Heart rate variability
Heart rate variability (HRV) refers to the physiological fluctuation in the time intervals between consecutive heartbeats, known as RR intervals on an electrocardiogram, which serves as a noninvasive indicator of autonomic nervous system function and cardiovascular adaptability.[1] This variation arises from the dynamic interplay between the sympathetic and parasympathetic branches of the autonomic nervous system, where parasympathetic (vagal) activity typically increases HRV through slower heart rates and greater beat-to-beat differences, while sympathetic activity reduces it during stress or exertion.[1] Higher HRV is generally associated with better health and resilience to physiological stressors, whereas reduced HRV signals potential autonomic imbalance, increased cardiovascular risk, or underlying disease.[2] HRV can be quantified through various time-domain, frequency-domain, and nonlinear metrics, each providing insights into different aspects of cardiac regulation.[3] Time-domain measures, such as the standard deviation of normal-to-normal intervals (SDNN) or the root mean square of successive differences (RMSSD), capture overall variability and short-term fluctuations, respectively.[3] Frequency-domain analysis decomposes HRV into high-frequency (HF) components (0.15–0.4 Hz), reflecting primarily parasympathetic influence, and low-frequency (LF) components (0.04–0.15 Hz), which involve both sympathetic and parasympathetic modulation.[1] Nonlinear methods, like Poincaré plots or entropy measures, assess the complexity and predictability of heart rate patterns, offering additional prognostic value beyond linear approaches.[3] Normal HRV values vary by age, sex, and measurement duration, with healthy adults typically showing SDNN values above 100 ms during 24-hour recordings, though these norms are influenced by factors such as physical fitness and sleep.[3] Clinically, HRV assessment has established roles in predicting outcomes after myocardial infarction, where depressed HRV indicates higher risk of arrhythmic events or sudden cardiac death.[2] It also serves as a marker for diabetic autonomic neuropathy, with progressive HRV reduction correlating to disease severity and poor prognosis.[2] In heart failure management, low HRV reflects impaired autonomic balance and guides therapeutic interventions like beta-blockers.[2] Beyond cardiology, HRV is applied in neurology to evaluate conditions like Parkinson's disease,[4] in sports science for monitoring training recovery,[5] and in occupational health to gauge stress responses,[6] underscoring its broad utility as a biomarker of overall physiological resilience.Fundamentals
Definition and Measurement
Heart rate variability (HRV) refers to the physiological phenomenon characterized by fluctuations in the time intervals between consecutive heartbeats, reflecting the dynamic regulation of cardiac function.[1] These variations occur naturally even in healthy individuals at rest and are quantified using interbeat intervals, primarily the RR intervals derived from electrocardiogram (ECG) signals or analogous pulse intervals from photoplethysmography (PPG) signals.[1] [7] The RR interval specifically measures the duration between successive R waves, which mark the onset of ventricular depolarization during the cardiac cycle.[8] This cycle begins with electrical activation from the sinoatrial node, propagating to produce the QRS complex as the key detectable feature for interval computation.[1] The measurement of HRV begins with signal acquisition via ECG, which records the heart's electrical activity through electrodes on the skin, or PPG, a non-invasive optical method that detects blood volume changes in peripheral arteries using light transmission or reflection.[1] [9] In ECG, automated algorithms detect QRS complexes to locate R peaks, ensuring accurate identification of normal sinus beats while filtering artifacts or ectopic beats.[10] The time difference between consecutive valid R peaks yields the RR interval series, typically expressed in milliseconds.[10] For PPG, peak detection in the pulsatile waveform provides inter-pulse intervals that approximate RR intervals under controlled conditions.[7] Instantaneous heart rate, which inversely relates to these intervals, is computed using the formula: \text{[HR](/page/HR) (bpm)} = \frac{60}{\text{[RR](/page/RR) (s)}} where RR is the interval duration in seconds.[1] This yields the beats per minute equivalent for each beat, highlighting how shorter intervals correspond to higher rates. The clinical significance of HRV was first recognized in 1965 by Hon and Lee, who observed that diminished beat-to-beat variations in fetal heart rate patterns preceded episodes of fetal distress during labor.[1] Their work using electronic fetal monitoring laid the groundwork for HRV as a non-invasive indicator of physiological stress.[1]Physiological Basis
