Content uploaded by Marek Malik
Author content
All content in this area was uploaded by Marek Malik on Mar 13, 2017
Content may be subject to copyright.
Heart rate variability: Standards of measurement, physiological interpretation, and clinical use
... Features should be chosen based on clinical or physiological motivations to enhance the contextual interpretation of model performance [16], [22]. Since HRV research offers well-established guidelines for key parameters [23], [24], comparing ML models using both selection methods is essential for detecting mental workload, and these should be evaluated alongside XML techniques. ...
... HRV, the variation in time between heartbeats, is influenced by the balance between SNS and PNS activity [24]. Under increased mental workload, SNS activity rises, leading to an elevated heart rate and reduced HRV as the body becomes more alert and engaged with the task. ...
... HRV analysis can be performed in the time, frequency, and non-linear indices [24], [23], [28]. Time-domain measures assess the variability in the intervals between normal heartbeats (normal-to-normal intervals), frequency-domain measures quantify power distribution across different frequency bands using spectral analysis, and non-linear indices capture the complexity and unpredictability of heart rate time series. ...
Detecting mental workload (MWL) is essential to optimize task performance, prevent cognitive overload, and improve well-being and safety, especially in contexts involving human-computer interaction (HCI) and complex task environments. This study introduces a novel data set for MWL detection using accessible physiological signals, specifically heart rate variability (HRV) and galvanic skin response (GSR), collected from 36 participants engaged in switch-and arithmetic tasks designed to induce varying levels of mental workload. To classify binary MWL states, we trained various machine learning (ML) algorithms, including Multi-Layer Perceptron (MLP), Gradient Boosting (GB), and Support Vector Machine (SVM). This approach provided significant insights for optimizing human-computer interaction (HCI) systems. To improve model performance, we employed three different feature selection techniques: correlation-based, minimum redundancy maximum relevancy (mRMR), and domain-expert selection. We optimized hyperparameters using grid search cross-validation (CV) and validated the results through nested CV. Among the models, the MLP with correlation-based feature selection demonstrated the highest performance, reaching an area under the curve (AUC) of 0.822. The GB and SVM models also performed well with mRMR and domain-expert feature selection (AUC = 0.760 and 0.741, respectively). To provide interpretability and a better understanding of feature importance in HCI contexts, Shapley Additive exPlanations (SHAP) identified several GSR features and heart rate as key predictors of mental workload, offering critical insights for designing adaptive HCI systems that account for cognitive load.
... Existe un delicado equilibrio entre los dos componentes del SNA a saber sistema simpático (SNS) y parasimpático (SNP). De la interracción entre ambos resulta la variabilidad de la frecuencia cardíaca (VFC), fluctuaciones latido a latido de la FC [1], [2]. Que se analiza a partir de las señales del electrocardiograma (ECG) o de fotopletismografía (PPG), conformando una serie temporal noequiespaciada de latidos RR, analizada, generalmente, en el dominio temporal, gráfico de Poincaré y frecuencial [1], [3], [4], [5]. ...
... De la interracción entre ambos resulta la variabilidad de la frecuencia cardíaca (VFC), fluctuaciones latido a latido de la FC [1], [2]. Que se analiza a partir de las señales del electrocardiograma (ECG) o de fotopletismografía (PPG), conformando una serie temporal noequiespaciada de latidos RR, analizada, generalmente, en el dominio temporal, gráfico de Poincaré y frecuencial [1], [3], [4], [5]. ...
... Recientes trabajos, relacionados con dispositivos de monitoreo, entre otros métodos, optimizan la frecuencia de muestreo de la señal de ECG o PPG, para incrementar la duración de sus baterias [9], [10], [11], [12], [13], [14], [15], [16]. Muy pocos de estos trabajos hacen referencia al procedimiento de remuestreo necesario para el equi-espaciamiento de la serie RR previo a su análisis de VFC en el dominio frecuencial [1], [17], [18]. Esta frecuencia de remuestreo (Fr), por un lado, debe ser lo suficientemente alta para evitar distorsiones dentro del rango de frecuencias donde el SNA tiene una respuesta significativa, que es alrededor de 1 segundo (1Hz, en frecuencia) y por otro lado, su valor aporta a la eficiencia energética y computacional [17], [19], [20], [21]. ...
