Morphological modeling of cardiac signals based on signal decomposition

ECG data enhancement method using generate adversarial networks based on Bi-LSTM and CBAM

Objective.The classification performance of electrocardiogram (ECG) classification algorithms is easily affected by data imbalance, which often leads to poor model prediction performance for a few classes and a consequent decrease in the overall performance of the model.Approach.To address this problem, this paper proposed an ECG data augmentation method based on a generative adversarial network (GAN) that combines bidirectional long short-term memory (Bi-LSTM) networks and convolutional block attention mechanism (CBAM) to improve the overall performance of ECG classification models. In this paper, we used two ECG databases, namely the MIT-BIH arrhythmia (MIT-BIH-AR) database and the Chinese cardiovascular disease database (CCDD). The quality of the ECG signals produced by the generated models was assessed using the percent relative difference, root mean square error, Frechet distance, dynamic time warping (DTW), and Pearson correlation metrics. In addition, we also validated the impact of our proposed data augmentation method on ECG classification performance on MIT-BIH-AR database and CCDD.Main results.On the MIT-BIH-AR database, the overall accuracy of the data-enhanced balanced dataset was improved to 99.46% for 15 types of heartbeat classification task. On the CCDD, which focuses on the detection of ventricular precession (PVC), the overall accuracy of PVC detection improved to 99.15% after performing data enhancement.Significance.The experimental results indicate that the data augmentation method proposed in this paper can further improve the ECG classification performance.

Deep Generative Models: The winning key for large and easily accessible ECG datasets?

Large high-quality datasets are essential for building powerful artificial intelligence (AI) algorithms capable of supporting advancement in cardiac clinical research. However, researchers working with electrocardiogram (ECG) signals struggle to get access and/or to build one. The aim of the present work is to shed light on a potential solution to address the lack of large and easily accessible ECG datasets. Firstly, the main causes of such a lack are identified and examined. Afterward, the potentials and limitations of cardiac data generation via deep generative models (DGMs) are deeply analyzed. These very promising algorithms have been found capable not only of generating large quantities of ECG signals but also of supporting data anonymization processes, to simplify data sharing while respecting patients' privacy. Their application could help research progress and cooperation in the name of open science. However several aspects, such as a standardized synthetic data quality evaluation and algorithm stability, need to be further explored.

CardioSim: a PC-based cardiac signal simulator using segmental modeling of electrocardiogram

Background: ECG modeling has wide application in signal representation, compression and synthetic ECG generation. Method: CardioSim generates synthetic ECG waveform in real-time using PC-based system. It provides dual facility of interface-based visualization with hardware-based waveform generation. It has two stages viz., development of reference model parameter database using Fourier model and generation of synthetic ECG waveform based on user defined parameters using normal and abnormal records (H, APC, PVC, LBBB, RBBB, P) from mitdb under PhysioNet. Result: It generates ten various ECG waveforms including one healthy and nine diseased rhythms from a single dynamic model with flexible user defined parameters. It gives higher reconstruction performance in terms of SNR and MSE. The mean SNR for different beat morphology is 89.2(H), 88.37(V), 86.32(A), 85.35(L), 97.22(P) and 83.3(R) and mean MSE is 2.45 × 10-6(H), 3.14 × 10-6(V), 8.98 × 10-6(A), 5.82 × 10-6(L), 0.43 × 10-6(P) and 0.25 × 10-6(R). Conclusion: It improves the performance parameters over published research work on ECG modeling and simulation. It can be used as a self-learning tool for entry level medical students.

Sensor Fusion of Cardiorespiratory Signals Using an Adaptive Kalman Filter. ANNUAL INTERNATIONAL CONFERENCE OF THE IEEE ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY. IEEE ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY. ANNUAL INTERNATIONAL CONFERENCE 2023;2023:1-4. [PMID: 38082963 DOI: 10.1109/embc40787.2023.10340942] [Citation(s) in RCA: 0] [Impact Index Per Article: 0]

For unobtrusive monitoring of vital signs, redundant sensors are beneficial to fuse several sensor measurements which can improve the estimation of, e.g. heart rate and respiratory rate. In this paper, an adaptive unscented Kalman filter is used to estimate respiratory rate and heart rate on a new simplified model for cardiorespiratory coupling. Additionally, the Kalman filter is tuned to incorporate the non-white system noise of the model. The Kalman filter is tested on synthesised data with variations regarding SNR, model mismatch and amount of sensors. For respiratory rate, a median squared error of as low as 0.02BPM2 and, for heart rate, a median squared error of as low as 0.2BPM2 for ideal assumptions is achieved.

