Contents lists available at ScienceDirect Autonomic Neuroscience: Basic and Clinical journal homepage: www.elsevier.com/locate/autneu Inconsistent relation of nonlinear heart rate variability indices to increasing vagal tone in healthy humans ☆ Felipe X. Cepeda a,b , Matthew Lapointe b , Can Ozan Tan b,c , J. Andrew Taylor b,c, ⁎ a Heart Institute (InCor-HCFMUSP) do Hospital das Clínicas da Faculdade de Medicina da Universidade de São Paulo, São Paulo, Brazil b Cardiovascular Research Laboratory, Spaulding Rehabilitation Hospital, Cambridge, MA, United States c Department of Physical Medicine and Rehabilitation, Harvard Medical School, Boston, MA, United States ARTICLE INFO Keywords: Autonomic control Heart rate variability Atropine Vagal outflow RR interval ABSTRACT Background: Prior work has found that linear heart rate variability (HRV) indices do not accurately reflect cardiac vagal control, and nonlinear indices of HRV have been proposed as alternative tools that may better capture cardiac vagal effects. We used progressive low dose atropine to induce changes in cardiac vagal tone to test the hypotheses that nonlinear HRV indices accurately reflect cardiac vagal control, and that their changes in response to low dose atropine correlate with those in RR interval. Methods: Changes in RR interval and HRV indices during intravenous injections of saline (control) and 6 cu- mulative doses of atropine (from 1.4 to 7.2 μg/kg) during controlled breathing at 15 breaths per minute were assessed in 14 young healthy individuals. Results: As expected, low dose atropine increased average RR interval (vagotonic effect). There was no strong association between vagotonic changes in RR interval and the majority of nonlinear HRV indices, either within or among subjects. Conclusions: These data suggest an inconsistent relationship between responses of nonlinear HRV indices and RR interval to changes in cardiac vagal tone. Therefore, nonlinear HRV indices may not be reliable indices of cardiac vagal control in healthy humans. 1. Introduction It is generally accepted that cardiac vagal modulation is among the primary mediators of fluctuations across heart beats (Task Force, 1996). Thus, heart rate variability (HRV) has been proposed to assess the au- tonomic influence on cardiac sinus rhythm (Task Force, 1996), and has been explored as a predictor of outcomes in cardiovascular diseases (Kleiger et al., 1987). However, in the last two decades, it has become evident that the magnitude of variability, assessed via traditional methods, may not directly relate to cardiac vagal modulation. For ex- ample, while respiratory sinus arrhythmia (RSA) seems to represent cardiac vagal control to some degree, sympathetic outflow may also affect the magnitude of RSA (Taylor et al., 2001). Similarly, time do- main estimates of global variability do not always track vagal tone (Picard et al., 2009). As a result, there have been attempts to employ alternative analyses to extract information from HRV that may more explicitly assess cardiac vagal modulation. Among these alternative analyses are nonlinear approaches to de- scribe patterns in HRV. For example, the Poincaré plot, a geometric plot of the current versus successive RR intervals (RRi), has been suggested to encompass a short-term correlation that reflects vagal modulation (Kamen et al., 1996). This has been extended to symbolic analysis of heart rate, which transforms a time series of R-R intervals (RRi) into discrete 3-beat patterns based on current and preceding beats (Guzzetti et al., 2005; Porto et al., 2016). It has been posited that sets of three beats demonstrating a pattern of two unequal changes represents vagal modulation. Somewhat similarly, deceleration capacity derives from a single increase in RRi from one beat to the next in a phase rectified, signal averaged time-series. This rapid ‘deceleration’ has been sug- gested to reflect vagal modulation (Bauer et al., 2006). An alternative https://doi.org/10.1016/j.autneu.2018.04.007 Received 4 January 2018; Received in revised form 5 April 2018; Accepted 30 April 2018 ☆ Institution where the work was performed: This work was performed in Cardiovascular Research Laboratory at Spaulding Rehabilitation Hospital. ⁎ Corresponding author at: Spaulding Rehabilitation Hospital, Cardiovascular Research Laboratory, 1575 Cambridge St, Cambridge, MA 02138, United States. E-mail address: jandrew_taylor@hms.harvard.edu (J. Andrew Taylor). Abbreviations: HRV, Heart rate variability; RSA, respiratory sinus arrhythmia; HR, Heart rate; RRi, RR intervals; SDNN, the standard deviation of all RRi; RMSSD, square root of the mean of the sum of the squares of differences between adjacent normal-to-normal intervals; SD1, the perpendicular dispersion to the line of identity; SD2, the length of the plot along the line of identity; 0 V%, patterns with no variation; 1 V%, patterns with one variation; 2LV%, patterns with two like variations; 2UV%, pattern with two unlike variations; DC, Deceleration capacity; AC, Acceleration capacity; PIP, The percentage of inflection points; IALS, The inverse of the average length; PSS, The percentage of short segments; PAS, The percentage of RRi in alternation segments Autonomic Neuroscience: Basic and Clinical 213 (2018) 1–7 1566-0702/ © 2018 Published by Elsevier B.V. T