Beat-to-beat variations in heart rate lead to heart rate variability (HRV), analysed from electrocardiogram or photoplethysmography signals, forming a non-equispaced time series of beats, which requires a resampling of 3 Hz or 4 Hz, for analysis in the frequency domain. HRV is considered a biomarker, predictor of the evolution of diseases in intensive care units (ICU). To enhance these HRV studies, it is necessary to monitor the patients health using portable devices, from admission to the ICU until discharge from it and subsequently at home. This requires monitoring devices that can minimise energy consumption and data storage. Reducing the sampling frequency in HRV can reduce energy consumption, computing power and to limite data storage. Therefore, the objective of this work is to prove that a series resampled at 1 Hz allows obtaining HRV indices, equivalent to a 3 Hz. Through concordance analysis, using a database of subjects with pharmacological autonomic blockade and postural changes. The results show equivalences between the indices, standard deviation (SDNN), total spectral power (PT), low frequency (LF) and long-term variability (SD2) and agree with those reported as predictors. This study has limitations, since only a small number of young men participated. Future studies should consider this. The reduction of SDNN, PT, LF values would be predictors of mortality in hospitals, so the equivalence found from series with 1Hz resampling, would allow the use of portable devices with optimized performance, to monitor the evolution of the disease in patients in ICUs.
... Changes in the psychophysiological state of individuals have been observed in various digital contexts on several physiological measures. Indicators such as heart rate, the standard deviation of intervals between normal R-peaks (SDNN), the root mean square of successive differences (RMSSD) of N-N intervals, and the power in low and high-frequency bands and its ratio (LF, HF, LF/HF) are often used to investigate changes in the physiological state of individuals 78,79 . Findings show rather negative effects of using digital devices on Heart Rate Variability (HRV) 80,81 . ...
... Various measures were computed from the raw EDA and ECG signals collected throughout the experiment. HRV measures such as HR, RMSSD, LF/HF were computed 78,80,[85][86][87][88]98 . The mean tonic level and the number of non-specific skin conductance responses (NS-SCRs) per minute (also called frequency) were selected as EDA indicators 23 ...
The technological advances in recent years are influencing and redefining our daily lives, communications, and social relationships. While these advances bring us many benefits, their negative effects may also cause concern. Although often studied, the potential benefits of digital deprivation are still disputed. This laboratory study investigates the impact of short digital deprivation (7 min and 30 s) on the psychophysiological state and time perception of 90 participants. Three experimental conditions were created for the task performed during the waiting period (30 subjects per condition). Participants had to either freely use their smartphone, perform a non-digital task (sudoku), or wait (i.e. passive digital deprivation). Indicators of electrodermal activity and heart rate variability were calculated for the baseline and waiting periods, along with measures of subjective affective state. Four measures of time perception were also collected after the waiting period. Regardless of their experimental condition, the participants underestimated the duration of the waiting period on average (5 min 44 vs. 7 min 30). Passive digitally deprived participants felt that the time passed more slowly and were more bored than participants engaged in a task, regardless of whether the task was digital or not. Sudoku induced more positive affect and was more cognitively engaging than the free use of a smartphone regarding heart rate variability measures. The results suggest that performing a digital task (free smartphone use) is less cognitively demanding than a non-digital task (sudoku) and alters time perception in the same way. The digital nature of a task might also impact one’s affective reaction. A similar study in the field with longer or repetitive digital deprivation periods and a different non-digital task to perform (e.g., reading news) should be conducted to confirm the results obtained in this study.
... For HRV analysis we used methods based on time-domain, spectral-domain and Poincaré plot 32 . From various time-domain indices we report the standard deviation of NN intervals (SDNN, ms), a measure validated in short-term recordings 33 with a documented prognostic significance, and the proportion of NN50 (the number of times successive heartbeat intervals exceed 50ms divided by the total number of NN intervals, pNN50, %) -a marker of parasympathetic phasic activity under resting conditions [34][35][36] . SDNN was calculated using Nevrokard BPV version 9.0.0 (Nevrokard Kiauta, Izola, Slovenia; https://www.nevrokard.eu) and pNN50 with LabChart Pro (ADInstruments) from 5 min period of an acceptable quality taken from the baseline recording and from the recording after atropine administration. ...
