Asymmetric detrended fluctuation analysis in neonatal stress

To detect stress in newborns by observing heart rate (HR) variability utilizing an asymmetric detrended fluctuation analysis (ADFA), we sought to determine the fractal structure of the series of inter-beat intervals, so as to distinguish the periods of acceleration of the HR from decelerations. Thus, two scaling exponents, α ⁺ and α ^-, representing decelerations and accelerations respectively, are obtained.

Forty healthy term newborns were included in this study, undergoing two different types of stress stimuli: routine heel lance blood sampling for metabolic screening purposes, and its simulation by applying dull pressure on the heel.

It appears that when newborns face stress, the scaling exponent related to accelerations significantly increases and becomes higher than the deceleration scaling exponent. To test the diagnostic properties of the scaling exponents, an ROC curve analysis was applied; α ^- showed good diagnostic performance with an AUC between 0.626 and 0.826, depending on the length of the time series. The joint use of α ⁺ and α ^- further increased the diagnostic performance, in particular for shorter series of RR intervals, with an AUC between 0.691 and 0.833.

ADFA, particularly of the acceleration scaling exponent, may be a useful clinical diagnostic tool for monitoring neonatal stress.