ECG Arrhythmia Research Articles

AttentivECGRU: GRU based autoencoder with attention mechanism and automated fuzzy thresholding for ECG arrhythmia detection

AttentivECGRU: GRU based autoencoder with attention mechanism and automated fuzzy thresholding for ECG arrhythmia detection

Sensor-Enhanced Deep Learning for ECG Arrhythmia Detection: Integrating Multiscale Convolutional and Self-Attentive GRU Networks

Sensor-Enhanced Deep Learning for ECG Arrhythmia Detection: Integrating Multiscale Convolutional and Self-Attentive GRU Networks

Cardiovascular disease detection from cardiac arrhythmia ECG signals using artificial intelligence models with hyperparameters tuning methodologies

Cardiovascular disease (CVD) is connected with irregular cardiac electrical activity, which can be seen in ECG alterations. Due to its convenience and non-invasive aspect, the ECG is routinely exploited to identify different arrhythmias and automatic ECG recognition is needed immediately. In this paper, enhancement for the detection of CVDs such as Ventricular Tachycardia (VT), Premature Ventricular Contraction (PVC) and ST Change (ST) arrhythmia using different dimensionality reduction techniques and multiple classifiers are presented. Three-dimensionality reduction methods, such as Local Linear Embedding (LLE), Diffusion Maps (DM), and Laplacian Eigen (LE), are employed. The dimensionally reduced ECG samples are further feature selected with Cuckoo Search (CS) and Harmonic Search Optimization (HSO) algorithms. A publicly available MIT-BIH (Physionet) - VT database, PVC database, ST Change database and NSR database were used in this work. The cardiac vascular disturbances are classified by using seven classifiers such as Gaussian Mixture Model (GMM), Expectation Maximization (EM), Non-linear Regression (NLR), Logistic Regression (LR), Bayesian Linear Discriminant Analysis (BDLC), Detrended Fluctuation Analysis (Detrended FA), and Firefly. For different classes, the average overall accuracy of the classification techniques is 55.65 % when without CS and HSO feature selection, 64.36 % when CS feature selection is used, and 75.39 % when HSO feature selection is used. Also, to improve the performance of classifiers, the hyperparameters of four classifiers (GMM, EM, BDLC and Firefly) are tuned with the Adam and Grid Search Optimization (GSO) approaches. The average accuracy of classification for the CS feature-based classifiers that used GSO and Adam hyperparameter tuning was 79.92 % and 85.78 %, respectively. The average accuracy of classification for the HSO feature-based classifiers that used GSO and Adam hyperparameter tuning was 86.87 % and 93.77 %, respectively. The performance of the classifier is analyzed based on the accuracy parameter for both with and without feature selection methods and with hyperparameter tuning techniques. In the case of ST vs. NSR, a higher accuracy of 98.92 % is achieved for the LLE dimensionality reduction with HSO feature selection for the GMM classifier with Adam's hyperparameter tuning approach. The GMM classifier with the Adam hyperparameter tuning approach with 98.92 % accuracy in detecting ST vs. NSR cardiac disease is outperforming all other classifiers and methodologies.

Complexity from ordinal pattern positioned slopes (COPPS)

Measuring complexity allows to characterize complex systems. Existing techniques are limited to simultaneously measure complexity from short length data sets, detect transitions and periodic dynamics. This paper presents an approach based on ordinal pattern positioned slopes (OPPS). It considers exclusively OPPS group occurrences to compute the complexity from OPPS (COPPS) as the average number of patterns and applies to short data series. The COPPS measure was successfully applied to simulation data for measuring complexity, detecting transition phases and regular dynamics, distinguishing between chaotic and stochastic dynamics; and to real-world data for detecting arrhythmia ECG beats.

Impact of artificial intelligence arrhythmia mapping on time to first ablation, procedure duration, and fluoroscopy use.