The autonomic regulation of heart rate (HR) reactivity to acute hypoxia remains unclear. Parasympathetic cardioneuroablation (PCNA) may serve as a novel model for the analysis of physiological consequences of reduced vagal influence over sinus node in humans. We studied 11 adult patients scheduled for PCNA for the treatment of vasovagal syncope. HR reactivity to hypoxia was studied before and after PCNA with brief nitrogen gas administrations. Each test was followed by an atropine challenge to evaluate the contribution of parasympathetic tone to the resting HR. Additionally, we assessed changes in cardiac baroreflex sensitivity and HR variability following the procedure. PCNA led to partial parasympathetic denervation of sinus node at rest (67.0 ± 20.1%). This translated into a significant change in HR reactivity to hypoxia (0.58 ± 0.21 vs. 0.22 ± 0.13 beats min− 1%SpO2 − 1, p = 0.0001) which was proportional to the degree of cardiac vagal denervation (R = 0.76, p = 0.01). There was no change in peak HR on atropine following PCNA implying unchanged sympathetic input to sinus node. This suggests that HR reactivity to acute hypoxia is significantly influenced by parasympathetic system. Additionally, despite incomplete vagal denervation PCNA resulted in profoundly depressed HR variability and cardiac baroreflex sensitivity. The clinical meaning of the latter should be explored in further studies.
The variation of heart rate in phase with breathing, known as ‘respiratory sinus arrhythmia’ (RSA), is a physiological phenomenon present in all air-breathing vertebrates. RSA arises from the interaction of several physiological mechanisms but is primarily mediated by rhythmic changes in cardiac parasympathetic (vagal) activity, increasing heart rate during inspiration and decreasing heart rate during expiration. RSA amplitude is an indicator of autonomic and cardiac health; RSA is diminished or absent in common pathological conditions such as chronic heart failure and hypertension. In this Expert Recommendation, we argue that the term ‘RSA’, although historically important, is semantically inaccurate and carries misleading pathological connotations, contributing to misunderstanding and misinterpretation of the origin and the physiological importance of the phenomenon. We propose replacing ‘RSA’ with the term ‘respiratory heart rate variability’ (RespHRV), which avoids pathological connotations and emphasizes the specific respiratory contribution to heart rate variability. We clarify that RespHRV encompasses respiratory-related heart rate variations in both the low-frequency and high-frequency bands traditionally defined in heart rate variability analysis, and that its amplitude should not be misconstrued as a measure of vagal tone. Adopting the proposed term ‘RespHRV’ is expected to unify understanding and stimulate further experimental and clinical research into the physiological mechanisms and functional importance of this phenomenon.
l. Introduction
Mathematical modeling is the process of using mathematics to represent, describe, and analyze real-world situations or systems. It involves creating equations, functions, or algorithms that simulate how things work in reality, so we can better understand them, make predictions, or find solutions.
It’s like building a bridge between the messy, complicated real world and the clean, logical world of math — so we can test ideas, make forecasts, and solve problems in a controlled, simplified way.
Human Heart Rate Variability (HRV) is a complex, dynamic biomarker reflecting the autonomic nervous system's regulation of cardiac function.
Recent advances in mathematical modeling of HRV have allowed for more in-depth studies of both healthy and pathological cardiovascular conditions.
Modern approaches leverage fractal analysis, stochastic processes, nonlinear dynamics, and time–frequency domain analyses to interpret HRV signals. These models are increasingly applied in stress testing, neurocardiology, sleep research, and wearable health monitoring.
Mathematical frameworks like fractal calculus, Markov models, entropy measures, and spectral analysis now help capture both short-term and long-term HRV fluctuations, crucial for predicting arrhythmias, cardiac stress responses, and cognitive workload. Innovative applications, such as integrating HRV modeling into real-time health gaming environments or AI-based diagnostic systems, mark an exciting future for precision cardiovascular medicine.
