Ultra-short-term heart rate variability analysis: Comparison between Poincare and frequency domain methods

Abstract

Short-term (5 min) heart rate variability analysis (HRV) demand for monitoring people's health and well-being status is increasing due to portable sensors and mobile applications. Five minutes analysis is extensive compared to that used in measurements of glucose, blood pressure, temperature, etc. To reduce costs and achieve optimal performance in personal health assessment, ultra-short-term monitoring less than 5 min is suggested. The aim of this work is to establish the minimum ultra-short-term HRV analysis time required for LF/HF (LH) (frequency domain sympathy-vagal balance indicator) and SD21 (Poincare graphic domain sympathy-vagal balance indicator) to be equivalent to those obtained by 5 min HRV. From electrocardiograms (ECGs), processing of 23 subjects in the experiment and series of 60, 90, 120, 180, 240 and 300 s (5 min) were obtained. The HRV time series was calculated and each LH and SD21 indicator in each of the ultra-short series was compared with the 300 s (Gold Standard), using different concordance analysis methods, such us Pearson correlation, Bland Altman, Cohen Delta. The ultra-short-term HRV LH indicator is equivalent to short-term HRV (5 min) in series equal or greater than 180 s while SD21 was equivalent to series equal or greater than 120 s.