BACKGROUND: Quantitative analysis of fetal heart rate variability (HRV) can be used to investigate the neural control mechanisms of fetal cardiac activity. However, conventional power spectrum methods do not reveal the full complexity of the time-varying sympatho-vagal balance in the fetus. AIM: This study was carried out to explore alternative digital signal processing methods of analysing fetal HRV in time domain (rather than frequency domain), in line with most types of physiological monitoring. METHODS: The beat-to-beat fetal heart rate was obtained by Superconducting Quantum Interference Device (SQUID) magnetocardiographic recording. These data were filtered within appropriately selected frequency bands: high frequency (HF) f>0.2 Hz, low frequency (LF) 0.05<f<0.2 Hz and very low frequency (VLF) f<0.05 Hz. To quantify the dynamics of sympathetic-parasympathetic interaction, the integrated amplitude of the HF curve (within a moving 4-s time window) was compared with the LF component. Also, the interaction of the VLF component with both HF and LF components was analysed. RESULTS: A high degree of correlation was observed between these components for extended periods of time, although the presence of a strong correlation was found to depend on the fetal behavioural state. It was further observed that the low-frequency oscillations lag high-frequency activity by 1-3 s. CONCLUSIONS: We propose a numerical parameter, based on the ratio of the LF to HF components of the fetal HRV as a quantitative measure of the relative gain of the parasympathetic system. Our results show this parameter to decrease with gestational age of the fetus.