Heart Rate Variability Classification Research Articles

A Hybrid Feature Dimension Reduction and SVM for a Classification of Heart Rate Variability

A Hybrid Feature Dimension Reduction and SVM for a Classification of Heart Rate Variability

Comparative Study of Machine Learning Algorithms in Classifying HRV for the Driver’s Physiological Condition

Heart Rate Variability (HRV) may be used as a psychological marker to assess drivers’ states from physiological signals such as an electrocardiogram (ECG), electroencephalogram (EEG), and photoplethysmography (PPG). This paper reviews HRV acquisition methods from drivers and machine learning approaches for driver cardiac health based on HRV classification. The study examines four publicly available ECG datasets and analyzes their HRV features, including time domain, frequency domain, short-term measures, and a combination of time and frequency domains. Eight machine learning classifiers, namely K-Nearest Neighbor, Decision Tree, Naive Bayes, Linear Discriminant Analysis, Support Vector Machine, Random Forest, Gradient Boost, and Adaboost, were used to determine whether the driver's state is normal or abnormal. The results show that K-Nearest Neighbor and Decision Tree classifiers had the highest accuracy at 92.86%. The study concludes by assessing the performance of machine learning algorithms in classifying HRV for the driver's physiological condition using the Man-Whitney U test in terms of accuracy and F1 score. We have statistical evidence to support that the prediction quality is different when HRV analysis applies these three sets: (i) time domain measures or frequency domain measures; (ii) frequency domain measures or short-term measures; and (iii) combining time and frequency domains or only frequency domains. Doi: 10.28991/CEJ-2023-09-09-013 Full Text: PDF

Telemedical transport layer security based platform for cardiac arrhythmia classification using quadratic time–frequency analysis of HRV signal

The heart rate variability signal is a valuable tool for cardiovascular system diagnostics. Processing this signal detects arrhythmia during long-term cardiac monitoring. It is also analyzed to recognize abnormalities within the autonomic nervous system. Processing this signal helps in detecting various pathologies, such as atrial fibrillation (AF), supraventricular tachycardia (SVT), and congestive heart failure (CHF). As a beneficial alternative to the commonly used HRV spectrum analysis, quadratic time–frequency analysis of HRV signals could be helpful in heart pathology detection. Indeed, in this study, we have created a client-server paradigm deployed as a telemedical platform for real-time remote monitoring of the cardiovascular function in patients suffering from arrhythmia. This platform detects arrhythmia in real-time by deploying time–frequency analysis, feature extraction, feature selection, and classification of Heart Rate Variability (HRV) signals. We gathered all these functionalities in a Graphical User Interface (GUI) in addition to data acquisition. As a client, a Raspberry Pi Zero ensures data acquisition and connects to a server over TCP/IP that involves a 4G/3G connection encrypted through the transport layer security (TLS). This telemedical tool continuously monitors the heart rate variability. In the case of an alarm, medical professionals may immediately interact with their patients in the hospital or at home.

Deriving electrophysiological phenotypes of paroxysmal atrial fibrillation based on the characteristics of heart rate variability

Aim. To analyze heart rate variability of patients with paroxysmal atrial fibrillation and identify electrophysiological phenotypes of the disease by using methods of exploratory analysis of twenty-four-hour electrocardiographic (Holter) recordings.
 Methods. 64 electrocardiogram recordings of patients with paroxysmal atrial fibrillation were selected from the open Long-Term Atrial Fibrillation Database (repository PhysioNet). 52 indices of heart rhythm variability were calculated for each recording, including new heart rate fragmentation and asymmetry indices proposed in the last 5 years. Data analysis was carried out with machine learning methods: dimensionality reduction with principal component analysis, hierarchical clustering and outlier detection. Feature correlation was checked by the Pearson criterion, the selected patients subgroups were confirmed by using MannWhitney and Student's tests.
 Results. For the vast majority of patients with paroxysmal atrial fibrillation, heart rate variability can be described by five parameters. Each of these parameters captures a distinct approach in heart rate variability classification: dispersion characteristics of interbeat intervals, frequency characteristics of interbeat intervals, measurements of heart rate fragmentation, indices based on heart rate asymmetry, mean and median of interbeat intervals. Two large phenotypes of the disease were derived based on these parameters: the first phenotype is a vagotonic profile with a significant increase of linear parasympathetic indices and paroxysmal atrial fibrillation lasting longer than 4.5 hours; the second phenotype with increased sympathetic indices, low parasympathetic indices and paroxysms lasting up to 4.5 hours.
 Conclusion. Our findings indicate the potential of nonlinear analysis in the study of heart rate variability and demonstrate the feasibility of further integration of nonlinear indices for arrhythmia phenotyping.