Heart rate variability (HRV) arises primarily from the dynamic interplay between the sympathetic and parasympathetic branches of the autonomic nervous system (ANS), which modulate the sinoatrial node's pacemaker activity to adapt cardiac output to physiological demands. The parasympathetic branch, mediated by the vagus nerve, promotes heart rate deceleration and increases beat-to-beat variability through acetylcholine release, enhancing cardiac adaptability during rest or recovery. In contrast, the sympathetic branch accelerates heart rate via norepinephrine, typically reducing variability to support heightened arousal or stress responses. This antagonistic balance allows HRV to serve as a noninvasive indicator of ANS integrity and cardiovascular flexibility.[1][11] A prominent parasympathetic influence on HRV is respiratory sinus arrhythmia (RSA), characterized by cyclic fluctuations in heart rate synchronized with breathing: acceleration during inspiration due to reduced vagal inhibition and deceleration during expiration from increased vagal tone. This mechanism optimizes pulmonary blood flow by matching cardiac output to respiratory demands, minimizing unnecessary heartbeats and enhancing gas exchange efficiency. Beyond respiration, the baroreflex contributes to HRV by detecting arterial pressure changes and eliciting rapid parasympathetic or sympathetic adjustments to maintain hemodynamic stability, while thermoregulation modulates variability through ANS-mediated responses to temperature shifts, such as vasoconstriction or sweating that indirectly alter cardiac rhythm.[12][13][14] Central nervous system inputs further orchestrate HRV via integrated control from the hypothalamus and brainstem, where nuclei like the nucleus tractus solitarius process sensory afferents and relay signals to modulate ANS outflow. The hypothalamus integrates environmental and emotional cues to fine-tune sympathetic and parasympathetic activity, while brainstem centers coordinate reflexive responses. Conceptually, HRV can be expressed as a function of vagal (high-frequency) and sympathetic (low-frequency) modulation, where overall variability reflects their relative contributions:\text{HRV} \approx f(\text{vagal modulation}, \text{sympathetic modulation})
This balance enables adaptive responses without delving into specific spectral decompositions. From an evolutionary standpoint, HRV likely evolved as a survival mechanism, allowing organisms to flexibly adjust cardiac function to environmental stressors, predators, or resource availability, thereby optimizing energy allocation and resilience in variable conditions.[15][16]
Analysis Techniques
Time-Domain Methods
Time-domain methods quantify heart rate variability (HRV) by applying statistical techniques directly to the series of normal-to-normal (NN) RR intervals extracted from electrocardiogram (ECG) recordings, providing measures of the magnitude and distribution of beat-to-beat fluctuations without decomposing the signal into frequency components.[17] These approaches are particularly valued for their computational simplicity and applicability to both short- and long-term recordings, as they do not require assumptions of signal stationarity.[17] A primary global index is the standard deviation of all NN intervals (SDNN), which captures overall HRV by assessing the dispersion of RR intervals around their mean, encompassing both sympathetic and parasympathetic influences over various time scales.[17] It is computed using the formula: \text{SDNN} = \sqrt{\frac{\sum_{i=1}^{N} (RR_i - \overline{RR})^2}{N}} where RR_i represents each NN interval, \overline{RR} is the mean NN interval, and N is the total number of intervals analyzed.[17] SDNN is widely used in 24-hour Holter monitoring to evaluate long-term variability, with values typically ranging from 50 ms in healthy adults during short recordings to over 140 ms in extended assessments.[17] For short-term variability, the root mean square of successive differences (RMSSD) measures the square root of the mean squared differences between adjacent NN intervals, serving as a robust indicator of high-frequency, parasympathetically mediated fluctuations.[17] The formula is: \text{RMSSD} = \sqrt{\frac{\sum_{i=1}^{N-1} \text{diff}_i^2}{N-1}} where \text{diff}_i = RR_i - RR_{i+1}.[17] RMSSD is less affected by respiratory influences compared to other metrics and correlates strongly with vagal tone, often yielding values around 30-50 ms in resting healthy individuals.[17] The percentage of adjacent NN intervals differing by more than 50 ms (pNN50) provides a binary threshold-based estimate of parasympathetic activity, counting the proportion of successive differences exceeding this criterion.[17] This metric complements RMSSD by highlighting episodic beat-to-beat changes, with normal values exceeding 10-20% in short-term analyses.[17] Geometric measures like the triangular index offer an alternative view of overall variability by constructing a histogram of NN intervals and computing the ratio of the integral of the density distribution to its maximum height, effectively approximating the mode of the distribution.