Convolutional neural network for classification of eight types of arrhythmia using 2D time–frequency feature map from standard 12-lead electrocardiogram
Ontology highlight
ABSTRACT:
SUBMITTER: Jeong D
PROVIDER: S-EPMC8516863 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:Most existing electrocardiogram (ECG) feature extraction methods rely on rule-based approaches. It is difficult to manually define all ECG features. We propose an unsupervised feature learning method using a convolutional variational autoencoder (CVAE) that can extract ECG features with unlabeled data. We used 596,000 ECG samples from 1,278 patients archived in biosignal databases from intensive care units to train the CVAE. Three external datasets were used for feature validation using two approaches. First, we explored the features without an additional training process. Clustering, latent space exploration, and anomaly detection were conducted. We confirmed that CVAE features reflected the various types of ECG rhythms. Second, we applied CVAE features to new tasks as input data and CVAE weights to weight initialization for different models for transfer learning for the classification of 12 types of arrhythmias. The f1-score for arrhythmia classification with extreme gradient boosting was 0.86 using CVAE features only. The f1-score of the model in which weights were initialized with the CVAE encoder was 5% better than that obtained with random initialization. Unsupervised feature learning with CVAE can extract the characteristics of various types of ECGs and can be an alternative to the feature extraction method for ECGs.
| S-EPMC8635334 | biostudies-literature
Project description:This newly inaugurated research database for 12-lead electrocardiogram signals was created under the auspices of Chapman University and Shaoxing People's Hospital (Shaoxing Hospital Zhejiang University School of Medicine) and aims to enable the scientific community in conducting new studies on arrhythmia and other cardiovascular conditions. Certain types of arrhythmias, such as atrial fibrillation, have a pronounced negative impact on public health, quality of life, and medical expenditures. As a non-invasive test, long term ECG monitoring is a major and vital diagnostic tool for detecting these conditions. This practice, however, generates large amounts of data, the analysis of which requires considerable time and effort by human experts. Advancement of modern machine learning and statistical tools can be trained on high quality, large data to achieve exceptional levels of automated diagnostic accuracy. Thus, we collected and disseminated this novel database that contains 12-lead ECGs of 10,646 patients with a 500?Hz sampling rate that features 11 common rhythms and 67 additional cardiovascular conditions, all labeled by professional experts. The dataset consists of 10-second, 12-dimension ECGs and labels for rhythms and other conditions for each subject. The dataset can be used to design, compare, and fine-tune new and classical statistical and machine learning techniques in studies focused on arrhythmia and other cardiovascular conditions.
| S-EPMC7016169 | biostudies-literature
Project description:Arrhythmia is a life-threatening cardiac condition characterized by irregular heart rhythm. Early and accurate detection is crucial for effective treatment. However, single-lead electrocardiogram (ECG) methods have limited sensitivity and specificity. This study propose an improved ensemble learning approach for arrhythmia detection using multi-lead ECG data. Proposed method, based on a boosting algorithm, namely Fine Tuned Boosting (FTBO) model detects multiple arrhythmia classes. For the feature extraction, introduce a new technique that utilizes a sliding window with a window size of 5 R-peaks. This study compared it with other models, including bagging and stacking, and assessed the impact of parameter tuning. Rigorous experiments on the MIT-BIH arrhythmia database focused on Premature Ventricular Contraction (PVC), Atrial Premature Contraction (PAC), and Atrial Fibrillation (AF) have been performed. The results showed that the proposed method achieved high sensitivity, specificity, and accuracy for all three classes of arrhythmia. It accurately detected Atrial Fibrillation (AF) with 100% sensitivity and specificity. For Premature Ventricular Contraction (PVC) detection, it achieved 99% sensitivity and specificity in both leads. Similarly, for Atrial Premature Contraction (PAC) detection, proposed method achieved almost 96% sensitivity and specificity in both leads. The proposed method shows great potential for early arrhythmia detection using multi-lead ECG data.
