FRACTAL ANALYSIS OF FETAL HEART-RATE-VARIABILITY

Current methods for analysing foetal heart rate (FHR) patterns have ye t to meet their full potential in the recognition of hypoxia in the fo etus. Following the recent suggestion that fractal analysis can be app lied to FHR recordings, the current paper describes a method for disti nguishing two simultaneous fractal dimensions in FHR variation. An irr egular line was plotted from 2500 consecutive foetal heart beat to bea t intervals derived from an ultrasound source. A window of 500 interva ls was moved along the line in steps of 20 intervals. At each step the Richardson technique was used to make estimates of the length of the line within the window using 40 different ruler lengths. When the esti mates were plotted against the ruler lengths on log-log axes the resul ting curve exhibited two distinct linear regions, each demonstrating a n inverse power relationship. From the two slopes the fractal dimensio ns were derived for unspecified low- and high-frequency FHR variation in the current window. The values of both fractal dimensions were plot ted simultaneously with the irregular FHR line and were found to accor d with perceived changes in FHR variation, The method described is sim ply a measure of the irregularity in a series of foetal heart beat to beat intervals: the existence of fractal properties in the irregular l ine does not of itself imply underlying deterministic dynamics (e.g. c haos). This new method of observing FHR variability requires no prepro cessing of the measured data, which are all taken into account. Not on ly does it represent a method for studying normal foetal behaviour but also has potential as a sensitive indicator of impending foetal compr omise.