A comparison of various linear and non-linear signal processing techniques to separate uterine EMG records of term and pre-term delivery groups

详细信息

下载全文

推荐本文 |

• 作者：G. Fele-沤or啪 (1) polz@fri.uni-lj.si
  G. Kav拧ek (2) gorazd.kavsek@obgyn.mf.uni-lj.si
  沤. Novak-Antoli膷 (2) ziva.novak@guest.arnes.si
  F. Jager (1) franc.jager@fri.uni-lj.si
• 关键词：Uterine EMG ; Electrohysterogram ; Pre ; term delivery prediction ; Linear signal processing techniques ; Non ; linear signal processing techniques
• 刊名：Medical & Biological Engineering & Computing
• 出版时间：September 2008
• 出版年：2008
• 期刊代码：114_1741-0444
• 类别：et
• 卷：46
• 期：9
• 页码：911-922
• 数据来源：sp

摘要

Various linear and non-linear signal-processing techniques were applied to three-channel uterine EMG records to separate term and pre-term deliveries. The linear techniques were root mean square value, peak and median frequency of the signal power spectrum and autocorrelation zero crossing; while the selected non-linear techniques were estimation of the maximal Lyapunov exponent, correlation dimension and calculating sample entropy. In total, 300 records were grouped into four groups according to the time of recording (before or after the 26th week of gestation) and according to the total length of gestation (term delivery records—pregnancy duration ≥37 weeks and pre-term delivery records—pregnancy duration <37 weeks). The following preprocessing band-pass Butterworth filters were tested: 0.08–4, 0.3–4, and 0.3–3 Hz. With the 0.3–3 Hz filter, the median frequency indicated a statistical difference between those term and pre-term delivery records recorded before the 26th week (p = 0.03), and between all term and all pre-term delivery records (p = 0.012). With the same filter, the sample entropy indicated statistical differences between those term and pre-term delivery records recorded before the 26th week (p = 0.035), and between all term and all pre-term delivery records (p = 0.011). Both techniques also showed noticeable differences between term delivery records recorded before and after the 26th week (p ≤ 0.001).