Establishing normative values for short-term heart rate variability indices in healthy infants in the emergency department

Introduction

Heart rate variability (HRV) is the term used to describe the differences in time intervals between consecutive heartbeats (1), which provides insights into the functioning of their autonomic nervous system (ANS) and ability to adapt. It has been recognised that HRV in the paediatric populations could be a significant indicator of various physiological and pathophysiological processes (2,3).

Studies suggest that by analysing HRV parameters, clinicians can gain insights into the balance between sympathetic and parasympathetic tone, which can reflect different medical conditions or potential complications in critically ill patients (4), such as the use of heart rate characteristic (HRC) index monitoring to determine the fold-increase of developing neonatal sepsis in the neonatal intensive care unit (NICU) (3) and changes in the low/high frequency (LF/HF) ratio for children with septic shock in the paediatric intensive care unit (PICU) (5). Moreover, HRV can potentially help predict the short- and long-term outcomes in specific conditions (6,7). Several studies show that through monitoring HRV, clinicians can potentially identify infants who may require additional support or early intervention (8). Therefore, the clinical use of HRV in infants can be crucial in the early identification of potential issues and in guiding appropriate interventions.

The measurement context, including recording period length, subject age, and sex, on baseline affect HRV values (9,10). Twenty-four-hour, short-term (5 minutes), and ultra-short-term (<5 minutes) normative values are not interchangeable. Specifically, the use of short-term and ultra-short-term HRV indices have been used in acute settings for practical reasons to identify critical illness (11).

One of the more common concerns and presentations for neonates and young infants in the emergency department (ED) was for fever and evaluation for serious bacterial infection. Intrinsically the risk stratification differs for infants <28 days (neonatal), 28 to <60 days and 60 to 90 days of life. Thus, we initiated the evaluation of short-term HRV indices in this age group and aimed to provide normative values. Currently, there is a wide variance in types of HRV analysis used and although there have been attempts to standardise the methodology, there are limited studies analysing infants up till 1 month old (3,12). Use of HRV patterns in adult patients who present with early sepsis at the ED has also been shown to predict progressive organ dysfunction and deterioration (13). To our knowledge there were very few studies using short-term HRV for infants beyond the neonatal age group in the ED setting (outpatient setting). Thus, there is novelty and practical applications for ambulatory use of HRV in infants <90 days of life. Explorative utilisation of short-term HRV indices as a clinical adjunct in the evaluation of serious bacterial infection febrile neonates and young infants has been published (4,8).

The aim of this study was to evaluate short-term HRV indices in awake, healthy young infants >48 hours and ≤90 days of life and to establish a reference range. We also aimed to produce a clinical calculator that can be used in this population for evaluation of short-term HRV variables in young infants in the ED setting that can be potentially used in future clinical validation and research. We present this article in accordance with the STROBE reporting checklist (available at https://atm.amegroups.com/article/view/10.21037/atm-24-180/rc).

Methods

Study design, population, and recruitment

We performed a prospective observational study of well young infants (≤90 days old) presenting to the paediatric emergency department (PED) of KK Women’s and Children’s Hospital in Singapore, between November 2017 and February 2021. KK Women’s and Children’s Hospital is a referral centre and has an annual ED attendance of about 150,000 children. In this study, we identified and obtained consent for electrocardiogram (ECG) recording and follow-up from a convenience sample of well infants >48 hours to ≤90 days old during office hours and the recruited patients were followed up till discharge.

We excluded infants with concerns of fever, significant respiratory distress or had injuries. We also excluded all infants with a history of non-sinus rhythm due to potential confounding on the HRV analysis, and preterm infants <35 weeks’ gestation. In neonates, abnormal HRV can be affected by factors other than sepsis, such as gestational age and underlying medical conditions (14,15) and as such, needs to be interpreted accordingly, hence premature neonates were excluded from the study. Thirty-five to 36 weeks’ gestational age neonates, although considered as late pre-terms, have been found to have low absolute risks for severe neonatal morbidities; they are typically managed in the maternity wards like term infants and are mostly discharged well after birth (16). Hence, including these late pre-term neonates will ensure that our aim to establish normative values for our population of interest can be interpreted more appropriately and makes the results more generalisable.