Artificial intelligence (AI) ECG arrhythmia mapping provides arrhythmia source localization using 12-lead ECG data; whether this information impacts procedural efficiency is unknown. We performed a retrospective, case-control study to evaluate the hypothesis that AI ECG mapping may reduce time to ablation, procedural duration, and fluoroscopy. Cases in which system output was used were retrospectively enrolled according to IRB-approved protocols at each site. Matched control cases were enrolled in reverse chronological order beginning on the last day for which the technology was unavailable. Controls were matched based upon physician, institution, arrhythmia, and a predetermined complexity rating. Procedural metrics, fluoroscopy data, and clinical outcomes were assessed from time-stamped medical records. The study group consisted of 28 patients (age 65 ± 11 years, 46% female, left atrialdimension 4.1 ± 0.9 cm, LVEF 50 ± 18%) and was similar to 28 controls. The most common arrhythmia types were atrial fibrillation (n = 10), premature ventricular complexes (n = 8), and ventricular tachycardia (n = 6). Use of the system was associated with a 19.0% reduction in time to ablation (133 ± 48 vs. 165 ± 49 min, p = 0.02), a 22.6% reduction in procedure duration (233 ± 51 vs. 301 ± 83 min, p < 0.001), and a 43.7% reduction in fluoroscopy (18.7 ± 13.3 vs. 33.2 ± 18.0 min, p < 0.001) versus controls. At 6 months follow-up, arrhythmia-free survival was 73.5% in the study group and 63.3% in the control group (p = 0.56). Use of forward-solution AI ECG mapping is associated with reductions in time to first ablation, procedure duration, and fluoroscopy without an adverse impact on procedure outcomes or complications.

Robust algorithm for the detection and classification of QRS complexes with different morphologies using the continuous spline wavelet transform with automatic scale detection

This work presents an algorithm for the detection and classification of QRS complexes based on the continuous wavelet transform (CWT) with splines. This approach can evaluate the CWT at any integer scale and the analysis is not restricted to powers of two. The QRS detector comprises four stages: implementation of CWT with splines, detection of QRS complexes, searching for undetected QRS complexes, and correction of the R wave peak location in detected QRS complexes. After, the onsets and ends of the QRS complexes are detected. The algorithm was evaluated with synthetic ECG and with the manually annotated databases: MIT-BIH Arrhythmia, European ST-T, QT and PTB Diagnostic ECG. Evaluation results of the QRS detector were: MIT-BIH arrhythmia database (109,447 beats analyzed), sensitivity Se = 99.72% and positive predictivity P+ = 99.87%; European ST-T database (790522 beats analyzed), Se = 99.92% and P+ = 99.55% and QT database (86498 beats analyzed), Se = 99.97% and P+ = 99.99%. To evaluate the delineation algorithm of the QRS onset (Qi) and QRS end (J) with the QT and PTB Diagnostic ECG databases, the mean and standard deviations of the differences between the automatic and manual annotated location of these points were calculated. The standard deviations were close to the accepted tolerances for deviations determined by the CSE experts. The proposed algorithm is robust to noise, artifacts and baseline drifts, classifies QRS complexes, automatically selects the CWT scale according to the sampling frequency of the ECG record used, and adapts to changes in the heart rate, amplitude and morphology of QRS complexes.

YOLO-ECG: multi-stage ECG arrhythmia classification using deep learning based YOLO network for portable monitoring

Cardiovascular disease is the leading cause of death and more than half million people were died around the world. However, cardiovascular health monitoring is crucial for effective heart disease diagnosis and management. In this paper, a novel deep learning-based YOLO-ECG model is proposed to ECG arrhythmia classification method for portable monitoring. Initially, the ECG signals are gathered using 12-lead electrodes in the real time and these signals are denoised using two-dimensional stationary wavelet transform (2D-SWT). In SWT, zeros are inserted between filter taps rather than decimal points to eliminate repetitions and increase robustness. The denoised ECG signals are fed into the deep learning-based YOLO network with Gaussian error linear unit (GELU) activation function for detecting the ECG abnormalities of arrythmia. ECG waveforms are analyzed for the local fractal dimension at each sample point before heartbeat waveforms are extracted within a set length window. A squeeze and excitation attention (SEAN) module is introduced in the YOLO network for selecting size of 1D convolution kernel, and the dimension is preserved during local cross-channel interactions, decrease network complexity and enhance model efficiency. The classification findings demonstrate that the proposed YOLO-ECG model performs better by ECG recordings from the MIT-BIH arrhythmia dataset. From the experimental analysis, the proposed YOLO-ECG model yields the overall accuracy of 99.16% for efficient classification of arrythmia ECG signals.

