Time domain correlation analysis of heart rate variability in preterm neonates

A fuller understanding of the neural control mechanisms of heart rate during the early stages of human development would be of great value to obstetric and neonatal management. In this paper, we investigate the correlation between heart rate variability (HRV) and other physiological parameters such as blood pressure and respiration in preterm neonates with the aim of developing a numerical model to explain and predict heart rate variability.

All the required data are readily available for premature babies who are routinely monitored while being nursed in intensive care, and we have collected large data sets for a random group of such neonates. For the quantitative analysis of the data, we have developed a time domain correlation method, which has a number of advantages over the more commonly used power spectral analysis. We have been able to study the dynamics of the different frequency components of HRV by this method.

Highly correlated behaviour of the different HRV components, previously observed in our work on fetal HRV, is also present in the neonate, with similar characteristic time constants. Furthermore, the correlation of high-frequency (HF) oscillations of HRV with respiration and that of low-frequency (LF) oscillations of HRV with blood pressure are demonstrated on timescales of a single oscillation. In neonates receiving artificial ventilation, the correlation between HRV and respiration depends on the type of ventilation involved and assumes opposite polarities for the two main types of equipment currently in use.

We demonstrate that it is possible to analyse HRV quantitatively by calculating the relative gains and characteristic time constants for the correlated parameters and components.