All infants with abnormal vital signs at triage or on any repeat evaluation, abnormal investigations suggesting significant pathologies (for example serious bacterial infections, severe viral infections and viral meningitis) or had significant pathological diagnoses on the infants’ discharge were also excluded.

Eligible infants were recruited in the ED by trained research coordinators.

The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013). The study was approved by the SingHealth Centralised Institutional Review Board in Singapore (No. CIRB 2017/2680). Informed written consent was also taken from the parents of all the young infants.

HRV application and processing

We recorded their heart rate (HR) by using the PMD-250 (TIIM Healthcare SG, Singapore, Singapore) to capture ECG data over 5 minutes. The PMD-250 patient monitor is non-invasive, colourised, portable and easy to use. Three-lead ECG electrodes were placed on the infant’s chest while the ECG signals were continuously measured for at least 5 minutes. Lead-II ECG signal was wirelessly transmitted in real-time via Bluetooth to an Android-based app on a tablet for display and recording purposes. The portable ECG device PMD-250 and corresponding Android-based App interface were shown in Figure 1.

Figure 1 Portable ECG device and attached three-lead ECG electrodes (left) and Android-based App interface (TIIM Healthcare SG) for real-time Lead-II ECG signal display and recording (right). DIA, diastolic blood pressure; ECG, electrocardiogram; HR, heart rate; LA, left arm; LL, left leg; NIBP, non-invasive blood pressure; RA, right arm; RESP, respiratory rate; SpO₂, peripheral arterial oxygen saturation; SYS, systolic blood pressure; TEMP, temperature.

To ensure accurate measurements and to correct for motion artefacts, additional steps were taken to ensure proper and secure placement of the ECG electrodes. The recruited infants were either placed supine in a cot or held supine in caregivers’ arms to optimise comfort. The research assistant was also trained to inspect the ECG tracing in real-time to ensure that sufficient good quality ECG tracings were obtained.

Regular quality control measures were implemented to check for any technical issues or anomalies in the data collection process. This involved monitoring electrode/sensor integrity, ensuring proper contact with the skin, and evaluating the overall signal quality.

For all patients in this study, the recorded 5-minute duration of Lead II ECG signal was subsequently exported for HRV analysis, where a customised software, HRnV-Calc (developed by Duke-NUS research group), was used for HRV analysis (17). HRnV-Calc is an open-source software based on the PhysioNet Cardiovascular Signal Toolbox (PCST) (18), featuring comprehensive step-by-step graphical user interfaces (GUIs). Wavelet de-noising and bandpass filtering with a finite impulse response (FIR) filter of order 41 (bandwidth between 0.25 and 30 Hz) to remove higher frequencies and baseline wander were employed as noise reduction techniques in HRnV-Calc as infants may exhibit movements or external interference that can introduce noise into the ECG. As shown in Figure 2, after noise removal, R peaks were automatically detected by HRnV-Calc software and visually inspected independently by study members (D.G. and R.P.). Where there were artefacts or difficulty in identifying R peaks, manual assignment of peaks were performed by study team members with consensus from the 3 individuals (D.G., R.P. and S.L.C.). This important work has also led to an improved prototype that can now handle fast HRs and presence of artefacts better. R-R intervals (RRIs) were then generated based on the confirmed R peaks for further HRV parameters derivation.

Figure 2 HRnV-Calc software with step-by-step graphical user interfaces developed by Duke-NUS research group (17).

