A novel method for detecting ST segment elevation myocardial infarction on a 12-lead electrocardiogram with a three-dimensional display

A periodic split attractor reconstruction method facilitates cardiovascular signal diagnoses and obstructive sleep apnea syndrome monitoring

Electrocardiogram (ECG) is a powerful tool to detect cardiovascular diseases (CVDs) and health conditions. We proposed a new method for evaluating ECG for efficient medical diagnosis in daily life. By splitting the signal according to the cardiac activity cycle, the periodic split attractor reconstruction (PSAR) method is proposed with time embedding, including three types of splitting methods to show its chaotic domain characteristics. We merged the CVDs dataset and the obstructive sleep apnea syndrome (OSAS) first-lead ECG signal dataset to validate the performance of PSAR for diagnosis and health monitoring using PSAR density maps as SE-ResNet input features. PSAR under 3 split methods showed different sensitivities for different CVDs. While in OSAS monitoring, PSAR showed good ability to recognize sleep abnormalities.

3D ECG display with deep learning approach for identification of cardiac abnormalities from a variable number of leads

Objective.The objective of this study is to explore new imaging techniques with the use of the deep learning method for the identification of cardiac abnormalities present in electrocardiogram (ECG) signals with 2, 3, 4, 6 and 12-lead in the framework of the PhysioNet/Computing in Cardiology Challenge 2021. The training set is a public database of 88,253 twelve-lead ECG recordings lasting from 6 s to 60 s. Each ECG recording has one or more diagnostic labels. The six-lead, four-lead, three-lead, and two-lead are reduced-lead versions of the original twelve-lead data.Approach.The deep learning method considers images that are built from raw ECG signals. This technique considers innovative 3D display of the entire ECG signal, observing the regional constraints of the leads, obtaining time-spatial images of the 12 leads, where the x-axis is the temporal evolution of ECG signal, the y-axis is the spatial location of the leads, and the z-axis (color) the amplitude. These images are used for training Convolutional Neural Networks with GoogleNet for ECG diagnostic classification.Main results.The official results of the classification accuracy of our team named 'Gio_new_img' received scores of 0.4, 0.4, 0.39, 0.4 and 0.4 (ranked 18th, 18th, 18th,18th, 18th out of 39 teams) for the 12-lead, 6-lead, 4-lead, 3-lead, and 2-lead versions of the hidden test set with the Challenge evaluation metric.Significance.The results indicated that all these algorithms have similar behaviour in the various lead groups, and the most surprising and interesting point is the fact that the 2-lead scores are similar to those obtained with the analysis of 12 leads. It permitted to test the diagnostic potential of the reduced-lead ECG recordings. These aspects can be related to the pattern recognition capacity and generalizability of the deep learning approach and/or to the fact that the characteristics of the considered cardiac abnormalities can be extracted also from a reduced set of leads.

Multilevel hybrid accurate handcrafted model for myocardial infarction classification using ECG signals

Myocardial infarction (MI) is detected using electrocardiography (ECG) signals. Machine learning (ML) models have been used for automated MI detection on ECG signals. Deep learning models generally yield high classification performance but are computationally intensive. We have developed a novel multilevel hybrid feature extraction-based classification model with low time complexity for MI classification. The study dataset comprising 12-lead ECGs belonging to one healthy and 10 MI classes were downloaded from a public ECG signal databank. The model architecture comprised multilevel hybrid feature extraction, iterative feature selection, classification, and iterative majority voting (IMV). In the hybrid handcrafted feature (HHF) generation phase, both textural and statistical feature extraction functions were used to extract features from ECG beats but only at a low level. A new pooling-based multilevel decomposition model was presented to enable them to create features at a high level. This model used average and maximum pooling to create decomposed signals. Using these pooling functions, an unbalanced tree was obtained. Therefore, this model was named multilevel unbalanced pooling tree transformation (MUPTT). On the feature extraction side, two extractors (functions) were used to generate both statistical and textural features. To generate statistical features, 20 commonly used moments were used. A new, improved symmetric binary pattern function was proposed to generate textural features. Both feature extractors were applied to the original MI signal and the decomposed signals generated by the MUPTT. The most valuable features from among the extracted feature vectors were selected using iterative neighborhood component analysis (INCA). In the classification phase, a one-dimensional nearest neighbor classifier with ten-fold cross-validation was used to obtain lead-wise results. The computed lead-wise results derived from all 12 leads of the same beat were input to the IMV algorithm to generate ten voted results. The most representative was chosen using a greedy technique to calculate the overall classification performance of the model. The HHF-MUPTT-based ECG beat classification model attained excellent performance, with the best lead-wise accuracy of 99.85% observed in Lead III and 99.94% classification accuracy using the IMV algorithm. The results confirmed the high MI classification ability of the presented computationally lightweight HHF-MUPTT-based model.