II. Key Research Works on this topic are presented.
In conclusion
Current HRV models aim to replicate HRV characteristics by simulating the statistical properties of measured HRV signals. These models play a vital role in predicting cardiovascular events and understanding the complexity of heart rate regulation.
The integration of mathematical modeling approaches with HRV data analysis offers valuable insights into cardiovascular health and can enhance the early detection of cardiac abnormalities.
Scientific research in this area continues intensively.
Mitral valve prolapse (MVP) is a common valvular abnormality especially in young female subjects. Recently, Boudulas had classified the patients with MVP into two types: the 1st type is MVP-anatomic (MVPa) with valvular malfunction and the 2nd MVP-syndrome (MVPs) without valvular disease. In the 2nd type, several authors have reported the occurrence of functional abnormalities of neuroendocrine or autonomic systems. Also, the syncope called 'neuromediated' has been found and is probably associated with modification of the autonomic nervous system. To understand the role of the autonomic nervous system, 39 female patients with MVPs (group 1) and 24 normal females as control group (group 2) were evaluated by a spectral analysis of heart rate variability (HRV) in time and frequency domains. The analysis between the two groups showed a low vagal tone in group 1 resulting in a lower high-frequency normalized unit (HF-NU) in comparison with the control group 2 (p < 0.03). Group 1 was divided into two subgroups with and without syncope. The analysis of HRV by Anova indicated no differences for all parameters between the two subgroups of MVPs patients. These results demonstrate that (1) MVPs has a low vagal tone; (2) the syncope in MVPs could not be only in relation to the autonomic nervous system malfunction, but probably other causes underlined this phenomenon.
Activation of the sympathetic nervous system has been extensively studied in patients with chronic heart failure, but the parasympathetic nervous system has received relatively little attention. The objective in this study was to investigate cardiac parasympathetic activity in chronic heart failure and to explore its relation to left ventricular function.
Heart rate variability was measured from 24 hour ambulatory electrocardiograms by counting the number of times each RR interval exceeded the preceding RR interval by more than 50 ms (counts). This method provided a sensitive index of cardiac parasympathetic activity.
Mean (range) of counts were: waking 48 (1-275)/h, sleeping 62 (0-360)/h, and total 1310 (31-7278)/24 h. These were lower than expected, and in 26 (60%) of the 43 patients counts fell below the lower 95% confidence intervals (95% CI) for RR counts in normal subjects. A significant correlation between total 24 hour RR counts and left ventricular ejection fraction was present (r = 0.49, p less than 0.05).
These results indicate that most patients with chronic heart failure have reduced heart rate variability and therefore reduced cardiac parasympathetic activity. The degree of parasympathetic dysfunction is related to the severity of left ventricular dysfunction. This may be relevant to the high incidence of ventricular arrhythmias and poor prognosis of patients with chronic heart failure.
Ambulatory ECG is an evolutionary cardiovascular discipline with regard to both its clinical applications and technology. With the technologic advances of the past decade, powerful personal computers have expanded the traditional clinical role of ambulatory ECG examination for cardiac arrhythmias to detection of ST-segment changes, heart-rate variability, QT interval measurement, and the signal-averaged ECG. These expanded clinical applications increase the practicality of a 24-hour ambulatory ECG on an individual cardiac disease patient and make the 24-hour ambulatory ECG one of the most cost-effective cardiovascular diagnostic tests when used properly. The current clinical developments make valid ST-segment detection and quantitation readily available and heart-rate variability an easily applied investigative tool to a wide spectrum of diseases. Automatic QT interval measurements and signal averaging of ambulatory ECG are just technologically emerging and can be expected to develop more in the near future.
