低采样率心电非稳态T波交替分析方法

ECG T-wave alternans detection using low sampling frequencies

心电T波交替现象与室性心动过速和室颤易发性有关,已成为预测心脏性猝死的一个无创的临床指标。目前T波交替变异性分析中常用平均功率谱分析方法,为弥补其不具备时间分辨率的不足,将Wigner-Ville时频分布(WVD)应用于T波交替变异性分析中,可反映T波交替随时间的动态变化。并提出基于相位谱修正的波形对准算法,降低了低采样率下心电信号误对准噪声对T波交替检测的影响。经仿真实验及MIT-BIH数据验证,该方法可使T波交替分析的应用范围从原先1kHz以上的高采样率心电数据扩展至200Hz及以上的低采样率心电数据。

Twave alternans (TWA) are associated with an increasing propensity to ventricular fibrillation and are an independent indicator of arrhythmic cardiac death. The spectral method commonly used to detect TWA does not provide temporal resolution. A new method based on the WignerVille distribution was developed to detect nonstationary TWA with both temporal and frequency resolution. A phase spectra alignment algorithm was also developed for lower sampling frequencies which reduce the influence caused by detector misalignment while detecting tiny TWA amplitudes. The ECG sampling frequency must generally be at least 1 kHz for TWA detection. The new method expands TWA detection to signals with sampling frequencies as low as 200 Hz. The method was tested using simulation experiments and the MITBIH clinical ECG database.