Skip to main content
NCBI home page
Annals of Noninvasive Electrocardiology logoLink to Annals of Noninvasive Electrocardiology
. 2018 Jun 4;23(5):e12565. doi: 10.1111/anec.12565

Ultra‐short heart rate variability recording reliability: The effect of controlled paced breathing

Hiago M Melo 1,2,3,✉, Thiago C Martins 1, Lucas M Nascimento 1, Alexandre A Hoeller 3,4, Roger Walz 2,3,4, Emílio Takase 1
PMCID: PMC6931441  PMID: 29863781

Abstract

Background

Recent studies have reported that Heart Rate Variability (HRV) indices remain reliable even during recordings shorter than 5 min, suggesting the ultra‐short recording method as a valuable tool for autonomic assessment. However, the minimum time‐epoch to obtain a reliable record for all HRV domains (time, frequency, and Poincare geometric measures), as well as the effect of respiratory rate on the reliability of these indices remains unknown.

Methods

Twenty volunteers had their HRV recorded in a seated position during spontaneous and controlled respiratory rhythms. HRV intervals with 1, 2, and 3 min were correlated with the gold standard period (6‐min duration) and the mean values of all indices were compared in the two respiratory rhythm conditions.

Results

rMSSD and SD1 were more reliable for recordings with ultra‐short duration at all time intervals (r values from 0.764 to 0.950, p < 0.05) for spontaneous breathing condition, whereas the other indices require longer recording time to obtain reliable values. The controlled breathing rhythm evokes stronger r values for time domain indices (r values from 0.83 to 0.99, p < 0.05 for rMSSD), but impairs the mean values replicability of domains across most time intervals. Although the use of standardized breathing increases the correlations coefficients, all HRV indices showed an increase in mean values (t values from 3.79 to 14.94, p < 0.001) except the RR and HF that presented a decrease (t = 4.14 and 5.96, p < 0.0001).

Conclusion

Our results indicate that proper ultra‐short‐term recording method can provide a quick and reliable source of cardiac autonomic nervous system assessment.

Keywords: autonomic nervous system, cardiac autonomic regulation, controlled paced breathing, heart rate variability, rMSSD, ultra‐short‐term recording

1. INTRODUCTION

Heart rate variability (HRV) is a commonly noninvasive autonomic biomarker used in clinical protocols (Beauchaine & Thayer, 2015; Task Force, 1996; Thayer & Lane, 2009; Thayer, Åhs, Fredrikson, Sollers, & Wager, 2012). The quantitative values obtained by R–R intervals analysis has reliability to distinguish patients with obesity (Iliescu, Tudorancea, Irwin, & Lohmeier, 2013; Yadavet al., 2017), diabetes (Arroyo‐Carmonaet al., 2016; Dimitropoulos, 2014), psychiatric disorders (Chalmers, Quintana, Abbott, & Kemp, 2014), and others specific situations (for a review see Thayer & Lane, 2009; Thayeret al., 2012;), when the impairment of vagal tone regulation is reported as a common biomarker for these clinical conditions (Laborde, Mosley, & Thayer, 2017; Thayer & Lane, 2000,2009). In order to obtain reliable data for the HRV quantitative analysis, it is recommended to carry out recordings with 5 min or longer, according to the research design (Labordeet al., 2017; Task Force, 1996).

Recent studies have reported that some time‐domain HRV indices remain reliable even in recordings shorter than 5 min, proposing the ultra‐short recording method (Esco & Flatt, 2014; Nakamuraet al., 2015; Thong, Li, McNames, Aboy, & Goldstein, 2003; Voss, Schulz, Schroeder, Baumert, & Caminal, 2009). In these experiments, the comparison of HRV indices, such as root mean square of standard deviation (rMSSD) or standard deviation of RR interval (SDNN), showed strong correlations between random periods shorter than 1 min (10, 30, and 60 s) and standard gold intervals (longer than 5 min; Munozet al., 2015; Nussinovitchet al., 2011). The reliability of ultra‐short recordings goes beyond measures at rest, being stable also to discriminate autonomic adaptability to physical training in elite athletes (Nakamura et al., 2015; Pereira, Flatt, Ramirez‐Campillo, Loturco, & Nakamura, 2016).

The time‐domain reliability is not replicated to the frequency domain indices, this may occur because most studies do not control the participants' breathing rhythm (Munozet al., 2015; Nussinovitchet al., 2011). The interindividual changes in respiratory rate during long‐term recording may impair frequency domain analysis based on the Fast Fourier Transform (FFT), so the alteration of HRV oscillatory morphology may produces inconstant data segments for analysis (Nussinovitchet al., 2011). To avoid this effect, some researchers have performed HRV recordings with standardized respiratory rhythms by using breathing pacer (Flatt & Esco, 2013).

Although breathing pacer can standardize heart rate oscillations, there are no evidences of its effects in data reliability improvement during ultra‐short recordings. In this sense, the aims of this study were: (a) To analyze ultra‐short recordings reliability including time, frequency, and Poincare geometric measures domains at different time intervals (1, 2, and 3 min) compared to the gold standard interval; and, (b) To compare the reliability of these indices in the spontaneous and controlled breathing condition (6 breaths/min) besides also evaluate the effects of controlled breathing in HRV indices.

2. METHODS

2.1. Participants

HRV recordings were carried out in 20 healthy male volunteers, aged between 18–28 years old (23.65 ± 3.24). The following exclusion criteria were applied: (a) Diagnostic of cardiac arrhythmia; (b) The intake of caffeine or energetic drinks three hours prior to data collection; (c) Alcoholic drinking 24 hr before the experiment; (d) Sleep deprivation in the night before the research; (e) Report of discomfort when performing the 6 breaths/min breathing rate. This study was approved by a local Ethics Committee, and was conducted according to the Helsinki Declaration. All participants were informed of experimental risks and research benefits, and signed an informed consent form agreeing with the procedures.

2.2. Physiologic recording

Participants’ chest skin was cleaned with isopropyl alcohol (70%) skin swab, followed by the fixation of a comfortable elastic straps to the chest for the acquisition of RR intervals (iRR, in ms) by a PolarH7 heart rate monitor (Polar, Finland). All acquisitions have occurred in the period between 9 and 12 hr in a bright and quiet room. The resting baseline was made in sitting position, followed by a breathing paced condition (6 breaths/min), both periods were 6 min long.

