O. Özgül, B. Hermans, A. Hunnik, S. Verheule, U. Schotten, P. Bonizzi, S. Zeemering
Highly complex and irregular atrial activation patterns during atrial fibrillation $(AF)$ can occasionally be interrupted by repetitive atrial activation patterns (RAAPs). These patterns are thought to be generated by mechanisms that initiate or maintain $AF$ episodes are therefore, might be more diverse in patients with more complex forms of $AF$ We quantified RAAP diversity by the half decay time of the ratio of the unprecedented RAAPs to the total num{###} ${it ber}$ of RAAPs in a goat model with different durations of sustained $AF[3$ weeks $(3wkAF, n=8)$ and 22 weeks $(22wkAF, n=8)]$. 32 recordings from left and right atria (LA/RA) of each goat were analyzed. 24 out of 32 curves could be modeled as exponential decay functions with adjusted R-squared $> 0.75$ while others presented more irregular decaying patterns $(3wkAF LA:2 RA:3,22wkAF$ LA: $1 RA:2)$. Half decay rates were significantly shorter in $LAs$ of $3wkAF$ goats $(delta_{3wkAF}=23.67s vs. delta_{22wkAF}=32.86s,p < 0.05$, Mann-Whitney U-test). There was no significant difference in RA.
{"title":"Incidence of Distinct Repetitive Atrial Activation Patterns as a Metric for Atrial Fibrillation Complexity","authors":"O. Özgül, B. Hermans, A. Hunnik, S. Verheule, U. Schotten, P. Bonizzi, S. Zeemering","doi":"10.22489/CinC.2022.394","DOIUrl":"https://doi.org/10.22489/CinC.2022.394","url":null,"abstract":"Highly complex and irregular atrial activation patterns during atrial fibrillation <tex>$(AF)$</tex> can occasionally be interrupted by repetitive atrial activation patterns (RAAPs). These patterns are thought to be generated by mechanisms that initiate or maintain <tex>$AF$</tex> episodes are therefore, might be more diverse in patients with more complex forms of <tex>$AF$</tex> We quantified RAAP diversity by the half decay time of the ratio of the unprecedented RAAPs to the total num{###} <tex>${it ber}$</tex> of RAAPs in a goat model with different durations of sustained <tex>$AF[3$</tex> weeks <tex>$(3wkAF, n=8)$</tex> and 22 weeks <tex>$(22wkAF, n=8)]$</tex>. 32 recordings from left and right atria (LA/RA) of each goat were analyzed. 24 out of 32 curves could be modeled as exponential decay functions with adjusted R-squared <tex>$> 0.75$</tex> while others presented more irregular decaying patterns <tex>$(3wkAF LA:2 RA:3,22wkAF$</tex> LA: <tex>$1 RA:2)$</tex>. Half decay rates were significantly shorter in <tex>$LAs$</tex> of <tex>$3wkAF$</tex> goats <tex>$(delta_{3wkAF}=23.67s vs. delta_{22wkAF}=32.86s,p < 0.05$</tex>, Mann-Whitney U-test). There was no significant difference in RA.","PeriodicalId":117840,"journal":{"name":"2022 Computing in Cardiology (CinC)","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131404610","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The medical system has been targeted by the cyber attackers, who aim to bring down the health security critical infrastructure. This research is motivated by the recent cyber-attacks happened during COVID 19 pandemics which resulted in the compromise of the diagnosis results. This study was carried to demonstrate how the medical systems can be penetrated using AI-based Directory Discovery Attack and present security solutions to counteract such attacks. We then followed the NIST (National Institute of Standards and Technology) ethical hacking methodology to launch the AI-based Directory Discovery Attack. We were able to successfully penetrate the system and gain access to the core of the medical directories. We then proposed a series of security solutions to prevent such cyber-attacks.
