0.90 (N), 0.81 (A), 0.72 (O), and 0.55 (P). Particular F1 scores within training set are 0.91 (N), 0.85 (A), 0.76 (O), and 0.73 (P).
These study results contribute to multidisciplinary areas, focusing on creation of robust and reliable cardiac monitoring systems in order to improve diagnosis, reduce unnecessary time-consuming expert ECG scoring and, consequently, ensure timely and effective treatment.
We present a novel wavelet-based ECG delineation method with robust classification of P wave and T wave. The work is aimed on an adaptation of the method to long-term experimental electrograms (EGs) measured on isolated rabbit heart and to evaluate the effect of global ischemia in experimental EGs on delineation performance. The algorithm was tested on a set of 263 rabbit EGs with established reference points and on human signals using standard Common Standards for Quantitative Electrocardiography Standard Database (CSEDB). On CSEDB, standard deviation (SD) of measured errors satisfies given criterions in each point and the results are comparable to other published works. In rabbit signals, our QRS detector reached sensitivity of 99.87% and positive predictivity of 99.89% despite an overlay of spectral components of QRS complex, P wave and power line noise. The algorithm shows great performance in suppressing J-point elevation and reached low overall error in both, QRS onset (SD = 2.8 ms) and QRS offset (SD = 4.3 ms) delineation. T wave offset is detected with acceptable error (SD = 12.9 ms) and sensitivity nearly 99%. Variance of the errors during global ischemia remains relatively stable, however more failures in detection of T wave and P wave occur. Due to differences in spectral and timing characteristics parameters of rabbit based algorithm have to be highly adaptable and set more precisely than in human ECG signals to reach acceptable performance.
The latest trends in clinical care and telemedicine suggest a demand for a reliable automated electrocardiogram (ECG) signal classification methods. In this paper, we present customized deep learning model for ECG classification as a part of the Physionet/CinC Challenge 2020. The method is based on modified ResNet type convolutional neural network and is capable to automatically recognize 24 cardiac abnormalities using 12-lead ECG. We have adopted several preprocessing and learning techniques including custom tailored loss function, dedicated classification layer and Bayesian threshold optimization which have major positive impact on the model performance. At the official phase of the Challenge, our team-BUTTeam-reached a challenge validation score of 0.696, and the full test score of 0.202, placing us 21 out of 40 in the official ranking. This implies that our method performed well on data from the same source (reached first place with validation score), however, it has very poor generalization to data from different sources.
BackgroundDetailed quantitative analysis of the effect of left ventricle (LV) hypertrophy on myocardial ischemia manifestation in ECG is still missing. The associations between both phenomena can be studied in animal models. In this study, rabbit isolated hearts with spontaneously increased LV mass were used to evaluate the effect of such LV alteration on ischemia detection criteria and performance.MethodsElectrophysiological effects of increased LV mass were evaluated on sixteen New Zealand rabbit isolated hearts under non-ischemic and ischemic conditions by analysis of various electrogram (EG) parameters. To reveal hearts with increased LV mass, LV weight/heart weight ratio was proposed. Standard paired and unpaired statistical tests and receiver operating characteristics analysis were used to compare data derived from different groups of animals, monitor EG parameters during global ischemia and evaluate their ability to discriminate between unchanged and increased LV as well as non-ischemic and ischemic state.ResultsSuccessful evaluation of both increased LV mass and ischemia is lead-dependent. Particularly, maximal deviation of QRS and area under QRS associated with anterolateral heart wall respond significantly to even early phase (the 1st-3rd min) of ischemia. Besides ischemia, these parameters reflect increased LV mass as well (with sensitivity reaching approx. 80%). However, the sensitivity of the parameters to both phenomena may lead to misinterpretations, when inappropriate criteria for ischemia detection are selected. Particularly, use of cut-off-based criteria defined from control group for ischemia detection in hearts with increased LV mass may result in dramatic reduction (approx. 15%) of detection specificity due to increased number of false positives. Nevertheless, criteria adjusted to particular experimental group allow achieving ischemia detection sensitivity of 89–100% and specificity of 94–100%, respectively.ConclusionsIt was shown that response of the heart to myocardial ischemia can be successfully evaluated only when taking into account heart-related factors (such as LV mass) and other methodological aspects (such as recording electrodes position, selected EG parameters, cut-off criteria, etc.). Results of this study might be helpful for developing new clinical diagnostic strategies in order to improve myocardial ischemia detection in patients with LV hypertrophy.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.