We reported HRV indices (HRV time-domain measures, HRV frequency-domain measures and HRV non-linear measures) commonly used in clinical studies (1,19,20). According to Task Force of The European Society of Cardiology and the North American Society of Pacing & Electrophysiology guidelines (19), HRV parameters were calculated in time domain, frequency domain and non-linear domain using HRnV-Calc. The following standard time-domain measures were derived: (I) standard deviation of NN intervals (SDNN); (II) root mean square of successive differences (RMSSD); (III) NN50 and pNN50: denotes RRIs differing more than 50 ms from the preceding one and their percentage; and (IV) triangular index, which is the integral of the density distribution (that is, the number of all RRIs) divided by the maximum of the density distribution. For frequency-domain measures, HRnV-Calc following PCST uses the Lomb Periodogram as the default method since it has a superior performance in handling unevenly sampled RRIs (21,22). After the power spectral density (PSD) is calculated by Lomb Periodogram, various frequency domain HRV measures were calculated where the frequency bands of interest for analysing HRV are defined as: very low frequency (VLF) (0–0.04 Hz), LF (0.04–0.15 Hz), HF (0.15–0.40 Hz) and total power (TP) (0.04–0.4 Hz). Frequency domain HRV measures include (I) frequency of peaks in each band ranges, i.e., VLF peak, LF peak, HF peak, (II) power in each band range and total band range, i.e., VLF milliseconds squared (MS), LF MS, HF MS and TP MS, (III) LF and HF power in normalised units (NU), LF NU and HF NU, (IV) ratio of LF power and HF power, i.e., LF/HF ratio. For non-linear domain measures, Poincare SD1 and Poincare SD2 from Poincare plot, approximation entropy (APP ENT) and sample entropy (SAM ENT) are derived.

HRV normalisation and analysis

There is a need to ensure that short-term HRV metrics truly reflect the variability in ANS activity and are not confounded by changes in the underlying HR itself. This is done by correcting HRV indices for HR to remove the dependency of HRV on HR, a mathematical adjustment known as HRV normalisation. To investigate the impact of HR on HRV, correlation analysis of standard HRV parameters with average HR was performed and Spearman correlation coefficient, r was reported. For those HRV parameters that were moderately (0.3≤ |r| <0.5) or strongly correlated (0.5≤ |r| ≤1) with average HR, they were corrected by the division of standard HRV indices by different powers of their corresponding average RRI (23).

The general formula for time and frequency domain HRVs is

ModHRVIndex=(HRVIndexaRR)n[1]

where aRR is the average RRI in milliseconds.

n=0 for VLF peak, LF peak, LF NU, HF peak, HF NU, LF/HF ratio, APP ENT and SAM ENT (these HRV parameters do not have HR dependence hence there is no need for correction).

n=2 for RMSSD, NN50, PNN50, triangular index, VLF MS, TP MS, Poincare SD1 and Poincare SD2.

n=4 for LF MS and HF MS.

For HRV parameters with low correlation (0≤ |r| <0.3) with HR, no correction was needed.

Statistical analysis

Demographic and clinical characteristics were presented as median and interquartile range (IQR) for continuous variables, and frequency and percentage for categorical variables. The Kolmogorov-Smirnov test was used to assess the normality of data distribution. Since all HRV parameters did not exhibit normal distribution, which were similarly reported in the literature (10,24,25), the data was presented as median and IQR.

Correlation of HRV and normalised HRV parameters with HR were reported via Spearman correlation coefficient. Correlation plots of HRV and modified HRV parameters with HR were done to illustrate presence of any correlation before and after normalisation of HRV.

Subsequently, correlation of normalised HRV parameters with age was also investigated. If no strong correlation (i.e., |r| ≥0.5) was found, then for each normalised HRV, the median and 90^th percentile reference range were calculated overall for all ages of infants. The 90% confidence interval for the lower 5^th percentile and upper 95^th percentile limits of the reference range were provided as well to show the precision in estimating the lower and upper limits with our present sample size. This was also analysed using Cohen’s f² where according to Cohen’s guidelines, f² ≥0.02, f² ≥0.15, and f² ≥0.35 represent small, medium, and large effect sizes, respectively (26). All analysis was performed using Microsoft Office 365 Excel and MedCalc statistical software, version 22.

