Heart Rate Variability: Influence of Pre-processing Methods in Identifying Single-Night Sleep-Deprived Subjects

Abstract

Purpose: Autonomic nervous system (ANS) activity can be non-invasively estimated from the heart rate variability (HRV) signal obtained from the electrocardiogram (ECG). The aim of this work is to find useful parameters that allow to establish the presence of a single night sleep deprivation in healthy subjects. Methods: The study included two experimental groups of subjects: Non-sleep-deprived (sleep ≥ 4 h, N = 13) and sleep-deprived (sleep <4 h, N = 10). The RR series extracted from 5 min resting ECG signals were pre-processed using four different algorithms to detect and edit artifacts. The RR series were analyzed in terms of time-domain, frequency-domain and using the Poincaré plot, in order to determine differences in HRV indexes across domains and subject groups. Statistical analyses were performed with the Friedman and Mann–Whitney tests, along, with correlation analysis. Results: Pre-processing methods showed a moderate level of agreement. In subjects with 4 h of single-night sleep deprivation, results differed according to the selected method. Conclusion: SD21 index derived from the Poincaré plot was the only HRV index showing differences between sleep-deprived and non-deprived subjects.