Analysis of heart rate variability of an anencephalic fetus using a new method to determine a fractal dimension of non-stationary time-serial data.

Human fetal heart rate (HR) variabilities were analyzed using the QIS-A, which we devised to determine a fractal dimension of non-stationary time series. Fifteen 10-min HR data of an anencephalic fetus at 23 weeks and 3 days of gestation and those of 10 normal fetuses at the same weeks of gestation were obtained by ultrasonic cardiography. The anencephalus preserved the spinal cord, medulla and partial anterior hypothalamus. The fractal scaling exponent alpha of the anencephalus was compared with that of each normal fetus by Student's t-test. In results, the scaling relationship in each case had a crossover pattern characterized by alpha(s) and alpha(l), which were slopes above and below a crossover point, respectively. Differences in mean alpha(s) and mean alpha(l) between the anencephalus and each normal fetus were significant (P < 0.01): mean alpha(s), 1.0 +/- 0.1 (+/-SD) (1/f fluctuation) and 1.6 +/- 0.2 (+/-SEM); mean alpha(l), 1.6 +/- 0.2 (+/-SD) and 1.4 +/- 0.1 (+/-SEM). There were six significant differences in mean crossover point between the anencephalus and each normal fetus: 13.8 +/- 5.7 s (+/-SD) and 15.3 +/- 5.6 s (+/-SEM). These results reveal the relationship between fractal structure of fetal HR variability and the developing central nervous system (CNS). In particular, the 1/f fluctuation of HR variability in an anencephalic fetus from the 1.25 to 13.8 s time scale might have a strong relation to the defect of the CNS.