Results

There were 120 eligible infants aged >48 hours to ≤90 days of life during the study period (November 2017 to February 2021) who had short-term HRV recordings performed after consent was obtained. Five infants (4.12%) did not have satisfactory ECG tracings for analyses or had incomplete data. Another seven infants (5.83%) who were initially identified as afebrile with minor concerns were excluded as they developed fever and/or serious bacterial infections on prospective follow-up. A final cohort of 108 infants (90.0%) were included in the study. Please refer to Figure 3.

Figure 3 STROBE flow-chart of participants. ECG, electrocardiogram; HRV, heart rate variability; STROBE, Strengthening the Reporting of Observational Studies in Epidemiology.

The patient characteristics were as shown in Table 1.

The result of correlation analysis of standard HRV parameters with HR and their Pearson’s correlation coefficient (r) and corresponding P values are shown in Table 2.

Figure 4 depicts the monotonic relationship between some HRV parameters and HR (correlation coefficients −0.615 to −0.286, P<0.05), and the Spearman correlation efficient was deemed as the most suitable measure for reporting the non-linear relationship between the two variables.

Figure 4 Correlations coefficients and corresponding P values between standard HRV parameters and HR. HF, high frequency; HR, heart rate; HRV, heart rate variability; LF, low frequency; MS, milliseconds squared; NN50, number of pairs of successive normal-to-normal inter-beat intervals that differ by more than 50 milliseconds; pNN50, proportion (or percentage) of pairs of successive normal-to-normal inter-beat intervals that differ by more than 50 milliseconds; RMSSD, root mean square of successive difference; TP, total power; VLF, very low frequency.

Following the normalisation formula listed in Table 2, HR dependence for these HRV parameters was successfully removed. There was no significant correlation between the normalised HRV parameters and HR; P>0.05 (Figure 5).

Figure 5 Correlations coefficients and corresponding P values between modified HRV parameters and HR. HF, high frequency; HR, heart rate; HRV, heart rate variability; LF, low frequency; Mod, modified; MS, milliseconds squared; NN50, number of pairs of successive normal-to-normal inter-beat intervals that differ by more than 50 milliseconds; pNN50, proportion (or percentage) of pairs of successive normal-to-normal inter-beat intervals that differ by more than 50 milliseconds; RMSSD, root mean square of successive difference; TP, total power; VLF, very low frequency.

For HRV parameters that did not correlate with the average HR, no normalisation was needed thus the corrected HRV parameters were kept the same as the original HRV parameters.

The corrected HRV parameters were not strongly associated with age, with all correlations being low (0≤ |r| <0.3) or moderate (0.3≤ |r| <0.5) (Table 3).

This is also demonstrated by using Cohen’s f² as a measure of the local effect of age on normalised HRV parameters within a regression model of modified HRV on age (27), whereby age represented a negligible, small or medium effect size for all modified HRV parameters as shown in Table 3. Thus, for each normalised HRV parameter, the 90^th percentile range with 90% confidence interval for the lower and upper limits of the reference range are summarised for the whole group of infants of all ages in Table 4.

Based on the above results of our study, we produced a clinical short-term HRV calculator that can be used to evaluate short-term HRV indices in awake, healthy young infants >48 hours and ≤90 days of life in the ED setting that can be potentially used in future clinical validation and research (material available at https://cdn.amegroups.cn/static/public/atm-24-180-1.xlsx).

Discussion

HRV has been explored as a potential tool for assessing ANS function and identifying critical conditions or prognostic indicators in paediatric patients (8,28). Despite its promise, there is a lack of well-designed studies focusing specifically on short-term HRV in paediatric populations, necessitating further validation of its clinical relevance in the ED setting. Establishing normative values for HRV in paediatric populations is challenging due to age-related variations, requiring accurate reference values for interpretation.

This study contributed short-term normative HRV data for well neonates and young infants (<90 days) in the ED and has the potential to facilitate future research by serving as a reference range. To facilitate this, we produced a clinical calculator for evaluation of the short-term HRV indices of young infants in ambulatory settings to assess if they fall within the normative range after correction. The potential clinical application of the use of non-invasive monitoring for evaluation in paediatric populations, especially unwell young infants, would be very attractive.