The intracellular mechanisms by which cardiac Ca current (ICa) and the delayed outward K current (IK) are modulated during beta-adrenergic or muscarinic stimulation were investigated at the level of both single-channel and whole-cell currents in single ventricular myocytes of guinea-pigs. Superfusion of cells with beta-adrenergic agonist increased the amplitude of whole-cell ICa in a dose-dependent manner. In the single-channel recording, neither the amplitude of elementary current nor the total number of active channels was affected but the number of blank records was markedly reduced resulting in a larger amplitude of the ensemble average current. Intracellular dialysis of cells with cyclic AMP (cAMP) or the catalytic (C) subunit of cAMP-dependent protein kinase (cAMP-PK) produced a dose-dependent increase in the amplitude of ICa and IK. A non-hydrolysable ATP analogue, AMP-PNP, reduced whereas ATP gamma S enhanced the effects of beta-agonist on ICa and IK, suggesting an involvement of protein phosphorylation during the enhancement of these currents. The regulatory subunit of cAMP-PK, the heat-stable protein-kinase inhibitor (PKI) and type-1 protein phosphatase antagonized the beta-adrenergic enhancement of ICa and IK, but did not eliminate ICa. Acetylcholine (ACh) reduced the amplitude of ICa when ICa was enhanced by either beta-adrenergic agonist, forskolin or 3-isobutyl-1-methyl-xanthine but did ACh not when ICa was enhanced by intracellular dialysis with cAMP or C subunit, suggesting that muscarinic inhibition occurs at the level of adenylate cyclase. Non-hydrolysable GTP analogue, GMP-PNP, uncoupled both beta-adrenergic and muscarinic modulation of ICa. Pertussis toxin selectively eliminated the effect of ACh on ICa. Based on these results, we concluded that the activities of the Ca channel and the delayed outward K channel are controlled by the action of neurotransmitters, which are mediated by GTP-binding proteins and cAMP-dependent protein phosphorylation. It is suggested that phosphorylation of 'Ca-channel-related protein' leads to an increased open probability without changing the total number of channels or the elementary current amplitude.
A detailed computerized analysis of heart rate (HR) behavior during ambulatory monitoring was performed in 19 long-term survivors of cardiac transplantation (12 orthotopic, 7 heterotopic) and in 10 normal volunteers. Compared with normal hearts, the transplanted denervated and the recipient innervated hearts in the heterotopic group showed higher average HRs over the whole 24 hours, the waking and sleeping periods. The rates of the denervated and the recipient hearts did not differ significantly. The maximal HR was similar in all 3 groups, but the minimal HR was lower in the normal hearts than in denervated and recipient hearts. The 24-hour RR interval variability was greatly reduced in the denervated hearts (60 +/- 22 ms) compared with the recipient (89 +/- 26 ms) and normal (151 +/- 38 ms) hearts. During arousal from sleep, the magnitude, time and rate of the HR change were significantly reduced in the denervated hearts compared with the innervated hearts and in the recipient compared with the normal hearts. These findings demonstrate that denervated hearts can show significant variations in HR during the 24-hour period, but to a lesser extent than normal hearts. The widest deviation from normal occurs during sleep, when the denervated heart shows minimal slowing probably due to lack of parasympathetic innervation. In heterotopic transplant recipients, the patients' own HR responses also are significantly blunted.
1. Voltage-clamp experiments were conducted on small specimens of rabbit sinoatrial node. In the same preparation the dose-response curve of the potassium current induced by application of different concentrations of acetylcholine (ACh), the time constant of relaxation and the current fluctuations were measured. From these measurements the apparent dissociation constant and the rate constants for the opening and closing of the ACh-activated potassium channel were estimated. 2. In the presence of neostigmine a measurable response was recorded at around 10(-8) M ACh, the saturation was reached at 10(-4) M, and the half saturation was attained at around 10(-6) M. 3. The time constant of relaxation at --35 mV decreased from 100 ms at 10(-8) M to 45 ms at 10(-4) M ACh. 4. The variance of the fluctuations of the ACh-activated current increased with increasing ACh concentration to a peak value of around 10(-5) M. 5. From the above 3 kinds of measurements, opening and closing rate constants of about 12s-1 and 10s-1, respectively, and a dissociation constant of 1.7 microM were calculated. 6. The Katz-Miledi model was considered to be appropriate to describe the reaction of ACh with the muscarinic receptor in the S-A node. 7. The current on ionophoretic application of ACh was computed using the rate constants and taking into account diffusion in the S-A node in which the density of receptors is low. The computed response had a similar time course to the recorded current.