[17] This index, similar to SDNN in scope, is robust to artifacts and typically ranges from 10-50 in healthy subjects, providing a visual and statistical summary of interval distribution.[17] The mean NN interval itself, while not a variability measure, contextualizes absolute heart rate alongside these indices, as higher means (longer intervals) often accompany greater variability in healthy states.[17] Overall, time-domain methods excel in ease of implementation and interpretability but are limited in resolving contributions from specific physiological oscillators, such as respiratory or baroreflex influences.[17] They are commonly applied to 5-minute ECG segments for short-term HRV evaluation in clinical settings, such as assessing autonomic balance during rest.[17]Frequency-Domain Methods
Frequency-domain methods analyze heart rate variability (HRV) by decomposing the RR interval time series into its frequency components, providing insights into the periodic oscillations influenced by autonomic nervous system activity. This approach quantifies the power spectral density (PSD) of the HRV signal, which represents the distribution of signal power across different frequencies. The RR tachogram is typically interpolated to a uniform sampling rate (e.g., 4 Hz) to enable spectral analysis, as unevenly spaced data from electrocardiograms require resampling for accurate Fourier-based methods.[1] PSD estimation can be performed using non-parametric techniques, such as the fast Fourier transform (FFT) combined with Welch's method, which segments the signal, applies windowing (e.g., Hanning window) to reduce spectral leakage, and averages periodograms for improved stability. Alternatively, parametric methods like autoregressive (AR) modeling fit a model to the data to estimate the spectrum, offering higher resolution for short recordings but requiring selection of model order (typically 10-20 for HRV). Both approaches assume signal stationarity, necessitating detrending to remove low-frequency trends and ensure the mean is zero before transformation.[1][18] The HRV power spectrum is divided into standard frequency bands: very low frequency (VLF, 0.003-0.04 Hz), low frequency (LF, 0.04-0.15 Hz), and high frequency (HF, 0.15-0.4 Hz). The VLF band reflects ultra-slow oscillations possibly linked to thermoregulation and peripheral vasomotor activity, though its physiological origins remain unclear. The LF band is modulated by both sympathetic and parasympathetic influences, including baroreflex activity, while the HF band primarily represents parasympathetic activity through respiratory sinus arrhythmia (RSA), where heart rate fluctuations synchronize with breathing cycles. Total power is calculated as the variance of the RR intervals (integral of PSD from 0 to 0.4 Hz), serving as an overall measure of HRV.[1][19] Key metrics include absolute powers in each band (in ms²/Hz) and normalized units to facilitate comparisons. Normalized LF (LFnu) is computed as: \text{LFnu} = \frac{\text{LF}}{\text{LF} + \text{HF}} \times 100 with HFnu = 100 - LFnu, expressing the relative contribution of LF to total spectral power excluding VLF. The LF/HF ratio is often used as an index of sympathovagal balance, where higher values suggest sympathetic dominance. However, this interpretation has caveats, as LF power is not exclusively sympathetic but includes significant parasympathetic contributions, and the ratio can be influenced by respiratory rate changes, limiting its specificity as a pure balance measure.[1][20] For reliable analysis, short-term recordings (5 minutes) in supine rest are recommended to capture LF and HF components adequately, with longer durations (e.g., 24 hours) needed for VLF. These methods enable real-time monitoring of autonomic imbalance, such as reduced HF in heart failure patients indicating parasympathetic withdrawal, or altered LF/HF in diabetes reflecting early autonomic neuropathy, aiding in risk stratification and therapeutic evaluation.[1][21]Nonlinear and Geometric Methods
Nonlinear and geometric methods provide advanced tools for analyzing heart rate variability (HRV) by modeling its chaotic, fractal, and deterministic properties, which arise from the interplay of multiple physiological control systems. These techniques complement linear approaches by revealing hidden structures in the RR interval time series, such as long-range correlations and irregularity, that are indicative of the heart's adaptive dynamics. Unlike time- or frequency-domain methods, they emphasize non-stationarity and complexity without assuming Gaussian distributions. The Poincaré plot is a fundamental geometric method that visualizes HRV as a scatterplot of successive RR intervals (RR_{n+1} versus RR_n), often forming an elliptical shape in healthy individuals due to short- and long-term fluctuations. The minor axis dispersion, denoted SD1, quantifies short-term beat-to-beat variability and is related to the root mean square of successive differences (RMSSD) by the formula \text{SD1} = \frac{\text{RMSSD}}{\sqrt{2}}, reflecting parasympathetic influences. The major axis dispersion, SD2, captures longer-term variability and approximates the standard deviation of all normal-to-normal intervals (SDNN) via \text{SD2} \approx \sqrt{2 \cdot \text{SDNN}^2 - \text{SD1}^2}. The ratio SD1/SD2 further indicates sympathovagal balance, with higher values in healthy states. This method, validated in early studies on autonomic modulation, offers a simple yet powerful visual and quantitative assessment of HRV geometry.