The brief duration needed for HRV analysis makes it a practical, non-invasive tool for quick assessment in EDs, especially for infants. Portable and user-friendly HRV monitoring devices facilitate rapid data collection especially in the paediatric population. Short-term HRV measurements, when integrated with other clinical parameters, may contribute to a more comprehensive assessment of paediatric patients in emergency settings, providing valuable insights into autonomic function dysregulation associated with serious conditions such as serious bacterial infections (4,8). This may potentially aid in identifying young infants at higher risk of adverse events or requiring closer monitoring.

In our study, when considering HR, we found that sex and age (potentially due to the limitations of the narrow age range in our cohort) did not contribute significantly to establishing normative short-term HRV. Hence, we calculated HRV independent of HR by correcting standard HRV indices. We did not consider respiratory rates in our analysis as it was found in earlier studies that the influence of respiratory rates on HRV appeared to be, at least in part, HR-dependent and after HRV correction for HR, the correlation between respiratory rates and HRV significantly decreased (24,29).

While we included VLF component in our study, we recognise that there have been concerns about its reliability as a short-term HRV parameter (1,19). However, VLF has been associated with thermoregulation and vasomotor activity (30-32). The association of thermoregulation and vasomotor tone may be useful in studying sepsis (including evolving and early sepsis). It is also unknown if trending short-term HRV may be useful after the initial ED disposition and interventions. We feel that it is useful to provide short-term normative values that may be of potential use for further study and validation.

Establishing normative values for short-term HRV in neonates and young infants could potentially benefit future clinical research in the ED by providing a baseline comparison which may help identify deviations that may signify underlying health issues. With a clearer understanding of what constitutes normal HRV ranges, clinicians can potentially explore means of detecting abnormalities that may indicate early stages of disease or physiological stress, allowing for timely intervention in this specific and more vulnerable age group in the ED setting. Furthermore, establishing normative values forms a foundational dataset that can be used in various research studies. It may also facilitate the comparison of data across different studies and populations, enhancing the robustness of research findings.

We have identified a few limitations to this study. As the study cohort is small, the findings from the clinical calculator for evaluation of short-term HRV will need to be externally validated before the results can be generalised. Additionally, the findings of our study may potentially be limited to a predominantly Asian young infant population. We advise caution when in the evaluation of normative short-term HRV values for non-Asian populations as it has been reported that Asian and Caucasian children display different frequency domain components of HRV (33).

We also recognise that standardising short-term HRV measurement protocols is crucial for reliable results. Inconsistencies in data collection methods, including recording duration and analysis techniques, can impact the interpretation of HRV in paediatric studies. While a specialised portable HRV monitor and recording machine were utilised in this study, future research should explore the feasibility of alternative means, such as physiological monitors, ECG recordings, or plethysmography. While there have been interests in the evaluation of NN10 and NN20 in neonates and infants (3), we instead used NN50, a standard time-domain measure that is often reported in all paediatric ages especially in the PED. Although we did not explore this in our current study, we agree that using more neonate and infant specific domains such as NN10 and NN20 will be useful and relevant for future studies involving this population.

Conclusions

We studied short-term HRV indices in awake, healthy young infants >48 hours and ≤90 days of life in the ED to establish a reference range. We also produced a clinical calculator that can be used in this population for evaluation of short-term HRV variables that can be potentially used in future clinical validation and research.

