Fetal heart rate variability: multiple regression models using autoregressive analysis and fast fourier transform

Abstract

The system is designed to measure the fetal heart rate variability for the evaluation of autonomic nervous system (ANS) indices in the normal and abnormal fetus using Doppler ultrasound method. We have tested the hypothesis that a LF/HF ratio [Parametric autoregressive (AR) and nonparametric fast Fourier transform (FFT) Based] as an index of fetal sympathetic activity is a function of ten variables, age, gestation week, body mass index, CVRR %, HR Mean, HR Std, RMSSD, NN50, pNN 50 and non linear index SD1/SD2 ratio, a multiple regression analysis was performed. The overall model explained 46.47 and 36.12 % of the variation in LF/HF ratio as an index of fetal sympathetic activity using AR and FFT based methods respectively. Age, CVRR %, HR mean, HR Std, and RMSSD are significant predictors (or significantly related to) of LF/HF ratio as an index of fetal sympathetic activity using AR based method and age, CVRR %, HR mean, HR Std and RMSSD are significant predictors using FFT based method. The standardized beta tells us the strength and direction of the relationships (interpreted like correlation coefficients). CVRR % is positively related to LF/HF ratio as an index of fetal sympathetic activity using both the methods