The Kubios software (v.2.2) (Tarvainen, Niskanen, Lipponen, & Ranta‐aho, & Karjalainen, 2014) was used to analyze the RR intervals data and artifacts that were manually filtered by visual inspection. Ectopic beats and artifacts were automatically detected and replaced by interpolated adjacent beats by applying an adaptive filter to generate normal‐to‐normal (NN) interval time series (as seen in Abad et al., 2017; Nakamura et al., 2015; Pereira et al., 2016; Tarvainen et al., 2014). HRV was analyzed in time, frequency and Poincare geometric measures domain (see Tarvainen et al., 2014 for detailed HRV indices calculation). The time domain included mean RR intervals (RR), standard deviation of RR interval (SDNN), root mean square of standard deviation (rMSSD), % of RR intervals with difference in successive RR intervals longer than 50 ms (pNN50). The frequency domain features were performed by Fast Fourier Transform (FFT), ascribing the % of low frequency (LF) (0.04–0.15 Hz), % of high frequency (HF) (0.15–0.4 Hz), and the low frequency‐to‐high frequency ratio (LF:HF). The Poincare geometric measures domain was composed by the perpendicular points dispersion to the line of instantaneous identity (SD1), points dispersion of the along the long‐term identity line (SD2) and correlation dimension (D2) (Brennan, Palaniswami, & Kamen, 2001; Kamen, Krum, & Tonkin, 1996; Tarvainen et al., 2014; Task Force, 1996).

For each HRV recording (6 min), the quantitative data analysis was performed for all time intervals (1, 2, 3, and 6 min) in the rest and controlled breathing conditions. Thus, for each HRV recording we extracted 1‐min (n = 6), 2‐min (n = 3), and 3‐min epochs (n = 2; Figure 1).

Figure 1.

Figure 1

Open in a new tab

Timeline ofepochs extraction. The numbers represent the time interval (1, 2, and 3 min) and the letters an analysis period (A, first epoch; B, second epoch; C, third epoch; D, fourth epoch; E, fifth epoch; F, sixth epoch)

2.3. Statistical analysis

The Pearson correlation was performed between the HRV indices with different time intervals (1, 2, 3, and 6 min) and the gold standard (6 min). The statistical test was performed separately for each respiratory rhythm condition. The reproducibility of time intervals mean values was verified by paired t‐test between the gold standard time intervals and other time epochs (1, 2, and 3 min). A paired t‐test was used to investigate the difference in HRV indices during the two breathing conditions, for this procedure the gold standard interval was used. All statistical procedures were performed in Stata 14©, for t‐tests the significance level was set at p < 0.05 and Pearson correlations p‐values were adjusted by Bonferroni correction.

3. RESULTS

Data from the correlation between ultra‐short time intervals and gold standard time during the breathing conditions for allHRV indices are presented in Tables 1and 2. For baseline spontaneous breathing period the time domain indices (mean RR, rMSSD and pNN50) showed better correlation results (r values from 0.298 to 0.972) and higher frequencies of p < 0.05 for 1‐min epochs, with emphasis in rMSSD for improved r values (from 0.768 to 0.919, p < 0.05) in 1‐min epochs and for all time intervals (r values from 0.768 to 0.960, p < 0.05). The frequency domain indices (LF, HF, and LF:HF) was poorly correlated in the 1‐, 2‐, or 3‐min periods (r values from 0.223 to 0.952). For nonlinear indices, the SD1 presented similar results to rMSSD (r values from 0.764 to 0.965, all time intervals with p < 0.05), and the correlation of the other nonlinear indices (SD2 and D2) showed few significant p‐values for all time intervals (as seen in Table 1).

Table 1.

Pearson correlation results (Bonferroni corrected p‐values) between all‐time intervals (1, 2, and 3 min) and the gold standard period (6 min) for spontaneous breathing condition

Heart rate variability indices 1A 1B 1C 1D 1E 1F 2A 2B 2C 3A 3B
RR 0.885* (0.001) 0.894* (0.001) 0.981* (0.001) 0.945* (0.001) 0.927* (0.001) 0.864* (0.001) 0.298 (1.0) 0.984* (0.001) 0.926* (0.001) 0.972* (0.001) 0.958* (0.001)
SDNN 0.822* (0.001) 0.448 (1.0) 0.598 (0.38) 0.543 (0.98) 0.514 (1.0) 0.577 (0.45) 0.826* (0.001) 0.676 (0.08) 0.775* (0.001) 0.841* (0.001) 0.798* (0.001)
rMSSD 0.891* (0.001) 0.919* (0.001) 0.875* (0.001) 0.906* (0.001) 0.890* (0.001) 0.768* (0.001) 0.934* (0.001) 0.950* (0.001) 0.890* (0.001) 0.960* (0.001) 0.924* (0.001)
pNN50 0.842* (0.001) 0.917* (0.001) 0.855* (0.001) 0.813* (0.001) 0.810* (0.001) 0.472 (1.0) 0.920* (0.001) 0.901* (0.001) 0.869* (0.001) 0.958* (0.001) 0.874* (0.001)
LF 0.291 (1.0) 0.569 (0.5) 0.829* (0.001) 0.101 (1.0) 0.275 (1.0) 0.605 (0.29) 0.281 (1.0) 0.729* (0.001) 0.544 (0.96) 0.637 (0.15) 0.436 (1.0)
HF 0.777* (0.001) 0.546 (0.95) 0.826* (0.001) 0.387 (1.0) 0.223 (1.0) 0.448 (1.0) 0.761* (0.001) 0.879* (0.001) 0.398 (1.0) 0.900* (0.001) 0.443 (1.0)
LF:HF 0.512 (1.0) 0.740* (0.001) 0.690* (0.02) 0.346 (1.0) 0.525 (1.0) 0.005 (1.0) 0.905* (0.001) 0.639 (0.31) 0.686* (0.01) 0.952* (0.001) 0.538 (0.65)
SD1 0.869* (0.001) 0.921* (0.001) 0.867* (0.001) 0.965* (0.001) 0.905* (0.001) 0.764* (0.003) 0.933* (0.001) 0.932* (0.001) 0.869* (0.001) 0.955* (0.001) 0.884* (0.001)
SD2 0.776* (0.001) 0.377 (1.0) 0.440 (1.0) 0.426 (1.0) 0.415 (1.0) 0.487 (1.0) 0.766* (0.004) 0.508 (1.0) 0.644 (0.15) 0.647 (0.15) 0.548 (0.72)
D2 0.623 (0.21) 0.433 (0.42) 0.628 (0.70) 0.796* (0.001) 0.351 (1.0) 0.307 (1.0) 0.698* (0.04) 0.825* (0.001) 0.463 (1.0) 0.886* (0.001) 0.457 (1.0)
Open in a new tab

Notes. (Number and Letter) The numbers represent the time interval (1, 2, and 3 min) and the letters an analysis period (A, first epoch; B, second epoch; C, third epoch; D, fourth epoch; E, fifth epoch; F, sixth epoch); D2 correlation dimension; HF: high frequency (0.15–0.4 Hz); HR: heart rate; LF: low frequency (0.04–0.15 Hz); LF:HF: low frequency‐to‐high frequency ratio; pNN50: % of RR intervals with difference in successive RR intervals longer than 50 ms; RR: mean of RR intervals; rMSSD: root mean square of standard deviation; SD: standard deviations; SD1: perpendicular points dispersion to the line of instantaneous identity; SD2: points dispersion of the along the long‐term identity line; SDNN: standard deviation of RR interval.

