Analysis of Reproducibility of Noninvasive Measures of Sympathetic Autonomic Control Based on Electrodermal Activity and Heart Rate Variability

Abstract

Heart rate variability (HRV) and electrodermal activity (EDA) are useful tools for assessing the central and peripheral dynamics of the sympathetic nervous system and detecting the effects of numerous systemic diseases and life-challenging situations. However, the indices of HRV and EDA are highly influenced by mental stress, environmental conditions, body position, and other physiological conditions that introduce variability. In this paper, we assessed the five-day reproducibility of HRV and EDA measures of sympathetic control, for N = 20 subjects undergoing 70° head-up tilt test (HUT) and Stroop task tests. We made the assessment in highly controlled conditions without environmental causes of variability, to have a good baseline understanding of the consistency of the various indices of HRV and EDA. Therefore, we assessed intra-subject variation (using the coefficient of variation, CV) and consistency (using the intra-class correlation coefficient, ICC) of the test-to-baseline differences produced by both tests on the studied measures. The low-frequency component of HRV (HRVLF), and its normalized variant was computed as HRV measures of sympathetic control. For EDA, the skin conductance level, frequency of non-specific skin conductance responses, spectral index (EDASympn), and time-varying index (TVSymp) were computed. TVSymp (ICC = 0.85) and HRV indices exhibited higher consistency during the HUT (ICC ≥ 0.8), compared to other EDA measures, and HRVLF was the least variable measure (CV = 85.4%). EDA indices exhibited higher consistency (except for the EDASympn) during the Stroop task (ICC ≥ 0.79) when compared to HRV, and TVSymp was the least variable measure (CV = 97.2%). Remarkably, TVSymp proved to be a reproducible measurement (low variation and high consistency) in both scenarios. These results are the necessary groundwork for studying the use of EDA and HRV in real-world conditions, as reproducibility of the indices has remarkable importance in clinical practice.