Deep Learning-Based Data Augmentation and Model Fusion for Automatic Arrhythmia Identification and Classification Algorithms

Automated ECG-based arrhythmia detection is critical for early cardiac disease prevention and diagnosis. Recently, deep learning algorithms have been widely applied for arrhythmia detection with great success. However, the lack of labeled ECG data and low classification accuracy can have a significant impact on the overall effectiveness of a classification algorithm. In order to better apply deep learning methods to arrhythmia classification, in this study, feature extraction and classification strategy based on generative adversarial network data augmentation and model fusion are proposed to address these problems. First, the arrhythmia sparse data is augmented by generative adversarial networks. Then, aiming at the identification of different types of arrhythmias in long-term ECG, a spatial information fusion model based on ResNet and a temporal information fusion model based on BiLSTM are proposed. The model effectively fuses the location information of the nearest neighbors through the local feature extraction part of the generated ECG feature map and obtains the correlation of the global features by autonomous learning in multiple spaces through the BiLSTM network in the part of the global feature extraction. In addition, an attention mechanism is introduced to enhance the features of arrhythmia-type signal segments, and this mechanism can effectively focus on the extraction of key information to form a feature vector for final classification. Finally, it is validated by the enhanced MIT-BIH arrhythmia database. The experimental results demonstrate that the proposed classification technique enhances arrhythmia diagnostic accuracy by 99.4%, and the algorithm has high recognition performance and clinical value.

Electrocardiogram Quality Assessment with a Generalized Deep Learning Model Assisted by Conditional Generative Adversarial Networks

The electrocardiogram (ECG) is widely used for cardiovascular disease diagnosis and daily health monitoring. Before ECG analysis, ECG quality screening is an essential but time-consuming and experience-dependent work for technicians. An automatic ECG quality assessment method can reduce unnecessary time loss to help cardiologists perform diagnosis. This study aims to develop an automatic quality assessment system to search qualified ECGs for interpretation. The proposed system consists of data augmentation and quality assessment parts. For data augmentation, we train a conditional generative adversarial networks model to get an ECG segment generator, and thus to increase the number of training data. Then, we pre-train a deep quality assessment model based on a training dataset composed of real and generated ECG. Finally, we fine-tune the proposed model using real ECG and validate it on two different datasets composed of real ECG. The proposed system has a generalized performance on the two validation datasets. The model's accuracy is 97.1% and 96.4%, respectively for the two datasets. The proposed method outperforms a shallow neural network model, and also a deep neural network models without being pre-trained by generated ECG. The proposed system demonstrates improved performance in the ECG quality assessment, and it has the potential to be an initial ECG quality screening tool in clinical practice.

Impulse Data Model For Solving The Inverse Problem of Electrocardiography

OBJECTIVE

To develop, train and test neural networks for predicting heart surface potentials (HSPs) from body surface potentials (BSPs). The method re-frames traditional inverse problems of electrocardiograpy into regression problems, constraining the solution space by decomposing signals with multidimensional Gaussian impulse basis functions.

METHODS

Impulse HSPs were generated with single Gaussian basis functions at discrete heart surface locations and projected to corresponding BSPs using a volume conductor torso model. Both BSP (inputs) and HSP (outputs) were mapped to regular 2D surface meshes and used to train a neural network. Predictive capabilities of the network were tested with unseen synthetic and experimental data.

RESULTS

A dense full connected single hidden layer neural network was trained to map body surface impulses to heart surface Gaussian basis functions for reconstructing HSP. Synthetic pulses moving across the heart surface were predicted from the neural network with root mean squared error of 9.1 +/ 1.4%. Predicted signals were robust to noise up to 20 dB and errors due to displacement and rotation of the heart within the torso were bounded and predictable. A shift of the heart 40 mm toward the spine resulted in a 4% increase in signal feature localization error. The set of training impulse function data could be reduced and prediction error remained bounded. Recorded HSPs from in-vitro pig hearts were reliably decomposed using space-time Gaussian basis functions. Activation times calculated from predicted HSPs for left-ventricular pacing had a mean absolute error of 10.4 +/ 11.4 ms. Other pacing scenarios were analyzed with similar success.

