Fractal dimension: a statistic for the normal irregularity of fetal heart rate

Fetal heart rate (FHR) has become a routine physiological measurement both during labour and during the antenatal period when certain pregnancy risk factors have been identified. The normal pattern of FHR is well established and easily recognised while other patterns (eg decelerations, loss of high frequency variability and pseudo-sinusoidal) can be indicative of fetal asphyxia. Most FHR recordings are unequivocal with patterns of normality or abnormality being easily recognised. However, many equivocal traces are encountered in the clinical environment and their interpretation will be dependent upon the experience and competence of the presiding clinician. Since the interpretation of FHR recordings is done predominantly by visual inspection there is potential for much observer variation. Trimbos and Kiersl demonstrated this by asking obstetricians to subjectively interpret a collection of equivocal FHR recordings. These consultants not only disagreed with each other in a significant number of recordings but also with themselves when asked to reinterpret the same recordings at a later time. FHR is a very complex signal. A reproducible interpretation of such a signal requires a systematic analytical approach. The purpose of this paper is to present a method for calculating the fractal dimension and to show how the fractal dimension can be used as an indicator of the amount of irregularity present in a sequential set of FHR numbers. This assumes that high fractal dimension and thus irregularity is healthy