*p < 0.05 for Bonferroni corrected Pearson correlation analysis.

Table 2.

Pearson correlation results (Bonferroni corrected p‐values) between all‐time intervals (1, 2, and 3 min) and the gold standard period (6 min) for controlled breathing condition

Heart rate variability 1A 1B 1C 1D 1E 1F 2A 2B 2C 3A 3B
RR 0.950* (0.001) 0.977* (0.001) 0.969* (0.001) 0.966* (0.001) 0.972* (0.001) 0.933* (0.001) 0.983* (0.001) 0.963* (0.001) 0.970* (0.001) 0.991* (0.001) 0.986* (0.001)
SDNN 0.923* (0.001) 0.923* (0.001) 0.926* (0.001) 0.895* (0.001) 0.785* (0.002) 0.704* (0.03) 0.952* (0.001) 0.941* (0.001) 0.850* (0.001) 0.980* (0.001) 0.920* (0.001)
rMSSD 0.905* (0.001) 0.904* (0.001) 0.921* (0.001) 0.987* (0.001) 0.849* (0.001) 0.825* (0.001) 0.955* (0.001) 0.970* (0.001) 0.906* (0.001) 0.973* (0.001) 0.949* (0.001)
pNN50 0.925* (0.001) 0.916* (0.001) 0.939* (0.001) 0.867* (0.001) 0.901* (0.001) 0.775* (0.003) 0.965* (0.001) 0.963* (0.001) 0.938* (0.001) 0.981* (0.001) 0.969* (0.001)
LF 0.184 (1.0) 0.637 (0.16) 0.674 (0.07) 0.204 (1.0) 0.447 (1.0) 0.157 (1.0) 0.632 (0.18) 0.570 (0.57) 0.276 (1.0) 0.794* (0.001) 0.355 (1.0)
HF 0.355 (1.0) 0.601 (0.33) 0.508 (1.0) 0.403 (1.0) 0.277 (1.0) 0.300 (1.0) 0.461 (1.0) 0.592 (0.39) 0.143 (1.0) 0.773* (0.004) 0.267 (1.0)
LF:HF 0.129 (1.0) 0.729* (0.01) 0.692* (0.04) 0.333 (1.0) 0.189 (1.0) 0.096 (1.0) 0.096 (1.0) 0.561 (0.66) 0.450 (1.0) 0.048 (1.0) 0.485 (1.0)
SD1 0.905* (0.001) 0.904* (0.001) 0.921* (0.001) 0.896* (0.001) 0.850* (0.001) 0.826* (0.001) 0.955* (0.001) 0.970* (0.001) 0.909* (0.001) 0.973* (0.001) 0.948* (0.001)
SD2 0.922* (0.001) 0.921* (0.001) 0.923* (0.001) 0.896* (0.001) 0.781* (0.003) 0.691* (0.04) 0.950* (0.001) 0.942* (0.001) 0.849* (0.001) 0.976* (0.001) 0.918* (0.001)
D2 0.546 (0.83) 0.321 (1.0) 0.264 (1.0) 0.215 (1.0) 0.094 (1.0) 0.507 (1.0) 0.632 (0.18) 0.360 (1.0) 0.518 (1.0) 0.597 (0.36) 0.485 (1.0)
Open in a new tab

Notes. (Number and Letter) The numbers represent the time interval (1, 2, and 3 min) and the letters an analysis period (A, first epoch; B, second epoch; C, third epoch; D, fourth epoch; E, fifth epoch; F, sixth epoch); D2 correlation dimension; HF: high frequency (0.15–0.4 Hz); HR: heart rate; LF: low frequency (0.04–0.15 Hz); LF:HF: low frequency‐to‐high frequency ratio; pNN50: % of RR intervals with difference in successive RR intervals longer than 50 ms; RR: mean of RR intervals; rMSSD: root mean square of standard deviation; SD: Standard Deviations; SD1: perpendicular points dispersion to the line of instantaneous identity; SD2: points dispersion of the along the long‐term identity line; SDNN: standard deviation of RR interval.

*p < 0.05 for Bonferroni corrected Pearson correlation analysis.

Following Pearson correlation, the controlled breathing period did not change the frequency domain of reliability indices (r values from 0.184 to 0.674) and did not improve de frequency of significant p‐values (as seen in Table 2). The time domain indices (RR, rMSSD, and pNN50) maintained their reliability for all time intervals, where rMSSD remained the most reliable due to higher r values (r values from 0.825 to 0.987, p < 0.05), whereas the SDNN index had its reliability increased for all time intervals (r values from 0.704 to 0.980, p < 0.05 for all time intervals). For nonlinear indices, the results were similar in that all indexes presented replicability for 3 or 3 min intervals, but theSD2 index had its reliability increased for all time intervals (r values from 0.691 to 0.976, p < 0.05) when compared with spontaneous breathing (r values from 0.415 to 0.776, only r = 0.776 has p < 0.05). It was only observed that the controlled breathing condition increased the r values for indices all domains (time, frequency, and nonlinear), keeping significantly r values from indices with p < 0.05 in spontaneous breathing condition and improving deSDNN and SD2 reliability.

The reproducibility of mean HRV indices during spontaneous breathing condition for 1‐, 2‐, and 3‐min was tested by paired samples t‐test between 6‐min interval and other time epochs (Table 3). For time domain indices, rMSSD and pNN50 presented few significant differences between different time intervals (only one time interval presented a significant difference from 6‐min epoch). Similar results were observed in LF:HF, SD1 and D2 indices. For controlled breathing condition, all indices presented an increase in the frequency of statistically different values, except the HF values that were reproducible at all time intervals (Table 4).

Table 3.

Comparison between mean heart rate variability (HRV) indices (p‐values) of all‐time intervals (1, 2, and 3 min) and the gold standard period (6 min) for spontaneous breathing condition