CONCLUSION

Impulses from Gaussian basis functions are potentially an effective and robust way to train simple neural network data models for reconstructing HSPs from decomposed BSPs.

SIGNIFICANCE

The HSPs predicted by the neural network can be used to generate activation maps that non-invasively identify features of cardiac electrical dysfunction and can guide subsequent treatment options.

ECG Signal Modeling Using Volatility Properties: Its Application in Sleep Apnea Syndrome

This study presents and evaluates the mathematical model to estimate the mean and variance of single-lead ECG signals in sleep apnea syndrome. Our objective is to use the volatility property of the ECG signal for modeling. ECG signal is a stochastic signal whose mean and variance are time-varying. So, we propose to decompose this nonstationarity into two additive components; a homoscedastic Autoregressive Integrated Moving Average (ARIMA) and a heteroscedastic time series in terms of Exponential Generalized Autoregressive Conditional Heteroskedasticity (EGARCH), where the former captures the linearity property and the latter the nonlinear characteristics of the ECG signal. First, ECG signals are segmented into one-minute segments. The heteroskedasticity property is then examined through various tests such as the ARCH/GARCH test, kurtosis, skewness, and histograms. Next, the ARIMA model is applied to signals as a linear model and EGARCH as a nonlinear model. The appropriate orders of models are estimated by using the Bayesian Information Criterion (BIC). We assess the effectiveness of our model in terms of mean square error (MSE), root mean square error (RMSE), mean absolute error (MAE), and mean absolute percentage error (MAPE). The data in this article is obtained from the Physionet Apnea-ECG database. Results show that the ARIMA-EGARCH model performs better than other models for modeling both apneic and normal ECG signals in sleep apnea syndrome.

The hidden waves in the ECG uncovered revealing a sound automated interpretation method

A novel approach for analysing cardiac rhythm data is presented in this paper. Heartbeats are decomposed into the five fundamental P, Q, R, S and T waves plus an error term to account for artifacts in the data which provides a meaningful, physical interpretation of the heart's electric system. The morphology of each wave is concisely described using four parameters that allow all the different patterns in heartbeats to be characterized and thus differentiated This multi-purpose approach solves such questions as the extraction of interpretable features, the detection of the fiducial marks of the fundamental waves, or the generation of synthetic data and the denoising of signals. Yet the greatest benefit from this new discovery will be the automatic diagnosis of heart anomalies as well as other clinical uses with great advantages compared to the rigid, vulnerable and black box machine learning procedures, widely used in medical devices. The paper shows the enormous potential of the method in practice; specifically, the capability to discriminate subjects, characterize morphologies and detect the fiducial marks (reference points) are validated numerically using simulated and real data, thus proving that it outperforms its competitors.

Predictive modeling of drug effects on electrocardiograms

Whole electrocardiogram (ECG) waveform analysis is a technique for evaluating aggregate arrhythmic risks of drugs. In this paper, we propose methods for exploring changes to ECG morphology due to drug effects using Gaussian model parameters, and predict patient specific post-drug ECG based on pre-drug ECG. We evaluate the proposed methods using clinical ECG recordings from subjects under the effect of anti-arrhythmic drugs Dofetilide, Quinidine, Ranolazine, and Verapamil, from the ECGRVDQ database on PhysioNet. Paired-sample t-test p-values (>0.05) suggest the proposed method can achieve similar results when compared to expert annotated J to T_peak and T_peak to T_end intervals for all four drug states. We employed a leave-one-out cross validation strategy to train the prediction model and produce the results. Mean Pearson correlations between all predicted and recorded post-drug waveform morphologies for all drug states across both the vector magnitude lead and Lead II is 0.94±0.05, with p-values <0.01 for all predictions; indicating significant predictions. Parameters from ECG models with Gaussian basis can be used to calculate clinically useful information and to capture or predict changes in cardiac signals due to drug effects.

