Early detection of sudden cardiac death using nonlinear analysis of heart rate variability

Early prediction of sudden cardiac death risk with Nested LSTM based on electrocardiogram sequential features

Electrocardiogram (ECG) signals are very important for heart disease diagnosis. In this paper, a novel early prediction method based on Nested Long Short-Term Memory (Nested LSTM) is developed for sudden cardiac death risk detection. First, wavelet denoising and normalization techniques are utilized for reliable reconstruction of ECG signals from extreme noise conditions. Then, a nested LSTM structure is adopted, which can guide the memory forgetting and memory selection of ECG signals, so as to improve the data processing ability and prediction accuracy of ECG signals. To demonstrate the effectiveness of the proposed method, four different models with different signal prediction techniques are used for comparison. The extensive experimental results show that this method can realize an accurate prediction of the cardiac beat's starting point and track the trend of ECG signals effectively. This study holds significant value for timely intervention for patients at risk of sudden cardiac death.

Machine learning of electrophysiological signals for the prediction of ventricular arrhythmias: systematic review and examination of heterogeneity between studies

BACKGROUND

Ventricular arrhythmia (VA) precipitating sudden cardiac arrest (SCD) is among the most frequent causes of death and pose a high burden on public health systems worldwide. The increasing availability of electrophysiological signals collected through conventional methods (e.g. electrocardiography (ECG)) and digital health technologies (e.g. wearable devices) in combination with novel predictive analytics using machine learning (ML) and deep learning (DL) hold potential for personalised predictions of arrhythmic events.

METHODS

This systematic review and exploratory meta-analysis assesses the state-of-the-art of ML/DL models of electrophysiological signals for personalised prediction of malignant VA or SCD, and studies potential causes of bias (PROSPERO, reference: CRD42021283464). Five electronic databases were searched to identify eligible studies. Pooled estimates of the diagnostic odds ratio (DOR) and summary area under the curve (AUROC) were calculated. Meta-analyses were performed separately for studies using publicly available, ad-hoc datasets, versus targeted clinical data acquisition. Studies were scored on risk of bias by the PROBAST tool.

FINDINGS

2194 studies were identified of which 46 were included in the systematic review and 32 in the meta-analysis. Pooling of individual models demonstrated a summary AUROC of 0.856 (95% CI 0.755-0.909) for short-term (time-to-event up to 72 h) prediction and AUROC of 0.876 (95% CI 0.642-0.980) for long-term prediction (time-to-event up to years). While models developed on ad-hoc sets had higher pooled performance (AUROC 0.919, 95% CI 0.867-0.952), they had a high risk of bias related to the re-use and overlap of small ad-hoc datasets, choices of ML tool and a lack of external model validation.

INTERPRETATION

ML and DL models appear to accurately predict malignant VA and SCD. However, wide heterogeneity between studies, in part due to small ad-hoc datasets and choice of ML model, may reduce the ability to generalise and should be addressed in future studies.

FUNDING

This publication is part of the project DEEP RISK ICD (with project number 452019308) of the research programme Rubicon which is (partly) financed by the Dutch Research Council (NWO). This research is partly funded by the Amsterdam Cardiovascular Sciences (personal grant F.V.Y.T).

Verhulst map measures: new biomarkers for heart rate classification

Recording, monitoring, and analyzing biological signals has received significant attention in medicine. A fundamental phase for understanding a bio-system under various conditions is to process the corresponding bio-signal appropriately. To this effect, different conventional and nonlinear approaches have been proposed. However, since the non-stationary properties of the bio-signals are not revealed by traditional linear methods, nonlinear dynamical techniques play a crucial role in examining the behavior of a bio-system. This work proposes new bio-markers based on the chaotic nature of the biomedical signals. These measures were introduced using the Verhulst map, a simple tool for characterizing the morphology of the reconstructed phase space. For this purpose, we extracted the features from the heart rate (HR) signals of six groups of meditators and non-meditators. For a typical classification problem, the performance of some conventional classifiers, including the k-nearest neighbor, support vector machine, and Naïve Bayes, was appraised separately. In addition, the competence of a hybrid classification strategy was inspected using majority voting. The results indicated a maximum accuracy, F1-score, and sensitivity of 100%. These findings reveal that the proposed framework is eminently capable of analyzing and classifying the HR signals of the groups. In conclusion, the Verhulst diagram-based measures are simple and based on the dynamics of the bio-signals, which can be served for quantifying different signals in medical systems.