[22] Detrended fluctuation analysis (DFA) quantifies long-range correlations in non-stationary HRV signals by integrating the RR series to form a random-walk-like profile, segmenting it into non-overlapping windows, detrending each locally via least-squares fitting, and computing the root-mean-square fluctuations as a function of window size. The scaling behavior follows a power law, F(n) ~ n^α, where the Hurst-like exponent α characterizes correlation strength: α ≈ 0.5 indicates uncorrelated white noise, while α ≈ 1 reflects persistent 1/f scaling typical of healthy HRV, signifying self-similar fractal patterns. Originally applied to HRV in seminal work demonstrating persistent correlations in healthy hearts versus anti-correlations in certain conditions, DFA excels at detecting scale-invariant properties over short recordings (e.g., 20 minutes). Sample entropy (SampEn) assesses the intrinsic irregularity and complexity of the RR series, providing a model-independent measure less biased than its predecessor, approximate entropy (ApEn). Defined for a time series of length N, embedding dimension m, and tolerance r (typically 0.15–0.25 times the standard deviation), SampEn is calculated as \text{SampEn}(m, r, N) = -\ln \left( \frac{A^m(r)}{B^m(r)} \right), where B^m(r) is the average number of matches between vectors of length m that remain close (within r) when extended to m+1 dimensions for A^m(r). Higher SampEn values denote greater irregularity, as seen in healthy HRV compared to reduced values in rigid systems; this entropy avoids self-matches, enhancing reliability for physiological data. Introduced to refine complexity estimation in time series like HRV, SampEn has become widely adopted for its robustness to finite data lengths.[23] Recurrence quantification analysis (RQA) probes deterministic patterns in HRV by reconstructing the phase space from the RR series using time-delay embedding and generating a recurrence plot—a binary matrix marking states closer than a threshold ε. Key metrics include recurrence rate (density of recurrent points), determinism (fraction of recurrent points forming diagonal lines, indicating predictability), and laminarity (proportion on vertical lines, reflecting intermittent laminar states). These quantify transitions between chaotic and ordered dynamics, revealing subtle nonlinear structures in physiological signals. Originating from recurrence plot visualization, RQA's application to HRV highlights recurrent motifs linked to autonomic regulation, offering insights into system stability.[24] These nonlinear and geometric methods offer distinct advantages over linear techniques by capturing nonlinearity, fractal scaling, and dynamical determinism that variance-based measures overlook, such as amplitude-independent complexity in HRV morphology. For instance, they better differentiate adaptive versus rigid physiological states through irregularity and correlation exponents, enabling detection of subtle transitions in system behavior with shorter data segments. Post-2020 advancements have integrated these features—e.g., DFA exponents and SampEn—with machine learning algorithms like random forests for automated pattern recognition, improving prognostic accuracy in real-time monitoring applications.[25][26]Clinical Applications
Disease Associations
Heart rate variability (HRV) alterations, particularly reductions in overall variability, serve as important biomarkers for various pathological conditions, reflecting disruptions in autonomic nervous system balance and increased cardiovascular risk. In many diseases, decreased HRV indices such as standard deviation of normal-to-normal intervals (SDNN) and root mean square of successive differences (RMSSD) indicate sympathetic dominance or parasympathetic withdrawal, correlating with disease severity and prognosis. These changes are often detected using time- and frequency-domain methods, providing non-invasive insights into autonomic dysfunction. In cardiovascular diseases, HRV is markedly reduced following acute myocardial infarction (AMI), where low SDNN values predict mortality. Specifically, an SDNN below 50 ms is associated with a 5.3-fold increased relative risk of death, as demonstrated in a seminal cohort study of post-AMI patients. In heart failure, particularly with reduced ejection fraction, high-frequency (HF) power is significantly lowered, indicating parasympathetic impairment and poorer prognosis independent of other risk