Developing of robust and high accurate ECG beat classification by combining Gaussian mixtures and wavelets features

Electrocardiogram (ECG) beat classification is a significant application in computer-aided analysis and diagnosis technologies. This paper proposed a method to detect, extract informative features, and classify ECG beats utilizing real ECG signals available in the standard MIT-BIH Arrhythmia database, with 10,502 beats had been extracted from it. The present study classifies the ECG beat into six classes, normal beat (N), Left bundle branch block beat, Right bundle branch block beat, Premature ventricular contraction, atrial premature beat, and aberrated atrial premature, using Gaussian mixture and wavelets features, and by applying principal component analysis for feature set reduction. The classification process is implemented utilizing two classifier techniques, the probabilistic neural network (PNN) algorithm and Random Forest (RF) algorithm. The achieved accuracy is 99.99%, and 99.97% for PNN and RF respectively. The precision is 99.99%, and 99.98% for PNN and RF respectively. The sensitivity is 99.99%, and 99.81% for PNN and RF respectively, while the specificity is 99.97%, 99.96% for PNN and RF respectively. It has been shown that the combination of Gaussian mixtures coefficients and the wavelets features have provided a valuable information about the heart performance and can be used significantly in arrhythmia classification.

ECG Signal De-noising and Baseline Wander Correction Based on CEEMDAN and Wavelet Threshold

A novel electrocardiogram (ECG) signal de-noising and baseline wander correction method based on complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) and wavelet threshold is proposed. Although CEEMDAN is based on empirical mode decomposition (EMD), it represents a significant improvement of the original EMD by overcoming the mode-mixing problem. However, there has been no previous study on using CEEMDAN to de-noise ECG signals, to the authors’ best knowledge. In the proposed method, the original noisy ECG signal is decomposed into a series of intrinsic mode functions (IMFs) sorted from high to low frequency by CEEMDAN. Each IMF is then analyzed by the autocorrelation method to find out the first few high frequency IMFs containing random noise, and these IMFs should be de-noised by the wavelet threshold. The zero-crossing rate (ZCR) of all IMFs, including final residue, are computed, and the IMFs with ZCR less than a certain value are removed. Finally, the remaining IMFs are reconstructed to obtain the clean ECG signal. The proposed algorithm is validated through experiments using the MIT–BIH ECG databases, and the results show that the random noise in the ECG signal can be effectively suppressed, and at the same time the baseline wander can be corrected efficiently.

An enhanced method for real-time modelling of cardiac related biosignals using Gaussian mixtures

Cardiac related biosignals modelling is very important for detecting, classification, compression and transmission of such health-related signals. This paper introduces a new, fast and accurate method for modelling the cardiac related biosignals (ECG and PPG) based on a mixture of Gaussian waves. For any signal, at first, the start and end of the ECG beat or PPG pulse is detected, then the baseline is detected then subtracted from the original signal, after that the signal is divided into two signals positive and negative, each modelled separately then incorporated together to form the modelled signal. The proposed method is applied in the MIMIC, and MIT-BIH Arrhythmia databases available online at PhysioNet.

A new algorithm for estimating the ν-index using sinusoidal basis functions

Recently it was shown that the spatial dispersion of ventricular repolarization (SHVR) can be assessed from the surface ECG using a metric termed ν-index. In this paper, a new algorithm is presented for estimating the ν-index, allowing the inclusion of higher order terms with ease, even in the presence of noise, leading to more accurate estimates. We first introduced a new analytical model for the derivative of the average transmembrane potentials during repolarization (the dominant T-wave) based on trigonometric functions. This functional set is closed under the operation of derivation. Therefore, the nonlinear iterative optimization required by previous methods is no longer necessary. Then, we suggested an iterative linear matrix factorization method to properly estimate the leads factors and the ν-index. Several synthetic SHVR (in the range 20 to 70 ms) were simulated, employing a publicly-available forward electrophysiological model (ECGSIM), leading to a total of 240 synthetic 8-lead electrocardiographical recordings (ECG), each composed of 128 beats. Then the ν-index was estimated with the newly introduced method and compared (root mean square error, RMSE) with the theoretical values, available for each series. The simulation results confirmed the theoretical expectations and indeed showed that the ν-index estimates were improved by increasing the number of lead factors included (RMSE=0:295±0:037 vs 0:280±0:038 for 2 and 8 lead factors respectively).