Formulation of a novel HRV classification model as a surrogate fraudulence detection schema

Lie detection has been studied since a few decades ago, usually for the purpose of producing a scheme to assist in the investigation of identifying the culprit from a list of suspects. Heart Rate Variability (HRV) may be used as a method in lie detection due to its versatility and suitability. However, since its analysis is not instantaneous, a new experiment is described in this paper to overcome the problem. Additionally, a preliminary HRV classification model is designed to further enhance the classification model which is able to distinguish the lie from the truth for up to 80%.

A unified non-linear approach based on recurrence quantification analysis and approximate entropy: application to the classification of heart rate variability of age-stratified subjects.

This paper presents a unified approach based on the recurrence quantification analysis (RQA) and approximate entropy (ApEn) for the classification of heart rate variability (HRV). In this paper, the optimum tolerance threshold (ropt) corresponding to ApEnmax has been used for RQA calculation. The experimental data length (N) of RR interval series (RRi) is optimized by taking ropt as key parameter. ropt is found to be lying within the recommended range of 0.1 to 0.25 times the standard deviation of the RRi, when N ≥ 300. Consequently, RQA is applied to the age stratified RRi and indices such as percentage recurrence (%REC), percentage laminarity (%LAM), and percentage determinism (%DET) are calculated along with ApEnmax, [Formula: see text], [Formula: see text], and an index namely the radius differential (RD). Certain standard HRV statistical indices such as mean RR, standard deviation of RR (or NN) interval (SDNN), and the square root of the mean squared differences of successive RR intervals (RMSSD) (Eur Hear J 17:354-381, 1996) are also found for comparison. It is observed that (i) RD can discriminate between the elderly and young subjects with a value of 0.1151 ± 0.0236 (mean ± SD) and 0.0533 ± 0.0133 (mean ± SD) respectively for the elderly and young subjects and is found to be statistically significant with p < 0.05. (ii) Similar significant discrimination was obtained using [Formula: see text] with a value of 0.1827 ± 0.0382 (mean ± SD) and 0.2248 ± 0.0320 (mean ± SD) (iii) other significant indices were found to be %REC, %DET, %LAM, SDNN, and RMSSD; however, ApEnmax was found to be insignificant with p > 0.05. The above features of RRi time series were tested for classification using support vector machine (SVM) and multilayer perceptron neural network (MLPNN). Higher classification accuracy was achieved using SVM with a maximum value of 99.71%. Graphical abstract.

Model-Based Classification of Heart Rate Variability.

Several Heart Rate Variability (HRV) based novel methodologies for describing heart rate dynamics have been proposed in the literature with the aim of risk assessment. One such methodology is ARFIMA-EGARCH modeling which allows the quantification of long range dependence and time-varying volatility with the aim of describing non-linear and complex characteristics of HRV. This study applies the ARFIMA-EGARCH modeling of HRV recordings from 30 patients of the Noltisalis database to investigate the discrimination power of a set of features comprising currently used linear HRV features (low and high frequency components) and new measures obtained from the modeling such as, long memory in the mean, and persistence and asymmetry in volatility. A subset of the multidimensional HRV features is selected in a two-step procedure using Principal Components Analysis (PCA). Additionally, supervised classification by quadratic discriminant analysis achieves 93.3% of discrimination accuracy between the groups using the new feature set created by PCA.

A Novel Feature Level Fusion for Heart Rate Variability Classification Using Correntropy and Cauchy-Schwarz Divergence.

Heart rate variability (HRV) analysis has become a widely used tool for monitoring pathological and psychological states in medical applications. In a typical classification problem, information fusion is a process whereby the effective combination of the data can achieve a more accurate system. The purpose of this article was to provide an accurate algorithm for classifying HRV signals in various psychological states. Therefore, a novel feature level fusion approach was proposed. First, using the theory of information, two similarity indicators of the signal were extracted, including correntropy and Cauchy-Schwarz divergence. Applying probabilistic neural network (PNN) and k-nearest neighbor (kNN), the performance of each index in the classification of meditators and non-meditators HRV signals was appraised. Then, three fusion rules, including division, product, and weighted sum rules were used to combine the information of both similarity measures. For the first time, we propose an algorithm to define the weights of each feature based on the statistical p-values. The performance of HRV classification using combined features was compared with the non-combined features. Totally, the accuracy of 100% was obtained for discriminating all states. The results showed the strong ability and proficiency of division and weighted sum rules in the improvement of the classifier accuracies.