factors. Low very low-frequency (VLF) power has also been linked to heightened risk of sudden cardiac death in susceptible populations, with VLF below 18 ms² elevating the relative risk by over 3.5-fold. Neurological conditions further illustrate HRV's role as a biomarker. In diabetic autonomic neuropathy, respiratory sinus arrhythmia (RSA) is impaired, and RMSSD is reduced, signaling early parasympathetic dysfunction and correlating with disease progression in type 1 and type 2 diabetes. Patients with tetraplegia due to spinal cord injury exhibit sympathetic denervation, leading to decreased low-frequency (LF) power and overall HRV loss compared to able-bodied individuals, with tetraplegic subjects showing LF values approximately 50% lower during postural challenges. Beyond cardiovascular and neurological disorders, HRV reductions occur in systemic conditions like sepsis, where an initial sharp drop in parameters such as SDNN and triangular interpolation of NN intervals (TINN) precedes multiple organ dysfunction and mortality, with SDNN ≤17 ms identifying non-survivors. In liver cirrhosis, vagal withdrawal manifests as decreased time- and frequency-domain HRV indices, inversely correlating with disease severity scores like Child-Pugh and independently predicting mortality. Chemotherapy for cancer induces autonomic changes, including reduced HRV and vagal activity, as seen in anthracycline-treated patients where both time-domain (e.g., SDNN) and frequency-domain (e.g., HF) measures decline post-treatment, potentially signaling cardiotoxicity. Mental health disorders, such as depression and anxiety, are associated with altered HRV profiles, including elevated LF/HF ratios reflecting sympathetic overactivity. In major depressive disorder, LF/HF increases at rest and during stress, correlating with symptom severity and distinguishing depressed patients from controls. Notably, in pregnancy, HRV patterns differ; during the third trimester, overall variability and HF power generally decrease due to autonomic adaptations, though this can be further altered in complicated cases. Recent research highlights HRV's utility in emerging conditions. In long COVID (post-acute sequelae of SARS-CoV-2), persistent low HRV—particularly reduced SDNN and RMSSD—is common in survivors, associating with fatigue and autonomic dysautonomia up to two years post-infection, as evidenced in studies from 2021 to 2024. For early neurodegeneration, wearable devices detect HRV declines in Parkinson's disease, with reduced time-domain metrics like RMSSD emerging as potential preclinical markers in 2023 cohort analyses, aiding non-invasive screening.Intervention Effects
Various pharmacological interventions modulate heart rate variability (HRV) to promote autonomic balance, particularly by enhancing parasympathetic tone. β-blockers, such as propranolol and atenolol, increase time-domain measures like the root mean square of successive differences (RMSSD) while reducing low-frequency (LF) power in the frequency domain, reflecting reduced sympathetic dominance and improved vagal activity in patients with hypertension or post-myocardial infarction (MI).[27][28] Scopolamine, acting as a muscarinic agonist, enhances high-frequency (HF) power—particularly in the 0.25-Hz respiratory band—through cholinergic stimulation, as observed in normal subjects and those with congestive heart failure after 24 hours of transdermal application.[29][30] Antiarrhythmic drugs produce variable effects on nonlinear HRV measures, with agents like encainide, flecainide, and moricizine altering approximate entropy and sample entropy to differentiate arrhythmia suppression, though outcomes depend on the specific drug and patient rhythm status.[31] Thrombolytic therapy following acute MI rapidly restores autonomic function, with HRV parameters improving within hours of reperfusion; for instance, standard deviation of normal-to-normal intervals (SDNN) rises significantly in the short term, correlating with better long-term prognosis compared to non-thrombolyzed patients.[32] Non-pharmacological approaches, such as aerobic exercise training, effectively enhance HRV by increasing HF power and overall vagal modulation, with systematic reviews indicating improvements in RMSSD and HF components after 3 months of moderate-intensity training in healthy adults and those with cardiovascular risk factors.[33][34] Heart rate variability biofeedback (HRVB), particularly coherence training involving resonant breathing, promotes autonomic balance by elevating HF power and reducing the LF/HF ratio, thereby fostering emotional regulation and stress resilience in chronic disease management.[35][36] Playing wind instruments mimics respiratory sinus arrhythmia (RSA) through controlled exhalation. In heart transplant recipients, surgical denervation initially results in profoundly low HRV due to absent autonomic innervation, but partial recovery occurs over years through reinnervation, with increases in HF oscillations during sleep reflecting mechanical respiratory influences on the sinoatrial node.