ECG arrhythmia detection in an inter-patient setting using Fourier decomposition and machine learning

ECG beat classification or arrhythmia detection through artificial intelligence (AI) is an active topic of research. It is vital to recognize and detect the type of arrhythmia for monitoring cardiac abnormalities. The AI-based ECG beat classification algorithms proposed in the literature suffer from two main drawbacks. Firstly, some of the works have not considered any unseen test data to validate the performance of their algorithms. Secondly, the accuracy of detecting superventricular ectopic beats (SVEB) needs to be improved. In this work, we address these issues by considering an inter-patient paradigm where the test dataset is collected from a different set of subjects than the training data. Also, the proposed methodology detects SVEB with an F1 score of 89.35%, which is better than existing algorithms. We have used the Fourier decomposition method (FDM) for multi-scale analysis of ECG signals and extracted time-domain and statistical features from the narrow-band signal components obtained using FDM. Feature selection techniques, including the Kruskal-Wallis test and minimum redundancy maximum relevance (mRMR) have been used to select only the relevant features and rank these features to remove any redundancy. Since the dataset used is highly imbalanced, Mathew's correlation coefficient (MCC) has also been used to analyze the performance of the proposed method. Support vector machine classifier with linear kernel achieves an overall 98.03% accuracy and 91.84% MCC for the MIT-BIH arrhythmia dataset.

Atrial arrhythmia predicts late events and mortality in patients with D-transposition of the great arteries and atrial switch repair

AimsPatients with D-transposition of the great arteries (D-TGA) and atrial switch experience late morbidity and mortality related to atrial arrhythmias and systemic right ventricular (SRV) failure. We sought to analyze the influence of atrial arrhythmias on long-term outcomes in this group. MethodsA retrospective review of all patients with D-TGA and atrial switch followed at a tertiary care center was performed. Results148 patients (63.5 % male; age 30.4 ± 10.6 years) were followed for 12 ± 9.8 years. Death or cardiac transplantation occurred in 22(15 %) patients and heart failure hospitalization occurred in 30(20 %) patients.Atrial arrhythmias were documented in 82(55.4 %) patients. Atrial fibrillation at the first visit (Kaplan-Meier estimate, p = 0.003) and atrial fibrillation as a time-dependent variable (HR 3.50, p = 0.006) predicted increased risk of death or cardiac transplantation. A triad of atrial fibrillation, prolonged QRS duration/RBBB, and severe SRV dysfunction (SRV EF < 35 %) emerged as a unique signature of a higher-risk population.Atrial tachycardia and flutter, while not associated with mortality, increased the risk of heart failure hospitalization (HR 3.5, p = 0.001). Moreover, 2/6 cases of resuscitated sudden cardiac arrest were caused by atrial flutter, and more patients received inappropriate shocks for atrial arrhythmias(16 %) than appropriate shocks(2.3 %). ConclusionIn D-TGA patients with atrial switch, there is a complex interplay between atrial arrhythmias and the SRV. Key ECG parameters, arrhythmia events and sequelae create a unique patient-specific fingerprint strongly associated with future events and mortality. This higher-risk cohort will need further characterization to delineate who may benefit from preemptive arrhythmia intervention.

A Novel Inter Patient ECG Arrhythmia Classification Approach with Deep Feature Extraction and 1D Convolutional Neural Network

A Novel Inter Patient ECG Arrhythmia Classification Approach with Deep Feature Extraction and 1D Convolutional Neural Network

Approach Based Lightweight Custom Convolutional Neural Network and Fine-Tuned MobileNet-V2 for ECG Arrhythmia Signals Classification

Approach Based Lightweight Custom Convolutional Neural Network and Fine-Tuned MobileNet-V2 for ECG Arrhythmia Signals Classification

SoC-Based Implementation of 1D Convolutional Neural Network for 3-Channel ECG Arrhythmia Classification via HLS4ML

SoC-Based Implementation of 1D Convolutional Neural Network for 3-Channel ECG Arrhythmia Classification via HLS4ML

Optimized multi-stage sifting approach for ECG arrhythmia classification with shallow machine learning models

Optimized multi-stage sifting approach for ECG arrhythmia classification with shallow machine learning models