HRV 6A 1A 1B 1C 1D 1E 1F 2A 2B 2C 3A 3B
RR 809.8 ± 36.7 821.6* ± 46.4 (0.02) 816.7 ± 46.4 (0.16) 811.7 ± 39.8 (0.29) 804.6* ± 35.9 (0.04) 803.7* ± 35.2 (0.04) 804.1 ± 39.4 (0.21) 800.6 ± 86.2 (0.62) 807.7 ± 36.7 (0.16) 802.8* ± 34.9 (0.03) 816.3* ± 41.8 (0.01) 803.7* ± 34.6 (0.01)
SDNN 36.4 ± 8.02 32.02* ± 12.34 (0.01) 28.7* ± 7.21 (0.004) 31.99* ± 10.1 (0.02) 32.4* ± 9.7 (0.05) 32.1 ± 11.6 (0.06) 29.48* ± 11.9 (0.004) 33.6* ± 9.9 (0.03) 33.6 ± 8.5 (0.07) 33.6* ± 9.4 (0.05) 34.5 ± 9.1 (0.09) 33.9* ± 8.4 (0.04)
rMSSD 20.1 ± 4.8 21.7 ± 7.6 (0.13) 20.6 ± 5.6 (0.57) 20.1 ± 5.4 (0.81) 20.2 ± 5.5 (0.81) 19.7 ± 4.6 (0.27) 18.49* ± 3.7 (0.01) 21.2 ± 6.4 (0.12) 20.2 ± 5.2 (0.87) 19.1* ± 4.1 (0.04) 20.9 ± 5.8 (0.13) 19.6 ± 4.4 (0.10)
pNN50 2.1 ± 2.7 3.7 ± 6.4 (0.10) 2.55 ± 3.8 (0.31) 2.07 ± 3.09 (0.69) 2.27 ± 1.74 (0.75) 1.7 ± 2.08 (0.27) 0.64* ± 1.2 (0.01) 3.1 ± 4.8 (0.07) 2.1 ± 2.9 (0.94) 1.1* ± 1.3 (0.03) 2.7 ± 4.07 (0.08) 1.5 ± 1.9 (0.09)
LF 33.5 ± 12.8 40.3 ± 19.7 (0.14) 42.9* ± 18.6 (0.01) 41.9* ± 19.6 (0.003) 51.1* ± 19.7 (0.002) 47.3* ± 18.3 (0.004) 43.4* ± 16.8 (0.009) 37.5 ± 15.5 (0.31) 37.5 ± 18.2 (0.16) 40.6* ± 13.6 (0.02) 34.6 ± 14.1 (0.69) 37.3 ± 9.01 (0.16)
HF 14.6 ± 8.7 20.9* ± 19.7 (0.005) 27.47* ± 15.9 (0.004) 17.23 ± 13.07 (0.14) 18.7 ± 10.76 (0.10) 20.01 ± 13.16 (0.10) 21.4 ± 12.05 (0.01) 19.6* ± 14.9 (0.03) 16.2 ± 10.12 (0.17) 17.78 ± 12.3 (0.25) 17.27* ± 11.52 (0.03) 15.24 ± 9.06 (0.77)
LF:HF 3.06 ± 1.8 2.9 ± 2.6 (0.89) 3.01 ± 2.7 (0.90) 3.9 ± 3.02 (0.08) 4.1 ± 3.4 (0.15) 3.8 ± 3.3 (0.22) 2.6 ± 2.1 (0.52) 3.1 ± 2.1 (0.79) 3.9 ± 4.4 (0.27) 3.3 ± 2.6 (0.44) 2.9 ± 2.06 (0.34) 3.6 ± 2.2 (0.22)
SD1 14.3 ± 3.4 15.4 ± 5.4 (0.10) 14.6 ± 4.03 (0.43) 14.3 ± 3.88 (0.99) 14.6 ± 4.1 (0.50) 14.08 ± 3.2 (0.40) 13.17* ± 2.7 (0.02) 15.07 ± 4.5 (0.10) 14.3 ± 3.7 (0.95) 13.6* ± 2.8 (0.05) 14.8 ± 4.1 (0.12) 13.7 ± 3.8 (0.14)
SD2 50.7 ± 12.5 42.07* ± 17.5 (0.002) 37.5* ± 10.7 (0.003) 42.6* ± 14.6 (0.02) 43.1* ± 14.21 (0.02) 42.8* ± 16.8 (0.04) 39.2* ± 17.5 (0.004) 44.7* ± 14.2 (0.008) 45.1* ± 12.2 (0.05) 45.4* ± 13.6 (0.004) 47.7 ± 14.8 (0.25) 47.1 ± 13.76 (0.21)
D2 1.1 ± 0.8 1.02 ± 0.8 (0.38) 0.87 ± 0.6 (0.08) 1.1 ± 0.8 (0.90) 1.5 ± 1.5 (0.09) 1.2 ± 0.9 (0.79) 0.7* ± 0.6 (0.03) 1.04 ± 0.7 (0.39) 1.05 ± 0.6 (0.27) 1.1 ± 0.8 (0.79) 1.06* ± 0.7 (0.03) 1.1 ± 0.7 (0.95)
Open in a new tab

Notes. (Number and Letter) The numbers represent the time interval (1, 2, and 3 min) and the letters an analysis period (A, first epoch; B, second epoch; C, third epoch; D, fourth epoch; E, fifth epoch; F, sixth epoch); D2 correlation dimension; HF: high frequency (0.15–0.4 Hz); HR: heart rate; LF: low frequency (0.04–0.15 Hz); LF:HF: low frequency‐to‐high frequency ratio; pNN50: % of RR intervals with difference in successive RR intervals longer than 50 ms; RR: mean of RR intervals; rMSSD: root mean square of standard deviation; SD: Standard Deviations; SD1: perpendicular points dispersion to the line of instantaneous identity; SD2: points dispersion of the along the long‐term identity line; SDNN: standard deviation of RR interval.

*p < 0.05 values for paired Student'st‐test between 6 min epoch and the respective time interval.

Table 4.

Comparison between mean heart rate variability (HRV) indices (p‐values) of all‐time intervals (1, 2, and 3 min) and the gold standard period (6 min) for controlled breathing condition