Cardiac and Autonomic Dysfunctions Assessed Through Recurrence Quantitative Analysis of Electrocardiogram Signals and an Application to the 6-Hydroxydopamine Parkinson's Disease Animal Model

A classic method to evaluate autonomic dysfunction is through the evaluation of heart rate variability (HRV). HRV provides a series of coefficients, such as Standard Deviation of n-n intervals (SDNN) and Root Mean Square of Successive Differences (RMSSD), which have well-established physiological associations. However, using only electrocardiogram (ECG) signals, it is difficult to identify proper autonomic activity, and the standard techniques are not sensitive and robust enough to distinguish pure autonomic modulation in heart dynamics from cardiac dysfunctions. In this proof-of-concept study we propose the use of Poincaré mapping and Recurrence Quantification Analysis (RQA) to identify and characterize stochasticity and chaoticity dynamics in ECG recordings. By applying these non-linear techniques in the ECG signals recorded from a set of Parkinson’s disease (PD) animal model 6-hydroxydopamine (6-OHDA), we showed that they present less variability in long time epochs and more stochasticity in short-time epochs, in their autonomic dynamics, when compared with those of the sham group. These results suggest that PD animal models present more “rigid heart rate” associated with “trembling ECG” and bradycardia, which are direct expressions of Parkinsonian symptoms. We also compared the RQA factors calculated from the ECG of animal models using four computational ECG signals under different noise and autonomic modulatory conditions, emulating the main ECG features of atrial fibrillation and QT-long syndrome.

Asymmetry of lagged Poincare plot in heart rate signals during meditation

Background and aim

Heart rate variability (HRV) quantifies the variability in the heart’s beat-to-beat intervals. This signal is a potential marker of cardiac function in normal, pathological, and psychological states. Signal asymmetry refers to an unequal distribution in the signal, which can be found by a two-dimensional Poincare plot. Earlier, heart rate asymmetry (HRA) was assessed using a conventional Poincare plot (lag of 1). In this study, we have investigated the effect of delay on the phase space asymmetry using lagged Poincare’s plot.

Experimental procedure

This study compared the presence/lack of asymmetries in the HRV data of 12 meditators (four Kundalini yoga (Yoga) at an advanced level of meditation, eight Chinese Chi meditators (Chi) ∼1–3 months) to 25 non-meditators (11 spontaneous nocturnal breathing (Normal) and 14 metronomic breathing (Metron)). Poincare’s plots were constructed with six different lags, and HRA was calculated. The analysis was conducted using HRV data provided in the Physionet database.

Results

The results showed that using conventional Poincare’s plot (lag of 1), the lowest HRA was observed in the Metron group. In addition, the HRA index was different between meditators and non-meditator groups. Moreover, as the most significant difference between groups was observed in a delay of 6, the role of the delay selection on the signal asymmetry was revealed.

Conclusion

The difference between lagged HRA responses on Yoga in comparison with other groups can be an emphasis on the importance of choosing the type of meditation technique and its effects on the cardiovascular system.

•

Asymmetries in HRV was assessed in different meditator and non-meditator groups.

•

The role of delay selection was explored on the phase space asymmetry using lagged Poincare plot.

•

A weaker asymmetry was observed in the metronomic breathing group.

•

The most significant difference between groups was perceived in a delay of six.

Detection of ventricular arrhythmia using hybrid time-frequency-based features and deep neural network

Sudden cardiac death (SCD) is a major cause of death among patients with heart diseases. It occurs mainly due to ventricular tachyarrhythmia (VTA) which includes ventricular tachycardia (VT) and ventricular fibrillation (VF) conditions. The main challenging task is to predict the VTA condition at a faster rate and timely application of automatic external defibrillator (AED) for saving lives. In this study, a VF/VT classification scheme has been proposed using a deep neural network (DNN) approach using hybrid time-frequency-based features. Two annotated public domain ECG databases (CUDB and VFDB) were used as training, test, and validation of datasets. The main motivation of this study was to implement a deep learning model for the classification of the VF/VT conditions and compared the results with other standard machine learning algorithms. The signal is decomposed with the wavelet transform, empirical mode decomposition (EMD) and variable mode decomposition (VMD) approaches and twenty-four are extracted to form a hybrid model from a window of length 5 s length. The DNN classifier achieved an accuracy (Acc) of 99.2%, sensitivity (Se) of 98.8%, and specificity (Sp) of 99.3% which is comparatively better than the results of the standard classifier. The proposed algorithm can detect VTA conditions accurately, hence could reduce the rate of misinterpretations by human experts and improves the efficiency of cardiac diagnosis by ECG signal analysis.