VAE 기반 데이터 증강과 CNN을 적용한 ECG 부정맥 분류 모델

VAE 기반 데이터 증강과 CNN을 적용한 ECG 부정맥 분류 모델

RNN Based Deep Learning Approach for ECG Beat Classification

The electrical patterns of the heart, captured through an electrocardiogram (ECG/EKG), serve as a diagnostic tool to identify potential issues such as heart attacks, irregular heart rhythms, heart failure, and arrhythmia, which manifests as irregularities in the heartbeat's rhythm. Deep Learning (DL) architectures have been successfully employed for arrhythmia detection and classification and offered superior performance to traditional shallow Machine Learning (ML) approaches. This paper introduces a novel approach utilizing deep learning techniques, specifically a Bidirectional Gated Recurrent Unit (Bi-GRU) model a variant of RNN, to classify ECG arrhythmia beats into distinct categories. The Bi-GRU model is employed in this work due to its capability to capture temporal dependencies in ECG signals, enabling a nuanced understanding of beat sequences for precise classification. Leveraging the MIT-BIH arrhythmia database, a comprehensive dataset containing annotated ECG signals, this study explores the efficacy of deep learning in accurately categorizing beats in to five super classes as per the standard of Association for the Advancement of Medical Instrumentation (AAMI). Evaluation metrics encompassing accuracy (Acc), specificity (Spe), sensitivity (Sen), positive predictive value (Ppv), and F1-score are utilized to assess the model performance in distinguishing between diverse arrhythmia classes.

A lightweight U-Net model for denoising and noise localization of ECG signals

ObjectiveElectrocardiogram (ECG) is a diagnostic tool for cardiovascular disease in clinical medicine, but noisy ECG signals can easily lead to misdiagnosis. This paper presents a lightweight U-Net model (LUNet) for denoising and noise localization of ECG signals of varying lengths. MethodsThe proposed model acts as a denoising autoencoder, which accepts an ECG signal containing noises and outputs the denoised signal. The locations of unacceptable noise are determined by comparing the difference between the input and output signals of the model. ResultsThe proposed method for ECG signal denoising achieved Root Mean Square Error (RMSE) scores of 0.0116 ± 0.0016, 0.0149 ± 0.0033, 0.0122 ± 0.0015, and 0.0098 ± 0.0012 for baseline wander (BW), muscle artifacts (MA), electrode motion artifacts (EM), and Additive white Gaussian noise (AWGN), respectively. For ECG signal noise localization, the proposed method obtained high Intersection over Union (IoU) scores and Dice scores. Meanwhile, our model showed excellent generalization performance on two public ECG databases. On the 12-lead ECG arrhythmia classification task, the results using the processed dataset with the proposed method were 2.4% higher in F1 score than using the original dataset. ConclusionThe proposed method is highly effective in removing acceptable noise from ECG signals of varying lengths and locating the positions of unacceptable noise. SignificanceThis paper presents a promising solution for denoising and noise localization of ECG signals, which can improve ECG interpretation and potentially lead to better diagnosis and treatment of cardiovascular diseases.

An intelligent deep neural network with Opposition based Laplacian Equilibrium Optimizer to improve feature extraction using ECG signals

ProblemThe leading cause of death in the world is cardiovascular disease (CVD). It is required to detect, prevent and manage cardiovascular diseases to reduce their effect and improve the outcome. Interpretation of ECG signals may vary from one expert to another. That’s why an automatic categorization of ECG recordings is needed. AimThe proposed study attempted to address the above issues by developing an automated method for predicting arrhythmias based on ECG signal categorization. This method would offer more precise and enhanced interpretation of ECGs while reducing small biases brought on by manual interpretation. MethodsIn this paper, a unique deep learning-based algorithm is used to efficiently and swiftly classify cardiac arrhythmias. We have used Discrete Wavelet Transformation (DWT) to appropriately classify ECG data into important arrhythmia categories during the pre-processing stage by removing intrinsic noise from ECG signals. To obtain optimal placement (equilibrium state) for better search space diversification, OB-L-EO uses Laplace distribution followed by opposition-based modelling to modify the intensity of search agents. A powerful type of machine learning technique called deep neural network (DNN) is used to accomplish signal categorization among sixteen categories using the MIT-BIH Arrhythmia ECG dataset. ResultsWith the help of analysing classification technique including optimized feature vector and the OBLEO method with DNN, the heart rate has been detected with a 98.779% accuracy, 98.67% sensitivity, 98.87% specificity and execution time with a value of 1.34 ms across the entire dataset. The application of 10-fold cross validation provides the better accuracy i.e 99.79%. ConclusionThe achieved performance parameters have demonstrated that the suggested ECG signal classification model is capable of categorizing arrhythmia classes appropriately using ECG signals.