{"title":"AI Based Directory Discovery Attack and Prevention of the Medical Systems","authors":"Ying He, Cunjin Luo, Jiyuan Zheng, Kuanquan Wang, Heng-Di Zhang","doi":"10.22489/CinC.2022.401","DOIUrl":"https://doi.org/10.22489/CinC.2022.401","url":null,"abstract":"The medical system has been targeted by the cyber attackers, who aim to bring down the health security critical infrastructure. This research is motivated by the recent cyber-attacks happened during COVID 19 pandemics which resulted in the compromise of the diagnosis results. This study was carried to demonstrate how the medical systems can be penetrated using AI-based Directory Discovery Attack and present security solutions to counteract such attacks. We then followed the NIST (National Institute of Standards and Technology) ethical hacking methodology to launch the AI-based Directory Discovery Attack. We were able to successfully penetrate the system and gain access to the core of the medical directories. We then proposed a series of security solutions to prevent such cyber-attacks.","PeriodicalId":117840,"journal":{"name":"2022 Computing in Cardiology (CinC)","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127567536","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Maximiliano Mollura, Christian Niklas, Stefanie Messner, M. Weigand, J. Larmann, R. Barbieri
Introduction: The high mortality and difficulty of diagnosis make Heart failure $(HF)$ a severe burden for the healthcare system, especially in intensive care units $(ICU)$. Goal: This work proposes a method to characterize $HF$ patients using autonomic indices from electrocardiogram $(ECG)$ recordings in the $ICU$ Methods: We considered 52 $ICU$ patients from the MIMIC-III database subjected to brain natriuretic peptide (NT-proBNP) laboratory measurement during their stay, of which 41 showed a positive reading for likely $HF$ due to elevated levels of the peptide $(NT-proBNP > 300 pg/mL)$. RR intervals from 1 hour $ECG$ recordings in the hour preceding NT-proBNP measurements were selected, and a point process framework was applied to extract time-varying estimates of indices related to autonomic nervous system activity. A general linear mixed-effects model was used to analyze the dynamics of the two populations.Results: Results showed an increasing average $RR$ interval in the negative population $(p < 0.001)$. In parallel, $RR$ variability increased in negative subjects $(p < 0.001)$ and decreased in positive patients $(p < 0.001)$. High frequency power $(p < 0.001)$ further showed different dynamics between the two populations. Conclusions: Results point at different autonomic cardiac control dynamics in patients with positive NT-proBNP test in the hour preceding the measurement.
{"title":"Characterization of Heart Rate Variability Dynamics in Heart Failure Patients Admitted to Intensive Care Unit","authors":"Maximiliano Mollura, Christian Niklas, Stefanie Messner, M. Weigand, J. Larmann, R. Barbieri","doi":"10.22489/CinC.2022.209","DOIUrl":"https://doi.org/10.22489/CinC.2022.209","url":null,"abstract":"Introduction: The high mortality and difficulty of diagnosis make Heart failure $(HF)$ a severe burden for the healthcare system, especially in intensive care units $(ICU)$. Goal: This work proposes a method to characterize $HF$ patients using autonomic indices from electrocardiogram $(ECG)$ recordings in the $ICU$ Methods: We considered 52 $ICU$ patients from the MIMIC-III database subjected to brain natriuretic peptide (NT-proBNP) laboratory measurement during their stay, of which 41 showed a positive reading for likely $HF$ due to elevated levels of the peptide $(NT-proBNP > 300 pg/mL)$. RR intervals from 1 hour $ECG$ recordings in the hour preceding NT-proBNP measurements were selected, and a point process framework was applied to extract time-varying estimates of indices related to autonomic nervous system activity. A general linear mixed-effects model was used to analyze the dynamics of the two populations.Results: Results showed an increasing average $RR$ interval in the negative population $(p < 0.001)$. In parallel, $RR$ variability increased in negative subjects $(p < 0.001)$ and decreased in positive patients $(p < 0.001)$. High frequency power $(p < 0.001)$ further showed different dynamics between the two populations. Conclusions: Results point at different autonomic cardiac control dynamics in patients with positive NT-proBNP test in the hour preceding the measurement.","PeriodicalId":117840,"journal":{"name":"2022 Computing in Cardiology (CinC)","volume":"42 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132587653","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
An action potential (AP) is an alteration in the membrane potential of an excitable cell. It occurs due to the size, shape, and type of cell excited. In literature, various differential equation (DE) based mathematical models have been proposed to emulate APs. More recently, a Fourier Series (FS) based technique has been proposed. This paper discusses the methodology to identify the parameters of the FS model for eventual implementation on an FPGA. Four DE models have been investigated. Two implementation techniques - direct digital synthesis (DDS) and double integrator based resonant model (RM) - have been compared in terms of FPGA resource usage. Our observations show that the FS model is an attractive alternative to the DE models. The FS implemented using the RM technique offers good accuracy with ease of FPGA implementation. The FS model has the potential for real-time tissues level emulation on an FPGA.