Acknowledgments

We would like to thank Dr. Liu Nan, Centre for Quantitative Medicine and Programme in Health Services and Systems Research, Duke-NUS Medical School, for his expert input and assistance in the review of the manuscript.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://atm.amegroups.com/article/view/10.21037/atm-24-180/rc

Data Sharing Statement: Available at https://atm.amegroups.com/article/view/10.21037/atm-24-180/dss

Peer Review File: Available at https://atm.amegroups.com/article/view/10.21037/atm-24-180/prf

Funding: This work was supported by the National Medical Research Council (Singapore) (No. CNIG18may-0002 to S.L.C.).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://atm.amegroups.com/article/view/10.21037/atm-24-180/coif). D.G. is a shareholder of TIIM Healthcare Singapore. S.L.C. reports grant funding from the National Medical Research Council (Singapore) (No. CNIG18may-0002) for investigator-initiated research. M.E.H.O. is a Scientific Advisor to TIIM Healthcare SG and holds TIIM Healthcare SG patents related to using HRV for risk stratification. He is the Director, Health Services Research Institute, SingHealth-Duke NUS, Academic Medical Center; Director, Health Services Research Center, SingHealth Services; Director, Unit for Pre-hospital Emergency Care, Ministry of Health, Singapore; and Vice-Chair, SingHealth Duke-NUS Emergency Medicine Academic Clinical Programme (EM ACP). The other authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013). The study was approved by the SingHealth Centralised Institutional Review Board in Singapore (No. CIRB 2017/2680). Informed written consent was also taken from the parents of all the young infants.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Shaffer F, Ginsberg JP. An Overview of Heart Rate Variability Metrics and Norms. Front Public Health 2017;5:258. [Crossref] [PubMed]
2. Oliveira V, von Rosenberg W, Montaldo P, et al. Early Postnatal Heart Rate Variability in Healthy Newborn Infants. Front Physiol 2019;10:922. [Crossref] [PubMed]
3. Latremouille S, Lam J, Shalish W, et al. Neonatal heart rate variability: a contemporary scoping review of analysis methods and clinical applications. BMJ Open 2021;11:e055209. [Crossref] [PubMed]
4. Chong SL, Ong GY, Allen JC, et al. Early prediction of serious infections in febrile infants incorporating heart rate variability in an emergency department: a pilot study. Emerg Med J 2021;38:607-12. [Crossref] [PubMed]
5. Wee BYH, Lee JH, Mok YH, et al. A narrative review of heart rate and variability in sepsis. Ann Transl Med 2020;8:768. [Crossref] [PubMed]
6. Marsillio LE, Manghi T, Carroll MS, et al. Heart rate variability as a marker of recovery from critical illness in children. PLoS One 2019;14:e0215930. [Crossref] [PubMed]
7. Piantino JA, Lin A, Crowder D, et al. Early Heart Rate Variability and Electroencephalographic Abnormalities in Acutely Brain-Injured Children Who Progress to Brain Death. Pediatr Crit Care Med 2019;20:38-46. [Crossref] [PubMed]
8. Chong SL, Niu C, Piragasam R, et al. Adding heart rate n-variability (HRnV) to clinical assessment potentially improves prediction of serious bacterial infections in young febrile infants at the emergency department: a prospective observational study. Ann Transl Med 2023;11:6. [Crossref] [PubMed]
9. Kazuma N, Otsuka K, Wakamatsu K, et al. Heart rate variability in normotensive healthy children with aging. Clin Exp Hypertens 2002;24:83-9. [Crossref] [PubMed]
10. Sharma VK, Subramanian SK, Arunachalam V, et al. Heart Rate Variability in Adolescents - Normative Data Stratified by Sex and Physical Activity. J Clin Diagn Res 2015;9:CC08-13. [Crossref] [PubMed]
11. Shaffer F, Meehan ZM, Zerr CL. A Critical Review of Ultra-Short-Term Heart Rate Variability Norms Research. Front Neurosci 2020;14:594880. [Crossref] [PubMed]