Textile-integrated polymer optical fibers for healthcare and medical applications

With ever growing interest in far-reaching solutions for pervasive healthcare and medicine, polymer optical fibers have been rendered into textile forms. Having both fiber-optic functionalities and traditional fabric-like comfort, textile-integrated polymer optical fibers have been advocated to remove the technical barriers for long-term uninterrupted health monitoring and treatment. In this context, this paper spotlights and reviews the recently developed textile-integrated polymer optical fibers in conjunction with fabrication techniques, applications in long-term continuous health monitoring and treatment, and future perspectives in the vision of mobile health (mHealth), as well as the introductory basics of polymer optical fibers. It is designed to serve as a topical guidepost for scientists and engineers on this highly interdisciplinary and rapidly growing topic.

Are signs and symptoms in cardiovascular rehabilitation correlated with heart rate variability? An observational longitudinal study

AIM

To analyze the correlation between the appearance of signs/symptoms during a cardiovascular rehabilitation program and linear indexes of the heart rate variability (HRV) at rest.

METHODS

To carry out the present observational longitudinal study, 48 patients were analyzed. The protocol was divided into two stages. First, the patients had their personal details collected, and the autonomic modulation at rest was evaluated by HRV. Second, they underwent 36 sessions of the cardiovascular rehabilitation program to evaluate signs/symptoms. Then, just for analysis of the data, they were divided into two groups: the group without signs/symptoms (n = 26; 65.15 ± 9.7 years); and the group with signs/symptoms (n = 22; 66.77 ± 14.4 years). The HRV indexes were compared by ancova. The effect size was measured through the partial eta-squared. Pearson's and Spearman's correlations (P < 0.05) were used to analyze the data, and linear regression was applied.

RESULTS

A total of 103 signs/symptoms occurred. The group with signs/symptoms presented lower values of HRV indexes when compared with the group without signs/symptoms, especially for the parasympathetic indexes with a large effect size. The root mean square of successive differences (rMSSD), percentage of adjacent RR intervals with a difference of duration >50 ms (pNN50), high-frequency spectral component (HF) varying from 0.15 to 0.4 Hz (expressed as ms² ), dispersion of the points perpendicular to the line of identity and represents the instantaneous record of the beat-to-beat variability (SD1) and SD1/scatter of points along the identity line and represents the HRV in long-term records (SD2) index presented a negative correlation with the appearance of signs/symptoms. When the linear regression was applied, the rMSSD, SD1 and SD1/SD2 showed negative values of β (P < 0.05).

CONCLUSIONS

Patients with lower HRV are more likely to have signs/symptoms. The rMSSD, pNN50, HF (expressed as ms² ), SD1 and SD1/SD2 index presented a negative correlation with the appearance of signs/symptoms. For rMSSD, SD1 and SD1/SD2, the lower the values of these HRV indexes, the greater the risk of appearance of signs/symptoms. Geriatr Gerontol Int 2020; 20: 853-859.

Human identification using a new matching Pursuit-based feature set of ECG

BACKGROUND AND OBJECTIVE

In recent years, many attempts have been made to design reliable systems for identifying individuals using biometrics. Electrocardiogram (ECG) biometric is one of the newest methods that not only offers unique characteristics of individuals for human identification, but also the possibility of counterfeiting it is negligible. In this paper, our objective was to develop an identification system using a non-fiducial one-lead ECG feature set based on a sparse algorithm.

METHODS

The ECG signals of 90 participants were decomposed using a matching pursuit (MP) and several statistical and nonlinear measures were extracted from the MP coefficients. Then, the performance of ECG characteristics delivered by MP analysis in human identification was evaluated by the probabilistic neural network (PNN) and k-nearest neighbor (kNN) with one vs. all strategy. The role of the feature set in classification rates was also tested in different modes, including linear attributes, nonlinear indices, all features, features selected by principal component analysis (PCA), and features selected by linear discriminant analysis (LDA).

RESULTS

Experimental results showed that (1) the highest recognition rate was 99.68%; (2) the performance of the PNN was superior to the kNN; and (3) selecting features with LDA resulted in higher identification rates.

CONCLUSIONS

The results are prominent from the performance perspective because it gives higher recognition rates over the group of 90 participants. The great performance of the proposed identification system advocates that it can be employed confidently in different smart systems.