脉搏传播时间变异性谱分析的预处理方法研究

Preprocessing methods for spectrum analysis of pulse transit time variability

研究了脉搏传播时间(PTT)变异性谱分析的预处理过程,提出了以平均心动频率作为重采样率的均匀重采样方法和基于平均心动频率控制的小波去趋势(WFC,wavelet detrend based on frequency control)方法。试验结果表明:1)采用平均心动频率作为重采样率,可以保证在信号不失真的情况下达到均匀采样的目的;2)相比于平滑先验(SDA,smoothness prior approach)方法和经验模态分解(EMD,empirical mode decomposition)方法,WFC方法得出了较好的去除非平稳趋势效果,同时在处理大数据量序列时有更高的运行效率。

This paper discusses the preprocessing procedures before power spectrum analysis of pulse transit time（PTT） variability,including the uniform resampling and removing the non-stationary trend.A resampling method using the average heartbeat frequency as resampling rate and a wavelet detrend method based on frequency control（WFC） are proposed.By analyzing the PTT sequence,we obtain the following results：1） using the average heartbeat frequency as resampling rate can both reach the goal of uniform sampling and ensure that the original PTT sequence is not distorted;2） comparing with the methods of smoothness prior approach（SPA）and empirical mode decomposition（EMD）,the method of WFC can obtain a better detrending result.Meanwhile,the computing time of WFC is shorter than EMD and SPA when dealing with long sequence.