12. Claiborne A, Williams A, Jolly C, et al. Methods for analyzing infant heart rate variability: A preliminary study. Birth Defects Res 2023;115:998-1006. [Crossref] [PubMed]
13. van Wijk RJ, Quinten VM, van Rossum MC, et al. Predicting deterioration of patients with early sepsis at the emergency department using continuous heart rate variability analysis: a model-based approach. Scand J Trauma Resusc Emerg Med 2023;31:15. [Crossref] [PubMed]
14. Chi D, Guo W, Liu J, et al. Advancements in understanding the association of sepsis with heart rate variability in premature infants. Egyptian Pediatric Association Gazette 2024;72:37.
15. Coggins SA, Weitkamp JH, Grunwald L, et al. Heart rate characteristic index monitoring for bloodstream infection in an NICU: a 3-year experience. Arch Dis Child Fetal Neonatal Ed 2016;101:F329-32. [Crossref] [PubMed]
16. Mitha A, Chen R, Altman M, et al. Neonatal Morbidities in Infants Born Late Preterm at 35-36 Weeks of Gestation: A Swedish Nationwide Population-based Study. J Pediatr 2021;233:43-50.e5. [Crossref] [PubMed]
17. Niu C, Guo D, Ong MEH, et al. HRnV-Calc: A Software for Heart Rate n-Variability and Heart Rate Variability Analysis. Journal of Open Source Software. 2023;8:5391.
18. Vest AN, Da Poian G, Li Q, et al. An open source benchmarked toolbox for cardiovascular waveform and interval analysis. Physiol Meas 2018;39:105004. [Crossref] [PubMed]
19. Heart rate variability: standards of measurement, physiological interpretation and clinical use. Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology. Circulation 1996;93:1043-65.
20. Pham T, Lau ZJ, Chen SHA, et al. Heart Rate Variability in Psychology: A Review of HRV Indices and an Analysis Tutorial. Sensors (Basel) 2021;21:3998. [Crossref] [PubMed]
21. Lomb NR. Least-squares frequency analysis of unequally spaced data. Astrophysics and Space Science 1976;39:447-62.
22. Clifford GD, Tarassenko L. Quantifying errors in spectral estimates of HRV due to beat replacement and resampling. IEEE Trans Biomed Eng 2005;52:630-8. [Crossref] [PubMed]
23. Sacha J, Barabach S, Statkiewicz-Barabach G, et al. How to strengthen or weaken the HRV dependence on heart rate--description of the method and its perspectives. Int J Cardiol 2013;168:1660-3. [Crossref] [PubMed]
24. Gąsior JS, Sacha J, Jeleń PJ, et al. Interaction Between Heart Rate Variability and Heart Rate in Pediatric Population. Front Physiol 2015;6:385. [Crossref] [PubMed]
25. Gąsior JS, Sacha J, Pawłowski M, et al. Normative Values for Heart Rate Variability Parameters in School-Aged Children: Simple Approach Considering Differences in Average Heart Rate. Front Physiol 2018;9:1495. [Crossref] [PubMed]
26. Cohen J, Cohen P, West SG, et al. Applied Multiple Regression/Correlation Analysis for the Behavioral Sciences. 3rd edition. New York, NY, USA: Routledge; 2002.
27. Cohen J. A power primer. Psychol Bull 1992;112:155-9. [Crossref] [PubMed]
28. Javorka K, Lehotska Z, Kozar M, et al. Heart rate variability in newborns. Physiol Res 2017;66:S203-14. [Crossref] [PubMed]
29. Gąsior JS, Sacha J, Jeleń PJ, et al. Heart Rate and Respiratory Rate Influence on Heart Rate Variability Repeatability: Effects of the Correction for the Prevailing Heart Rate. Front Physiol 2016;7:356. [Crossref] [PubMed]
30. Coenen AJ, Kitney RI, Rompelman O. Thermally elicited entertainment of heart rate variability [proceedings] J Physiol 1977;270:5P-6P.
31. Bouzida N, Bendada A, Maldague XP. Visualization of body thermoregulation by infrared imaging. Journal of Thermal Biology 2009;34:120-6.
32. Machado C, Estevez M, Perez-Nellar J, et al. Residual vasomotor activity assessed by heart rate variability in a brain-dead case. BMJ Case Rep 2015;2015: [Crossref] [PubMed]
33. Reed KE, Warburton DE, Whitney CL, et al. Differences in heart rate variability between Asian and Caucasian children living in the same Canadian community. Appl Physiol Nutr Metab 2006;31:277-82. [Crossref] [PubMed]