HRV 6A 1A 1B 1C 1D 1E 1F 2A 2B 2C 3A 3B
RR 754.4 ± 47.1 781.4* ± 52.04 (0.001) 761.8* ± 54.2 (0.01) 747.6* ± 52.4 (0.03) 746.1* ± 44.8 (0.007) 743.8* ± 48.6 (0.006) 746.8* ± 45.5 (0.06) 771.3* ± 52.2 (0.001) 743.1* ± 54.1 (0.003) 745.3* ± 46.2 (0.002) 763.1* ± 51.6 (0.001) 745.5* ± 45.01 (0.001)
SDNN 73.1 ± 11.7 75.9 ± 16.06 (0.08) 71.7 ± 15.7 (0.37) 69.6* ± 13.5 (0.005) 68.7* ± 12.3 (0.002) 70.2 ±  9.8 (0.09) 68.4* ± 11.9 (0.03) 74.9 ±  15.7 (0.18) 69.8* ± 12.3 (0.002) 69.9* ± 9.3 (0.03) 74.4 ±  14.2 (0.12) 69.8* ± 10.08 (0.006)
rMSSD 33.2 ± 6.08 37.1* ± 7.5 (0.001) 34.2 ± 7.5 (0.19) 32.4 ± 7.5 (0.23) 31.2* ± 6.1 (0.004) 31.7* ± 5.2 (0.04) 32.06 ±  6.4 (0.16) 35.7* ± 7.1 (0.001) 31.9* ± 6.5 (0.001) 31.9* ± 5.4 (0.03) 34.6* ± 7.1 (0.003) 31.8* ± 5.5 (0.04)
pNN50 13.6 ± 7.5 19.3* ± 10.5 (0.001) 14.6 ± 9.8 (0.27) 12.9 ± 7.9 (0.29) 12.2 ± 8.6 (0.16) 11.7* ± 6.9 (0.01) 11.8 ± 7.9 (0.15) 16.9* ± 9.7 (0.001) 12.3* ± 7.5 (0.01) 11.7* ± 6.6 (0.006) 15.5* ± 8.8 (0.005) 11.7* ± 6.6 (0.005)
LF 81.4 ± 6.2 87.7* ± 8.4 (0.008) 84.7 ± 13.4 (0.18) 85.02* ± 8.7 (0.02) 79.5 ± 16.6 (0.6) 85.1 ± 11.4 (0.12) 83.5 ± 15.9 (0.57) 87.9* ± 8.2 (0.003) 78.3 ± 11.4 (0.15) 82.03 ± 12.01 (0.83) 84.6* ± 8.6 (0.01) 82.8 ± 9.03 (0.48)
HF 2.8 ± 0.9 3.07 ± 1.9 (0.64) 4.3 ± 1.8 (0.24) 2.8 ± 1.2 (0.95) 2.6 ± 1.2 (0.44) 2.8 ± 1.3 (0.99) 3.2 ± 1.8 (0.42) 2.6 ± 1.07 (0.44) 2.8 ± 0.9 (0.67) 2.9 ± 1.7 (0.81) 3.06 ± 1.6 (0.45) 2.8 ± 0.9 (0.96)
LF:HF 31.4 ± 11.3 42.3 ± 27.1 (0.09) 31.09 ± 15.9 (0.84) 35.4 ± 15.1 (0.12) 36.1 ± 14.7 (0.18) 35.5 ± 14.1 (0.28) 32.8 ± 17.5 (0.75) 36.9* ± 12.8 (0.05) 30.3 ± 9.6 (0.62) 34.6 ± 16.3 (0.35) 32.6 ± 12.5 (0.40) 33.1 ± 12.4 (0.54)
SD1 23.5 ± 4.3 26.4*±5.3 (0.001) 24.3 ± 5.3 (0.13) 23.07 ± 5.4 (0.34) 22.2* ± 4.4 (0.01) 22.57 ± 3.7 (0.07) 22.8 ± 4.6 (0.23) 25.3* ± 5.1 (0.001) 22.6* ± 4.6 (0.002) 22.7* ± 3.8 (0.04) 25.5* ± 5.03 (0.002) 22.5* ± 3.9 (0.004)
SD2 100.6 ± 16.1 104.09 ± 22.3 (0.13) 98.5 ± 21.9 (0.32) 95.7* ± 18.6 (0.006) 94.6* ± 16.9 (0.002) 96.7 ± 13.8 (0.10) 94.1* ± 16.6 (0.03) 103.8 ± 21.9 (0.26) 95.6* ± 17.3 (0.001) 96.09* ± 13.06 (0.02) 101.8 ± 20.1 (0.36) 96.2* ± 13.8 (0.006)
D2 2.1 ± 0.1 1.7* ± 0.2 (0.001) 2.2 ± 2.1 (0.77) 1.7* ± 0.1 (0.001) 1.8* ± 0.1 (0.001) 1.7* ± 0.1 (0.001) 1.8* ± 0.2 (0.001) 1.8* ± 0.2 (0.001) 2.01* ± 0.1 (0.04) 2.07* ± 0.2 (0.05) 1.9* ± 0.2 (0.01) 2.09 ± 0.3 (0.87)
Open in a new tab

Notes. (Number and Letter) The numbers represent the time interval (1, 2, and 3 min) and the letters an analysis period (A, first epoch; B, second epoch; C, third epoch; D, fourth epoch; E, fifth epoch; F, sixth epoch); D2 correlation dimension; HF: high frequency (0.15–0.4 Hz); HR: heart rate; LF: low frequency (0.04–0.15 Hz); LF:HF: low frequency‐to‐high frequency ratio; pNN50: % of RR intervals with difference in successive RR intervals longer than 50 ms; RR: mean of RR intervals; rMSSD: root mean square of standard deviation; SD: Standard Deviations; SD1: perpendicular points dispersion to the line of instantaneous identity; SD2: points dispersion of the along the long‐term identity line; SDNN: standard deviation of RR interval.

*p < 0.05 values for paired t‐test between 6 min epoch and the respective time interval.

According with paired t‐test, significant changes were observed in all HRV indices in time, frequency, and Poincare geometric measures domains (p < 0.001) in the comparison between rest and controlled breathing conditions (Table 5). Accordingly, a significant increase in SDNN (t = 11.52, p = 0.0001), rMSSD (t = 7.41, p = 0.0001), pNN50 (t = 6.36, p = 0.0001), LF (t = 14.94, p = 0.0001), LF:HF (t = 11.03, p = 0.0001), SD1 (t = 7.41, p = 0.0001), SD2 (t = 9.75, p = 0.0001), and D2 (t = 3.79, p = 0.001) were noticed and associated with a significant decrease in the RR (t = 4.14, p = 0.0001) and HF (t = 5.96, p = 0.0001) in the controlled breathing condition.

Table 5.

Comparison between mean heart rate variability values during the spontaneous and controlled breathing conditions

Indices Spontaneous mean ± SD Controlled mean ± SD t P
RR 809.85 ± 36.78 754.46 ± 47.14 4.14 0.0001*
SDNN 36.48 ± 8.02 73.11 ± 11.73 11.52 0.0001*
rMSSD 20.31 ± 4.89 33.26 ± 6.08 7.41 0.0001*
pNN50 2.14 ± 2.76 13.61 ± 7.56 6.36 0.0001*
LF 33.55 ± 12.89 81.47 ± 6.27 14.94 0.0001*
HF 14.64 ± 8.76 2.88 ± 0.93 5.96 0.0001*
LF:HF 3.06 ± 1.89 31.486 ± 11.35 11.03 0.0001*
SD1 14.38 ± 3.47 23.55 ± 4.30 7.41 0.0001*
SD2 42.07 ± 17.52 100.66 ± 16.18 9.75 0.0001*
D2 1.02 ± 0.18 2.10 ± 0.16 3.79 0.001*
Open in a new tab

Notes. D2 correlation dimension; HF: high frequency (0.15–0.4 Hz); HR: heart rate; LF: low frequency (0.04–0.15 Hz); LF:HF: low frequency‐to‐high frequency ratio; pNN50: % of RR intervals with difference in successive RR intervals longer than 50 ms; RR: mean of RR intervals; rMSSD: root mean square of standard deviation; SD: Standard Deviations; SD1: perpendicular points dispersion to the line of instantaneous identity; SD2: points dispersion of the along the long‐term identity line; SDNN: standard deviation of RR interval.

*p < 0.05 values for paired t‐test.