Mode by Mode Classification of Congestive Heart Failure from Long Term HRV Analysis

In this work a new approach for automatic discrimination of Congestive Heart Failure (CHF) from Normal Sinus Rhythm (NSR) using long-term Heart Rate Variability (HRV) signals is presented. Unlike classical approaches where features are extracted from the input HRV signal, in the present work the hidden information of the signal is retrieved from its different narrow-band components. These components are extracted using the Multivariate Empirical Modes Decomposition (MEMD). Being fully data driven approach, the MEMD is well suited for processing non-stationary signals such as HRV signals. The idea is to identify physically meaningful components that convey pertinent information about the CHF. From these components statistical and geometrical features extracted and combined. We investigate the potential of the higher order statistics (Skewness and Kurtosis) combined with the Hjorth and Poincare plot parameters as features for HRV classification. Different combinations of these features are analyzed, and feature vectors are selected as an input into the binary Support Vector Machines (SVM) classifier. To evaluate the performance of the proposed strategy, experimental tests have been conducted. Results of classification show an accuracy rate around 100%. We expected that the proposed method would be helpful, in clinical diagnosis, to discriminate CHF patients from NSR subjects using long-term HRV recordings.

Electrocardiogram analysis using a combination of statistical, geometric, and nonlinear heart rate variability features

The paper addresses a common and recurring problem of electrocardiogram (ECG) classification based on heart rate variability (HRV) analysis. Current understanding of the limits of HRV analysis in diagnosing different cardiac conditions is not complete. Existing research suggests that a combination of carefully selected linear and nonlinear HRV features should significantly improve the accuracy for both binary and multiclass classification problems. The primary goal of this work is to evaluate a proposed combination of HRV features. Other explored objectives are the comparison of different machine learning algorithms in the HRV analysis and the inspection of the most suitable period T between two consecutively analyzed R-R intervals for nonlinear features. We extracted 11 features from 5min of R-R interval recordings: SDNN, RMSSD, pNN20, HRV triangular index (HTI), spatial filling index (SFI), correlation dimension, central tendency measure (CTM), and four approximate entropy features (ApEn1-ApEn4). Analyzed heart conditions included normal heart rhythm, arrhythmia (any), supraventricular arrhythmia, and congestive heart failure. One hundred patient records from six online databases were analyzed, 25 for each condition. Feature vectors were extracted by a platform designed for this purpose, named ECG Chaos Extractor. The vectors were then analyzed by seven clustering and classification algorithms in the Weka system: K-means, expectation maximization (EM), C4.5 decision tree, Bayesian network, artificial neural network (ANN), support vector machines (SVM) and random forest (RF). Four-class and two-class (normal vs. abnormal) classification was performed. Relevance of particular features was evaluated using 1-Rule and C4.5 decision tree in the cases of individual features classification and classification with features' pairs. Average total classification accuracy obtained for top three classification methods in the two classes' case was: RF 99.7%, ANN 99.1%, SVM 98.9%. In the four classes' case the best results were: RF 99.6%, Bayesian network 99.4%, SVM 98.4%. The best overall method was RF. C4.5 decision tree was successful in the construction of useful classification rules for the two classes' case. EM and K-means showed comparable clustering results: around 50% for the four classes' case and around 75% for the two classes' case. HTI, pNN20, RMSSD, ApEn3, ApEn4 and SFI were shown to be the most relevant features. HTI in particular appears in most of the top-ranked pairs of features and is the best analyzed feature. The choice of the period T for nonlinear features was shown to be arbitrary. However, a combination of five different periods significantly improved classification accuracy, from 70% for a single period up to 99% for five periods. Analysis shows that the proposed combination of 11 linear and nonlinear HRV features gives high classification accuracy when nonlinear features are extracted for five periods. The features' combination was thoroughly analyzed using several machine learning algorithms. In particular, RF algorithm proved to be highly efficient and accurate in both binary and multiclass classification of HRV records. Interpretable and useful rules were obtained with C4.5 decision tree. Further work in this area should elucidate which features should be extracted for the best classification results for specific types of cardiac disorders.