算法,基于稀疏表示的MCA可以用来提取癫痫脑电棘波。

This paper proposed a morphological component analysis （MCA） method, which is based on sparse representation, to detect the spike wave in electroencephalogram （EEG） signals. It takes the advantage of MCA being able to extract the background waves and the spike waves from the EEG signals, respectively,as the dictionaries and chooses the discrete cosine transform （DCT） and the daubechies order 4 wavelet （db4） transformation as the dictionaries of MCA to detect the spike waves from the epileptic EEG. The experiment results showed that the MCA could detect epileptic spike waves in EEG signals very effectively, and it yielded high selectivity of 89.01% and sensitivity of 90.71%. As a feature extraction/decomposition algorithm, MCA can be used to extract the spike waves from EEG signals.