4. DISCUSSION

This study investigated the effects of controlled breathing rhythm on short‐term HRV indices reliability. The results suggest that rMSSD and SD1 have the higher r coefficients and more replicable mean values for all time intervals in spontaneous breathing conditions. The breathing rate standardization impairs the HRV indices reliability, except in the case of HF. Other HRV indices showed poorly r coefficients and high frequency of mean values with statistically significant difference (when compared to 6‐min epochs), suggesting that the gold standard time interval (>5 min) are required for reliable recordings. The comparison between mean HRV indices during breathing conditions suggests that controlled breathing protocol presents a significant effect on all HRV indices.

Previous studies that compare HRV indices in different time intervals support the same relation between time domain indices and higher replicability in short‐term time intervals. The comparison between HRV recordings with 10, 60, and 300 s, showed that only rMSSD presents reliability to short‐term measurements, whereas other indices (including other time domain indices as SDNN) require longer measurements (Nussinovitchet al., 2011; Thonget al., 2003). Important to say, although these studies found significant correlations only in rMSSD, they have analyzed only a random time interval within the total sample, which can cause artifact epochs to be chosen for analysis by decreasing the r values of the indices that depend of the signal morphology. Recent studies, with large samples (n = 3,387) using more than one signal epoch for short‐term comparisons, found that rMSSD presents reliability to short‐term comparison, requiring at least 30 s for replicable data (Munozet al., 2015).

The rMSSD higher reliability makes it commonly reported in studies using ultra‐short recordings. Studies that compare elite athletes in rest and postexercise conditions showed that the reliability to this measurement modality can discriminate physical adaptation to train load and fatigue (Esco & Flatt, 2014; Nakamuraet al., 2015; Schmitt, Regnard, & Millet, 2015). As in elite sport training, short‐term rMSSD values have been used to clinical samples, such as observed in patients with diabetes (Chen, Yang, Liu, & Tang, 2015) or mental stress (Salahuddin, Cho, Jeong, & Kim, 2007).

The high replicability of rMSSD during ultra‐short recordings can be understood by its mathematical formula. Calculated by the difference of consecutive R intervals (Task Force, 1996), the rMSSD values are poorly influenced by fluctuations in heart rate values, making the measurement more stable even in periods with different heart oscillations (Saboul, Pialoux, & Hautier, 2013). The identical mathematical metrics between rMSSD and SD1 explains the similar reliability of these indices (Ciccone et al., 2017). Furthermore, indices such as SDNN or those of the frequency domain that are not derived from the difference between RR intervals, have a direct influence on fluctuations of heart rate values, so that small changes in one specific period of the record are not replicable in the others (Nussinovitchet al., 2011).

Although the frequency and Poincare geometric measures domain indices are more influenced by heart rate oscillations, the results suggest that breathing standardization does not improve the indices reliability during short‐term recordings. This may have occurred because there is no synchronization between waves peaks and valleys of different epochs, a factor that directly influences the values based on FFT (Labordeet al., 2017). Therefore, the lack of synchrony between the beginning of respiratory cycle can contribute to reduce the reliability of these indices at different epochs. In addition, there is no certainty that all individuals have been able to adapt equally to the breathing pace and maintain it for a prolonged time, which may justify impairment in some time domain indices (rMSSD and pNN50).

Important to say, these results are supported by the correlation analysis, despite the mean values comparison of HF suggests that all time intervals values are similar during controlled breathing condition. Thus, our hypothesis is that HF decrease during controlled respiration condition may reduce the interindividual variation, as all participants have a significant reduction (nearly around to zero values). Although there is better replicability of mean values, the robust effect of controlled respiration in HF may impair some experimental effects (the participant may still be influenced by the controlled breathing effect). Thus, the assessment of HF under these breathing conditions may only reflects the individual's adaptability to the breathing rhythm and not to reflect the proper resting baseline values.

During the fixed breathing pacer of 6 breaths/min there is a sinusoidal variation induction in RR intervals (Vaschillo, Vaschillo, & Lehrer, 2006). This pattern is related to the synchrony between heart rate and breathing frequency caused by resonant frequency rhythm (Lehrer, Vaschillo, & Vaschillo, 2000; Lehreret al., 2004), denominated Respiratory Sinus Arrhythmia (ASR), a phenomenon that represents the increase in HR during inspiration and its reduction during expiration (Billman, Huikuri, Sacha, & Trimmel, 2015; Yasuma & Hayano, 2004).

According to our findings, studies comparing controlled and spontaneous breathing rhythm conditions reported significant changes in all HRV domains (Barth & Vecchio, 2014; Tavareset al., 2016). These findings suggest that if breathing rate control is adopted during a baseline rest record, the values obtained will already be influenced by a variable that may change the HRV values, compromising the investigation of other variables that would have the potential to change the HRV indices. In addition, the induction of a fixed breathing rhythm may evoke different effects on participants who are familiar with paced respiration, such as yoga or meditation practitioners (Krygieret al., 2013), with a significant increase in the data variability.

Based in our results, some practices to the ultra‐short HRV measurement are suggested: (a) to use the spontaneous breathing rhythm and apply this same condition to all individuals of the research for ultra‐short measurements duo to the RSA influence on HRV indices; and, (b) to use the rMSSD (orSD1) to perform ultra‐short record analysis as it presents greater reliability at different time intervals.

Important to note, this study has some limitations to be considered. First, the sample size of the present manuscript should be larger and diversified (our sample is only composed by 20 healthy young men) in order to investigate the effect of variables associated to HRV such as age, sex, physical activity, sleep quality, and other behavioral data in ultra‐short HRV indices. Second, in our experiment the standard gold record adopted was of 6‐min duration. We believe that this time interval is sensitive for the assessment of acute conditions (e.g., recovery period after intense physical exercise), although future studies may investigate the stability of the ultra‐short recording indices over a day or a week to verify whereas short‐term recording is suitable only for acute events or can be extrapolated as a stable autonomic marker of the individual. Last, our sample is only composed of healthy individuals, we propose the investigation of ultra‐short HRV recording reliability in clinical samples (e.g., diabetes) compared to healthy individuals, this could encourage the use of HRV in the clinical routine as the short‐term HRV recording represent a noninvasive, low‐cost, and practical technique.

Heart rate variability is a technique to quantify the autonomic activity, being commonly reported as possible biomarker for several diseases, including psychiatric disorders (Laborde et al., 2017; Thayer & Brosschot, 2005; Thayer & Lane, 2000,2009; Thayer et al., 2012). This study encourages the use of ultra‐short HRV recording as complementary resource for clinical reasoning, as the HRV data analysis is based in reliable indices and the assessment carried out in appropriated breathing conditions. The 1‐min recording do not take much time when compared with classical long‐term monitoring systems or require expansive apparatus (to 24 hr recording) doing this technique more accessible to less developed countries. Therefore, the electrocardiogram recording measurement that is commonly performed in the hospital routine can be used for other clinical areas besides the classical use in cardiology, as the use of short‐term registries optimizes the use of the equipment by a clinicians group.