Ensembled Feature based Multi-Label ECG Arrhythmia Classification.

Arrhythmia is one of the major contributors to sudden cardiac death. Compared to single-label arrhythmia classification, multi-label classification provides the advantage of simultaneous detection of different cardiac ailments. Existing multi-label approaches employ deep learning models, which have several limitations, e.g., data dependency, overfitting, complexity, and lack of interpretability. In this work, we propose a multi-label classification framework with ensembled hand-crafted features added with label powerset, which better preserves the correlation between classes. Experimentation with CPSC 2018 database shows that the proposed model provides accuracy and F1-score of 88.52% and 88.45% for single-label data and 88.11% and 86.32% for multi-label data, respectively. The proposed method improves individual class performance for some important diseases compared to the existing multi-label state-of-the-art methods.Clinical relevance- Early detection of life-threatening multi-label ECG patterns consisting of atrial fibrillation, atrioventricular block, and right bundle branch block with high accuracy can reduce the sudden cardiac death rate.

A 510μW 0.738-mm2 6.2-pJ/SOP Online Learning Multi-Topology SNN Processor with Unified Computation Engine in 40-nm CMOS.

Implementing neural networks (NN) on edge devices enables AI to be applied in many daily scenarios. The stringent area and power budget on edge devices impose challenges on conventional NNs with massive energy-consuming Multiply Accu- mulation (MAC) operations and offer an opportunity for Spiking Neural Networks (SNN), which can be implemented within sub- mW power budget. However, mainstream SNN topologies varies from Spiking Feedforward Neural Network (SFNN), Spiking Recurrent Neural Network (SRNN), to Spiking Convolutional Neural Network (SCNN), and it is challenging for the edge SNN processor to adapt to different topologies. Besides, online learning ability is critical for edge devices to adapt to local environments but comes with dedicated learning modules, further increasing area and power consumption burdens. To alleviate these prob- lems, this work proposed RAINE, a reconfigurable neuromorphic engine supporting multiple SNN topologies and a dedicated trace- based rewarded spike-timing-dependent plasticity (TR-STDP) learning algorithm. Sixteen Unified-Dynamics Learning-Engines (UDLEs) are implemented in RAINE to realize a compact and reconfigurable implementation of different SNN operations. Three topology-aware data reuse strategies are proposed and analyzed to optimize the mapping of different SNNs on RAINE. A 40-nm prototype chip is fabricated, achieving energy-per- synaptic-operation (SOP) of 6.2 pJ/SOP at 0.51V, and power consumption of 510 μW at 0.45V. Finally, three examples with dif- ferent SNN topologies, including SRNN-based ECG arrhythmia detection, SCNN-based 2D image classification, and end-to-end on-chip learning for MNIST digit recognition, are demonstrated on RAINE with ultra-low energy consumption of 97.7nJ/step, 6.28μJ/sample, and 42.98μJ/sample respectively. These results show the feasibility of obtaining high reconfigurability and low power consumption simultaneously on a SNN processor.

Deep Learning-Based ECG Arrhythmia Classification: A Systematic Review

Deep learning (DL) has been introduced in automatic heart-abnormality classification using ECG signals, while its application in practical medical procedures is limited. A systematic review is performed from perspectives of the ECG database, preprocessing, DL methodology, evaluation paradigm, performance metric, and code availability to identify research trends, challenges, and opportunities for DL-based ECG arrhythmia classification. Specifically, 368 studies meeting the eligibility criteria are included. A total of 223 (61%) studies use MIT-BIH Arrhythmia Database to design DL models. A total of 138 (38%) studies considered removing noise or artifacts in ECG signals, and 102 (28%) studies performed data augmentation to extend the minority arrhythmia categories. Convolutional neural networks are the dominant models (58.7%, 216) used in the reviewed studies while growing studies have integrated multiple DL structures in recent years. A total of 319 (86.7%) and 38 (10.3%) studies explicitly mention their evaluation paradigms, i.e., intra- and inter-patient paradigms, respectively, where notable performance degradation is observed in the inter-patient paradigm. Compared to the overall accuracy, the average F1 score, sensitivity, and precision are significantly lower in the selected studies. To implement the DL-based ECG classification in real clinical scenarios, leveraging diverse ECG databases, designing advanced denoising and data augmentation techniques, integrating novel DL models, and deeper investigation in the inter-patient paradigm could be future research opportunities.