| S-EPMC11003640 | biostudies-literature
Project description:BackgroundThe standard configuration's set of twelve electrocardiogram (ECG) leads is optimal for the medical diagnosis of diverse cardiac conditions. However, it requires ten electrodes on the patient's limbs and chest, which is uncomfortable and cumbersome. Interlead conversion methods can reconstruct missing leads and enable more comfortable acquisitions, including in wearable devices, while still allowing for adequate diagnoses. Currently, methodologies for interlead ECG conversion either require multiple reference (input) leads and/or require input signals to be temporally aligned considering the ECG landmarks.MethodsUnlike the methods in the literature, this paper studies the possibility of converting ECG signals into all twelve standard configuration leads using signal segments from only one reference lead, without temporal alignment (blindly-segmented). The proposed methodology is based on a deep learning encoder-decoder U-Net architecture, which is compared with adaptations based on convolutional autoencoders and label refinement networks. Moreover, the method is explored for conversion with one single shared encoder or multiple individual encoders for each lead.ResultsDespite the more challenging settings, the proposed methodology was able to attain state-of-the-art level performance in multiple target leads, and both lead I and lead II seem especially suitable to convert certain sets of leads. In cross-database tests, the methodology offered promising results despite acquisition setup differences. Furthermore, results show that the presence of medical conditions does not have a considerable effect on the method's performance.ConclusionsThis study shows the feasibility of converting ECG signals using single-lead blindly-segmented inputs. Although the results are promising, further efforts should be devoted towards the improvement of the methodologies, especially the robustness to diverse acquisition setups, in order to be applicable to cardiac health monitoring in wearable devices and less obtrusive clinical scenarios.
| S-EPMC9710059 | biostudies-literature
Project description:BackgroundMost electrocardiogram (ECG) studies still take advantage of traditional statistical functions, and the results are mostly presented in tables, histograms, and curves. Few papers display ECG data by visual means. The aim of this study was to analyze and show data for electrocardiographic left ventricular hypertrophy (LVH) with ST-segment elevation (STE) by a heat map in order to explore the feasibility and clinical value of heat mapping for ECG data visualization.MethodsWe sequentially collected the electrocardiograms of inpatients in the First Affiliated Hospital of Shantou University Medical College from July 2015 to December 2015 in order to screen cases of LVH with STE. HemI 1.0 software was used to draw heat maps to display the STE of each lead of each collected ECG. Cluster analysis was carried out based on the heat map and the results were drawn as tree maps (pedigree maps) in the heat map.ResultsIn total, 60 cases of electrocardiographic LVH with STE were screened and analyzed. STE leads were mainly in the V1, V2 and V3 leads. The ST-segment shifts of each lead of each collected ECG could be conveniently visualized in the heat map. According to cluster analysis in the heat map, STE leads were clustered into two categories, comprising of the right precordial leads (V1, V2, V3) and others (V4, V5, V6, I, II, III, aVF, aVL, aVR). Moreover, the STE amplitude in 40% (24 out of 60) of cases reached the threshold specified in the STEMI guideline. These cases also could be fully displayed and visualized in the heat map. Cluster analysis in the heat map showed that the III, aVF and aVR leads could be clustered together, the V1, V2, V3 and V4 leads could be clustered together, and the V5, V6, I and aVL leads could be clustered together.ConclusionHeat maps and cluster analysis can be used to fully display every lead of each electrocardiogram and provide relatively comprehensive information.
| S-EPMC7281952 | biostudies-literature
Project description:18F-Fluorodeoxyglucose-positron emission tomography (FDG-PET) is commonly used in clinical practice and clinical drug development to identify and quantify metabolically active tumors. Manual or computer-assisted tumor segmentation in FDG-PET images is a common way to assess tumor burden, such approaches are both labor intensive and may suffer from high inter-reader variability. We propose an end-to-end method leveraging 2D and 3D convolutional neural networks to rapidly identify and segment tumors and to extract metabolic information in eyes to thighs (whole body) FDG-PET/CT scans. The developed architecture is computationally efficient and devised to accommodate the size of whole-body scans, the extreme imbalance between tumor burden and the volume of healthy tissue, and the heterogeneous nature of the input images. Our dataset consists of a total of 3664 eyes to thighs FDG-PET/CT scans, from multi-site clinical trials in patients with non-Hodgkin's lymphoma (NHL) and advanced non-small cell lung cancer (NSCLC). Tumors were segmented and reviewed by board-certified radiologists. We report a mean 3D Dice score of 88.6% on an NHL hold-out set of 1124 scans and a 93% sensitivity on 274 NSCLC hold-out scans. The method is a potential tool for radiologists to rapidly assess eyes to thighs FDG-avid tumor burden.
