Sensitivity of detrended fluctuation analysis applied to heart rate variability of preterm newborns.

Detrended fluctuation analysis (DFA), a fractal analysis method which is widely used in heart rate variability (HRV) studies, is used to analyze the scaling behaviour of RR interval series of preterm neonates. The average scaling behaviour, calculated using 30000 RR intervals (3 - 4 hours), is characterized by a scaling exponent of 1.4 ± 0.1 at small scales (n ≤ 20) and a smaller exponent of 1.0 ± 0.1 at larger scales. It is shown that the scaling behaviour is not constant over such long segments and how heart rate patterns, associated with specific physiological mechanisms, contribute to the observed variation of the scaling exponents. The effect of the two most important patterns, spikes (either due to faulty peak detection or true decelerations in heart rate) and periodic fluctuations, on the scaling behaviour is investigated.