5. CONCLUSION

The rMSSD and SD1 present the best reliability to ultra‐short‐term HRV recordings, being stable for the 1‐, 2‐, and 3‐min recordings. For the measurement of frequency and Poincare geometric measures domains are recommended longer recording periods. The controlled breathing rhythm evokes stronger r values for time domain indices, but impairs the mean values replicability across all time intervals. The comparison between HRV indices during the breathing conditions showed that all HRV indices present an increase in mean values, except the HF that significantly decreased. Therefore, our results suggest that rMSSD (or SD1) and spontaneous breathing rate are more appropriate to measure HRV with the ultra‐short‐term method.

DISCLOSURES

The authors declare no conflict of interest.

ACKNOWLEDGMENTS

This work was supported by PRONEX Program (Programa de Núcleos de Excelência– NENASC Project) of FAPESC‐CNPq‐MS, Santa Catarina Brazil (process number 56802/2010). RW is Researcher Fellow from CNPq (Brazilian Council for Scientific and Technologic Development, Brazil), AAH is supported by scholarship from CAPES/PNPD and HMM is supported by CAPES/DS scholarship.

Melo HM, Martins TC, Nascimento LM, et al. Ultra‐short heart rate variability recording reliability: The effect of controlled paced breathing. Ann Noninvasive Electrocardiol. 2018;23:e12565 10.1111/anec.12565

