Gonçalves Hernâni, Rocha Ana Paula, Ayres-de-Campos Diogo, Bernardes João
Departamento de Matemática Aplicada, Faculdade de Ciências da Universidade do Porto, Portugal.
Physiol Meas. 2006 Mar;27(3):307-19. doi: 10.1088/0967-3334/27/3/008. Epub 2006 Feb 6.
The effect of foetal heart rate (FHR) acquisition mode on linear and nonlinear parameters is still largely unknown. In 33 normal labouring women, FHR signals were acquired simultaneously by an external ultrasound sensor applied to the maternal abdomen and an internal scalp electrode, in the minutes preceding delivery. For each case, the initial and final 5, 10 and 20 min segments were analysed, considering FHR signals at a frequency of 4 Hz (the frequency at which they are transmitted by the majority of commercialized foetal monitors). Several time and frequency domain linear and nonlinear FHR indices were computed in these segments, namely mean FHR, very low frequency (VLF), low frequency (LF), high frequency (HF), approximate entropy (ApEn) and sample entropy (SampEn). Parametric confidence intervals, statistical tests and correlation coefficients were calculated in order to evaluate the effect of internal versus external FHR monitoring modes on the considered indices. The whole evaluation was repeated using FHR signals at a frequency of 2 Hz. Most time domain linear indices were similar with external and internal monitoring in the initial and final segments of the tracings. However, linear frequency domain indices were poorly correlated in the final segments and had significantly different mean values in the initial segments. Nonlinear indices were significantly different in both initial and final segments. The correlation between 4 and 2 Hz sampled parameters was high for both linear and nonlinear indices (most correlation coefficient values ranging between 0.95 and 1) but nonlinear index values were significantly higher at 2 Hz. In conclusion, the mode used to acquire FHR signals and the sampling rate employed can significantly affect most FHR indices.
胎儿心率（FHR）采集模式对线性和非线性参数的影响在很大程度上仍不为人知。在33名正常分娩的妇女中，在分娩前几分钟，通过应用于产妇腹部的外部超声传感器和内部头皮电极同时采集FHR信号。对于每个病例，分析了最初和最后的5、10和20分钟片段，考虑了频率为4Hz的FHR信号（大多数商业化胎儿监护仪传输这些信号的频率）。在这些片段中计算了几个时域和频域的线性和非线性FHR指标，即平均FHR、极低频（VLF）、低频（LF）、高频（HF）、近似熵（ApEn）和样本熵（SampEn）。计算了参数置信区间、统计检验和相关系数，以评估内部与外部FHR监测模式对所考虑指标的影响。使用频率为2Hz的FHR信号重复了整个评估。大多数时域线性指标在描记图的初始和最终片段中，外部和内部监测结果相似。然而，线性频域指标在最终片段中相关性较差，在初始片段中均值有显著差异。非线性指标在初始和最终片段中均有显著差异。对于线性和非线性指标，4Hz和2Hz采样参数之间的相关性都很高（大多数相关系数值在0.95和1之间），但非线性指标值在2Hz时显著更高。总之，用于采集FHR信号的模式和所采用的采样率会显著影响大多数FHR指标。