[37][38] Psychological interventions like meditation and yoga increase nonlinear HRV complexity, as evidenced by elevated sample entropy (SampEn) in practitioners compared to controls, indicating greater dynamic adaptability in heart rate dynamics following regular sessions.[39] Emerging digital therapeutics, such as HRV-guided mobile apps, deliver biofeedback training remotely and have shown feasibility in trials from 2022 onward, improving vagal activity and subjective well-being in young adults and those with stress-related conditions through short daily sessions. In 2025, HRV biofeedback has demonstrated efficacy in reducing negative affect and supporting treatment for substance use disorders.[40][41][42] Non-invasive vagus nerve stimulation, particularly transcutaneous auricular methods, boosts RSA and overall HRV by shifting toward parasympathetic predominance, with 2023-2024 studies confirming increased time- and frequency-domain metrics in response to stimulation.[43][44]Advanced Considerations
Normal Values and Variability Factors
Heart rate variability (HRV) exhibits well-established reference ranges in healthy populations, which vary by measurement duration and specific metrics. For short-term recordings (approximately 5 minutes), the standard deviation of normal-to-normal intervals (SDNN) typically averages around 50 ms in young adults, while root mean square of successive differences (RMSSD) averages about 42 ms, and high-frequency (HF) power is approximately 975 ms². In contrast, 24-hour recordings yield higher values, with SDNN often exceeding 100 ms in healthy individuals under 50 years, dropping to around 70-80 ms in those over 70. These norms are derived from large systematic reviews of healthy cohorts, emphasizing the need to distinguish short-term (resting) from long-term (ambulatory) assessments for accurate interpretation.[45][3] Age is a primary determinant of HRV decline in healthy individuals, reflecting reduced autonomic flexibility over time. Short-term SDNN decreases progressively from approximately 50 ms in individuals in their 20s to about 25-30 ms by the 70s, while HF power roughly halves every decade after age 30 due to diminished parasympathetic tone. This age-related reduction follows a linear pattern for overall variability (e.g., SDNN reaching 46% of young adult levels by the ninth decade) but is more pronounced for vagal indices like RMSSD, which drop rapidly in early adulthood before stabilizing. Large cohort studies, including those spanning nine decades, confirm these trends in non-diseased populations, highlighting the importance of age-stratified norms.[46][47] Sex differences in HRV are observed, with mixed findings across studies; some report higher parasympathetic activity in females (e.g., greater HF power), potentially attributed to estrogen's modulatory effects on vagal tone during reproductive years, while others show higher values in males or no significant differences. This disparity often diminishes post-menopause.[48] Circadian rhythms further modulate HRV, with values peaking at night (often 2-3 times higher than daytime levels) due to vagal dominance during sleep, as evidenced by increased HF power and RMSSD in nocturnal segments of 24-hour recordings.[49] HRV is inversely proportional to mean heart rate (HR), with a logarithmic relationship where log(SDNN) ≈ -0.5 × log(HR), meaning variability diminishes as HR rises due to physiological constraints on interbeat intervals. Lifestyle factors amplify this: endurance athletes exhibit 20-50% higher HRV (e.g., elevated SDNN and HF) than sedentary peers, reflecting enhanced autonomic balance from chronic training. Conversely, smoking reduces HF power by 15-20% in habitual users, impairing vagal modulation via nicotine's sympathoexcitatory effects. Ethnic variations exist but are less pronounced, with some studies noting slightly higher baseline HRV in individuals of African ancestry.[50] The Autonomic Tone and Reflexes After Myocardial Infarction (ATRAMI) study provides key 24-hour norms from large cohorts, establishing SDNN >70 ms as a healthy threshold, though short-term values require separate standardization. Recent 2020s data from wearables like Apple Watch, analyzed in population-scale studies, offer percentile benchmarks: an SDNN of 36 ms represents the median, with values below 18 ms in the 10th percentile and above 76 ms in the 90th, enabling real-time healthy range assessments.[51][52][53][54]| Factor | Effect on HRV | Example Metric Change | Source |
|---|---|---|---|
| Age (per decade after 30) | Decline | HF power halves; SDNN ↓ ~10-20% | [46] |
| Sex (females vs. males) | Mixed; some higher in females | HF ↑ in some studies | [48] |
| Circadian (night vs. day) | Higher at night | Overall HRV ↑ 2-3× | [49] |
| Mean HR (log relation) | Inverse | log(SDNN) ≈ -0.5 log(HR) | [51] |
| Athletes (vs. sedentary) | Higher | SDNN/HF ↑ 20-50% | [52] |
| Smoking (habitual) | Reduced | HF ↓ 15-20% | [53] |