| S-EPMC7522127 | biostudies-literature
Project description:BackgroundThis is the first case report demonstrating the use of a smartphone device, enabling the diagnosis of an arrhythmia in the sports cardiology literature.Case summaryA 17-year-old semi-professional rugby player presented with recurrent episodes of palpitations terminated by vagal manoeuvres. The rugby player's resting 12-lead electrocardiogram (ECG), echocardiogram, and exercise stress test were normal. Due to his suggestive history and an ECG trace from a smartphone device, demonstrating a narrow complex tachycardia, an electrophysiological study was arranged. The study demonstrated a slow-fast atrioventricular nodal re-entrant tachycardia which was successfully ablated.DiscussionThe ambulatory use of a smartphone ECG device assisted in the timely diagnosis and management of an undiagnosed paroxysmal arrhythmia in a rugby player. This resulted in an expedited return to play.
| S-EPMC8142154 | biostudies-literature
Project description:A method is presented that modifies a 2mFobs - DFmodel ?A-weighted map such that the resulting map can strengthen a weak signal, if present, and can reduce model bias and noise. The method consists of first randomizing the starting map and filling in missing reflections using multiple methods. This is followed by restricting the map to regions with convincing density and the application of sharpening. The final map is then created by combining a series of histogram-equalized intermediate maps. In the test cases shown, the maps produced in this way are found to have increased interpretability and decreased model bias compared with the starting 2mFobs - DFmodel ?A-weighted map.
| S-EPMC4356370 | biostudies-literature
Project description:Background Accurate detection of arrhythmic events in the intensive care units (ICU) is of paramount significance in providing timely care. However, traditional ICU monitors generate a high rate of false alarms causing alarm fatigue. In this work, we develop an algorithm to improve life threatening arrhythmia detection in the ICUs using a deep learning approach. Methods and Results This study involves a total of 953 independent life-threatening arrhythmia alarms generated from the ICU bedside monitors of 410 patients. Specifically, we used the ECG (4 channels), arterial blood pressure, and photoplethysmograph signals to accurately detect the onset and offset of various arrhythmias, without prior knowledge of the alarm type. We used a hybrid convolutional neural network based classifier that fuses traditional handcrafted features with features automatically learned using convolutional neural networks. Further, the proposed architecture remains flexible to be adapted to various arrhythmic conditions as well as multiple physiological signals. Our hybrid- convolutional neural network approach achieved superior performance compared with methods which only used convolutional neural network. We evaluated our algorithm using 5-fold cross-validation for 5 times and obtained an accuracy of 87.5%±0.5%, and a score of 81%±0.9%. Independent evaluation of our algorithm on the publicly available PhysioNet 2015 Challenge database resulted in overall classification accuracy and score of 93.9% and 84.3%, respectively, indicating its efficacy and generalizability. Conclusions Our method accurately detects multiple arrhythmic conditions. Suitable translation of our algorithm may significantly improve the quality of care in ICUs by reducing the burden of false alarms.
| S-EPMC9075394 | biostudies-literature
Project description:BACKGROUND:Experimental studies have demonstrated that unstable repolarization dynamics is a risk factor of arrhythmia. We have recently developed an algorithm to detect QT interval (QTI) instability from the clinical electrocardiogram (ECG). OBJECTIVE:To develop a clinical arrhythmia risk stratification index based on the detection of QTI instability. METHODS:Intracardiac ECGs were recorded at rest in 114 patients with implanted implantable cardioverter-defibrillators (ICDs). Patients were followed up until appropriate implantable cardioverter-defibrillator therapy or death occurred, whichever came first. Each ECG recording was divided into 1-minute episodes (minECGs); the instability in QTI dynamics, if any, of each minECG was detected with our algorithm. An arrhythmia risk index termed QTI instability index (QTII) was defined as the number of minECGs with unstable QTI dynamics normalized by the number of minECGs with premature activations. The performance of QTII in arrhythmia risk stratification was examined with survival analysis and was compared with other risk indices, such as the mean RR interval (RRI), the standard deviation of the RRI and the QTI, and the frequency of premature activation. We hypothesized that the index QTII, which accounts for multiple risk factors and their interdependence, performs better than indices quantifying individual arrhythmia risk factors in the stratification of arrhythmia risk. RESULTS:The results of survival analysis show that QTII outperformed all other studied indices in arrhythmia risk stratification and was the only independent indicator of arrhythmia propensity in a multivariate survival model. CONCLUSION:QTII is a promising arrhythmia risk stratification index.
| S-EPMC3703156 | biostudies-literature