Automatic arrhythmia detection with multi-lead ECG signals based on heterogeneous graph attention networks

Automatic arrhythmia detection is very important for cardiovascular health. It is generally performed by measuring the electrocardiogram (ECG) signals of standard multiple leads. However, the correlations of multiple leads are often ignored. In addition, an extensive and complex feature extraction process is usually needed in most existing studies. Therefore, these challenges will not only lead to the loss of overall lead information, but also cause the detection performance to depend on the quality of features. To solve these challenges, a novel multi-lead arrhythmia detection model based on a heterogeneous graph attention network is proposed in this paper. We have modeled the multi-lead data as a heterogeneous graph to integrate diverse information and construct intra-lead and inter-lead correlations in multi-lead data, providing a reasonable and effective the data model. A heterogeneous graph network with a dual-level attention strategy has been utilized to capture the interactions among diverse information and information types. At the same time, our model does not require any feature extraction process for the ECG signals, which avoids out complex feature engineering. Extensive experimental results show that multi-lead information and complex correlations can be well captured, thus confirming that the proposed model results in significant improvements in multi-lead arrhythmia detection.

A visually interpretable detection method combines 3-D ECG with a multi-VGG neural network for myocardial infarction identification

BACKGROUND AND OBJECTIVE

The automatic recognition of myocardial infarction (MI) by artificial intelligence (AI) has been an emerging topic of academic research and an existing classification method that can recognize conventional electrocardiogram (ECG) signals with high accuracy. However, they are employed to classify one-dimensional (1-D) ECG signals rather than three-dimensional (3-D) ECG images, and it is limited to provide physicians with significant recommendations to aid in diagnosis like highlighting abnormal leads. Other studies on 3-D ECG images either did not achieve high accuracy or did not employ an inter-patient classification scheme. By proposing a multi-VGG deep neural network, this study aims to develop an automatic classification method for identifying myocardial infarction with inter-patient high accuracy and proper interpretability using 3-D ECG image and a Grad-CAM++ method.

METHODS

We apply a multi-VGG deep convolutional neural network to top-view images of 3-D ECG, which are generated from simply denoised standard 12 leads ECG signals for classification. The multi-network method, which separately classifies QRS areas, ST areas, and whole heartbeats, was applied to improve classification performance. Furthermore, the Grad-CAM++ method was used to provide visually interpretable heatmaps for user's attention to improve network interpretability and assist physicians in MI diagnosis RESULTS: The proposed method achieved 95.65% inter-patient accuracy and exactly perfect inner-patient accuracy in the Physikalisch-Technische Bundesanstalt (PTB) diagnostic ECG database experiment. In the PTB-XL diagnostic ECG database, the proposed method achieved 97.23% inter-patient accuracy. The Grad-CAM++ experiment results also showed that the highlighted areas matched the medical diagnosis criteria for myocardial infarction.

CONCLUSIONS

Our method demonstrates that 3-D ECG images with AI classification can be efficiently employed for heart disease diagnosis with both high accuracy and visual interpretability.

Interpretable Detection and Location of Myocardial Infarction Based on Ventricular Fusion Rule Features

Myocardial infarction (MI) is one of the most common cardiovascular diseases threatening human life. In order to accurately distinguish myocardial infarction and have a good interpretability, the classification method that combines rule features and ventricular activity features is proposed in this paper. Specifically, according to the clinical diagnosis rule and the pathological changes of myocardial infarction on the electrocardiogram, the local information extracted from the Q wave, ST segment, and T wave is computed as the rule feature. All samples of the QT segment are extracted as ventricular activity features. Then, in order to reduce the computational complexity of the ventricular activity features, the effects of Discrete Wavelet Transform (DWT), Principal Component Analysis (PCA), and Locality Preserving Projections (LPP) on the extracted ventricular activity features are compared. Combining rule features and ventricular activity features, all the 12 leads features are fused as the ultimate feature vector. Finally, eXtreme Gradient Boosting (XGBoost) is used to identify myocardial infarction, and the overall accuracy rate of 99.86% is obtained on the Physikalisch-Technische Bundesanstalt (PTB) database. This method has a good medical diagnosis basis while improving the accuracy, which is very important for clinical decision-making.

A review of risk prediction models in cardiovascular disease: conventional approach vs. artificial intelligent approach

Cardiovascular disease (CVD) is the leading cause of death worldwide and is a global health issue. Traditionally, statistical models are used commonly in the risk prediction and assessment of CVD. However, the adoption of artificial intelligent (AI) approach is rapidly taking hold in the current era of technology to evaluate patient risks and predict the outcome of CVD. In this review, we outline various conventional risk scores and prediction models and do a comparison with the AI approach. The strengths and limitations of both conventional and AI approaches are discussed. Besides that, biomarker discovery related to CVD are also elucidated as the biomarkers can be used in the risk stratification as well as early detection of the disease. Moreover, problems and challenges involved in current CVD studies are explored. Lastly, future prospects of CVD risk prediction and assessment in the multi-modality of big data integrative approaches are proposed.