REFERENCES

  1. Abad, C. , Kobal, R. , Kitamura, K. , Gil, S. , Pereira, L. , Loturco, I. , & Nakamura, F. (2017). Heart rate variability in elite sprinters: Effects of gender and body position. Clinical Physiology and Functional Imaging, 37(4), 442–447. 10.1111/cpf.12331. [DOI] [PubMed] [Google Scholar]
  2. Arroyo‐Carmona, R. E. , López‐Serrano, A. L. , Albarado‐Ibañez, A. , Mendoza‐Lucero, F. M. F. , Medel‐Cajica, D. , López‐Mayorga, R. M. , & Torres‐Jácome, J. (2016). Heart rate variability as early biomarker for the evaluation of diabetes mellitus progress. Journal of Diabetes Research, 2016, 1–8. 10.1155/2016/8483537. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Barth, J. , & Del Vecchio, F. B. (2014). Efeitos da frequência ventilatória sobre os índices da variabilidade da frequência cardíaca. Revista Iberoamericana De Arritmología, 5(1), 185–193. 10.5031/v5i1.RIA10245. [DOI] [Google Scholar]
  4. Beauchaine, T. P. , & Thayer, J. F. (2015). Heart rate variability as a transdiagnostic biomarker of psychopathology. International Journal of Psychophysiology, 98(2), 338–350. 10.1016/j.ijpsycho.2015.08.004. [DOI] [PubMed] [Google Scholar]
  5. Billman, G. E. , Huikuri, H. V. , Sacha, J. , & Trimmel, K. (2015). An introduction to heart rate variability: Methodological considerations and clinical applications. Frontiers in Physiology, 6(2), 2013–2015. 10.3389/fphys.201400177. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Brennan, M. , Palaniswami, M. , & Kamen, P. (2001). Do existing measures of Poincare plot geometry reflect nonlinear features of heart rate variability? IEEE Transactions on Biomedical Engineering, 48(11), 1342–1347. 10.1109/10.959330. [DOI] [PubMed] [Google Scholar]
  7. Chalmers, J. A. , Quintana, D. S. , Abbott, M.‐J.‐A. , & Kemp, A. H. (2014). Anxiety Disorders are Associated with Reduced Heart Rate Variability: A Meta‐Analysis. Frontiers in Psychiatry, 5(7), 1–11. 10.3389/fpsyt.2014.00080. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Chen, J. , Yang, S.‐B. , Liu, J. , & Tang, Z.‐H. (2015). Diagnostic performance analysis for diabetic cardiovascular autonomic neuropathy based on short‐term heart rate variability using Bayesian methods: Preliminary analysis. Diabetology & Metabolic Syndrome, 7(1), 74 10.1186/s13098-015-0070-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Ciccone, A. B. , Siedlik, J. A. , Wecht, J. M. , Deckert, J. A. , Nguyen, N. D. , & Weir, J. P. (2017). Reminder: RMSSD and SD1 are identical heart rate variability metrics. Muscle and Nerve, 56(4), 674–678. 10.1002/mus.25573. [DOI] [PubMed] [Google Scholar]
  10. Dimitropoulos, G. (2014). Cardiac autonomic neuropathy in patients with diabetes mellitus. World J Diabetes, 5(1), 17 10.4239/wjd.v5.i1.17. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Esco, M. R. , & Flatt, A. A. (2014). Ultra‐short‐term heart rate variability indexes at rest and post‐exercise in athletes: Evaluating the agreement with accepted recommendations. Journal of Sports Science and Medicine, 13(3), 535–541. [PMC free article] [PubMed] [Google Scholar]
  12. Flatt, A. A. , & Esco, M. R. (2013). Validity of the ithleteTM Smart Phone Application for Determining Ultra‐Short‐Term Heart Rate Variability. J. Hum. Kinet., 39(1), 85–92. 10.2478/hukin-2013-0071. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Iliescu, R. , Tudorancea, I. , Irwin, E. D. , & Lohmeier, T. E. (2013). Chronic baroreflex activation restores spontaneous baroreflex control and variability of heart rate in obesity‐induced hypertension. AJP: Heart and Circulatory Physiology, 305(7), H1080–H1088. 10.1152/ajpheart.00464.2013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Kamen, P. W. , Krum, H. , & Tonkin, A. M. (1996). Poincaré plot of heart rate variability allows quantitative display of parasympathetic nervous activity in humans. Clinical Science (London, England), 91(2), 201–208. 10.1042/cs0910201. [DOI] [PubMed] [Google Scholar]
  15. Krygier, J. , Heathers, J. , Shahrestani, S. , Abbott, M. , Gross, J. , & Kemp, A. (2013). Mindfulness meditation, well‐being, and heart rate variability: A preliminary investigation into the impact of intensive Vipassana meditation. International Journal of Psychophysiology, 89(3), 305–313. 10.1016/j.ijpsycho.2013.06.017. [DOI] [PubMed] [Google Scholar]
  16. Laborde, S. , Mosley, E. , & Thayer, J. F. (2017). Heart rate variability and cardiac vagal tone in psychophysiological research – recommendations for experiment planning, data analysis, and data reporting. Frontiers in Psychology, 8(8), 1–18. 10.3389/fpsyg.2017.00213. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Lehrer, P. M. , Vaschillo, E. , & Vaschillo, B. (2000). Resonant frequency biofeedback training to increase cardiac variability: Rationale and manual for training. Applied Psychophysiology and Biofeedback, 25(3), 177–191. [DOI] [PubMed] [Google Scholar]
  18. Lehrer, P. M. , Vaschillo, E. , Vaschillo, B. , Lu, S.‐E. , Scardella, A. , Siddique, M. , & Habib, R. H. (2004). Biofeedback treatment for asthma. Chest, 126(2), 352–361. 10.1378/chest.126.2.352. [DOI] [PubMed] [Google Scholar]
  19. Munoz, M. L. , Van Roon, A. , Riese, H. , Thio, C. , Oostenbroek, E. , Westrik, I. , & Snieder, H. (2015). Validity of (Ultra‐) Short recordings for heart rate variability measurements. PLoS ONE, 10(9), 1–15. 10.1371/journal.pone.0138921. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Nakamura, F. Y. , Flatt, A. A. , Pereira, L. A. , Ramirez‐Campillo, R. , Loturco, I. , & Esco, M. R. (2015). Ultra‐short‐term heart rate variability is sensitive to training effects in team sports players. Journal of Sports Science & Medicine, 14(3), 602–605. [PMC free article] [PubMed] [Google Scholar]
  21. Nussinovitch, U. , Elishkevitz, K. P. , Katz, K. , Nussinovitch, M. , Segev, S. , Volovitz, B. , & Nussinovitch, N. (2011). Reliability of ultra‐short ECG indices for heart rate variability. Annals of Noninvasive Electrocardiology, 16(2), 117–122. 10.1111/j.1542-474X.2011.00417.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Pereira, L. A. , Flatt, A. A. , Ramirez‐Campillo, R. , Loturco, I. , & Nakamura, F. Y. (2016). Assessing shortened field‐based heart‐rate‐variability‐data acquisition in team‐sport athletes. International Journal of Sports Physiology and Performance, 11(2), 154–158. 10.1123/ijspp.2015-0038. [DOI] [PubMed] [Google Scholar]
  23. Saboul, D. , Pialoux, V. , & Hautier, C. (2013). The impact of breathing on HRV measurements: Implications for the longitudinal follow‐up of athletes. European Journal of Sport Science, 13(5), 534–542. 10.1080/17461391.2013.767947. [DOI] [PubMed] [Google Scholar]
  24. Salahuddin, L. , Cho, J. , Jeong, M. G. , & Kim, D. (2007). Ultra short term analysis of heart rate variability for monitoring mental stress in mobile settings. Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society 2007, EMBS, 4656–4659. 10.1109/IEMBS.2007.4353378. [DOI] [PubMed] [Google Scholar]
  25. Schmitt, L. , Regnard, J. , & Millet, G. P. (2015). Monitoring fatigue status with HRV measures in elite athletes: An avenue beyond RMSSD? Frontiers in Physiology, 6(11), 2013–2015. 10.3389/fphys.2015.00343. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Tarvainen, M. P. , Niskanen, J.‐P. , Lipponen, J. A. , Ranta‐aho, P. O. , & Karjalainen, P. A. (2014). Kubios HRV – Heart rate variability analysis software. Computer Methods and Programs in Biomedicine, 113(1), 210–220. 10.1016/j.cmpb.2013.07.024. [DOI] [PubMed] [Google Scholar]
  27. Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology (1996). Heart rate variability: Standards of measurement, physiological interpretation, and clinical use. European Heart Journal, 17, 354–381. [PubMed] [Google Scholar]
  28. Tavares, B. S. , de Paula Vidigal, G. , Garner, D. M. , Raimundo, R. D. , de Abreu, L. C. , & Valenti, V. E. (2016). Effects of guided breath exercise on complex behaviour of heart rate dynamics. Clinical Physiology and Functional Imaging, 37(6), 622–629. 10.1111/cpf.12347. [DOI] [PubMed] [Google Scholar]
  29. Thayer, J. F. , Åhs, F. , Fredrikson, M. , Sollers, J. J. , & Wager, T. D. (2012). A meta‐analysis of heart rate variability and neuroimaging studies: Implications for heart rate variability as a marker of stress and health. Neuroscience & Biobehavioral Reviews, 36(2), 747–756. 10.1016/j.neubiorev.2011.11.009. [DOI] [PubMed] [Google Scholar]
  30. Thayer, J. F. , & Brosschot, J. F. (2005). Psychosomatics and psychopathology: Looking up and down from the brain. Psychoneuroendocrinology, 30(10), 1050–1058. 10.1016/j.psyneuen.2005.04.014. [DOI] [PubMed] [Google Scholar]
  31. Thayer, J. F. , & Lane, R. D. (2000). A model of neurovisceral integration in emotion regulation and dysregulation. Journal of Affective Disorders, 61(3), 201–216. [DOI] [PubMed] [Google Scholar]
  32. Thayer, J. F. , & Lane, R. D. (2009). Claude Bernard and the heart‐brain connection: Further elaboration of a model of neurovisceral integration. Neuroscience and Biobehavioral Reviews, 33(2), 81–88. 10.1016/j.neubiorev.2008.08.004. [DOI] [PubMed] [Google Scholar]
  33. Thong, T. , Li, K. , McNames, J. , Aboy, M. , & Goldstein, B. (2003). Accuracy of ultra‐short heart rate variability measures. Proceedings of the 25th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (IEEE Cat. No.03CH37439), 3, 2424–2427. 10.1109/IEMBS.2003.1280405 [DOI] [Google Scholar]
  34. Vaschillo, E. G. , Vaschillo, B. , & Lehrer, P. M. (2006). Characteristics of resonance in heart rate variability stimulated by biofeedback. Applied Psychophysiology and Biofeedback, 31(2), 129–142. 10.1007/s10484-006-9009-3. [DOI] [PubMed] [Google Scholar]
  35. Voss, A. , Schulz, S. , Schroeder, R. , Baumert, M. , & Caminal, P. (2009). Methods derived from nonlinear dynamics for analysing heart rate variability. Philosophical Transactions. Series A Mathematical, Physical, and Engineering Sciences, 367(1887), 277–296. 10.1098/rsta.2008.0232 [DOI] [PubMed] [Google Scholar]
  36. Yadav, R. L. , Yadav, P. K. , Yadav, L. K. , Agrawal, K. , Sah, S. K. , & Islam, M. N. (2017). Association between obesity and heart rate variability indices: An intuition toward cardiac autonomic alteration — a risk of CVD. Diabetes, Metabolic Syndrome and Obesity: Targets and Therapy, 10, 57–64. 10.2147/DMSO.S123935. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Yasuma, F. , & Hayano, J. (2004). Why Does the Heartbeat Synchronize With Respiratory Sinus Arrhythmia. Chest, 125(2), 683–690. 10.1378/chest.125.2.683. [DOI] [PubMed] [Google Scholar]

Articles from Annals of Noninvasive Electrocardiology : The Official Journal of the International Society for Holter and Noninvasive Electrocardiology, Inc are provided here courtesy of International Society for Holter and Noninvasive Electrocardiology, Inc. and Wiley Periodicals, Inc.

RESOURCES