{"title":"Emulation of Biological Cells","authors":"Jerry Jacob, Nitish D. Patel, Sucheta Sehgal","doi":"10.22489/CinC.2022.245","DOIUrl":"https://doi.org/10.22489/CinC.2022.245","url":null,"abstract":"An action potential (AP) is an alteration in the membrane potential of an excitable cell. It occurs due to the size, shape, and type of cell excited. In literature, various differential equation (DE) based mathematical models have been proposed to emulate APs. More recently, a Fourier Series (FS) based technique has been proposed. This paper discusses the methodology to identify the parameters of the FS model for eventual implementation on an FPGA. Four DE models have been investigated. Two implementation techniques - direct digital synthesis (DDS) and double integrator based resonant model (RM) - have been compared in terms of FPGA resource usage. Our observations show that the FS model is an attractive alternative to the DE models. The FS implemented using the RM technique offers good accuracy with ease of FPGA implementation. The FS model has the potential for real-time tissues level emulation on an FPGA.","PeriodicalId":117840,"journal":{"name":"2022 Computing in Cardiology (CinC)","volume":"79 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132619230","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Cardiac auscultation is an effective method to screen hemodynamic abnormalities. As part of the George B. Moody PhysioNet Challenge 2022, this paper aims to propose an automated algorithm to identify the presence of murmurs in heart sounds from multiple auscultation locations and to determine whether the heart sounds signal is normal. Two methods are explored. In method one, we perform a series of pre-processing such as denoising and segmentation of the heart sounds signal, extract Log Mel-Spectrogram as features, and use fastai's built-in xResNet 18 pre-trained model for classification. In method two, we extract Mel-frequency cepstral coefficients (MFCCs) as features without any pre-processing and build a customized model based on deep residual networks using one-dimensional convolutional neural layers. Our team, USST_ Med, received a challenging score of weighted accuracy of 0.642 (ranked 26th out of 40 teams) and cost of 14529 (ranked 30th out of 39 teams) on the final hidden test set.
心脏听诊是筛查血流动力学异常的有效方法。作为George B. Moody PhysioNet Challenge 2022的一部分，本文旨在提出一种自动算法，以识别来自多个听诊位置的心音中是否存在杂音，并确定心音信号是否正常。本文探讨了两种方法。方法一是对心音信号进行去噪和分割等一系列预处理，提取Log Mel-Spectrogram作为特征，并使用fastai内置的xResNet 18预训练模型进行分类。在方法二中，我们在不进行任何预处理的情况下提取Mel-frequency cepstral系数(MFCCs)作为特征，并使用一维卷积神经层构建基于深度残差网络的定制模型。我们的团队USST_ Med在最终的隐藏测试集中获得了具有挑战性的分数，加权准确率为0.642(在40支球队中排名第26)，成本为14529(在39支球队中排名第30)。
{"title":"Detection of Murmurs from Heart Sound Recordings with Deep Residual Networks","authors":"Leigang Hu, Wenjie Cai, Xinyue Li, Jia Li","doi":"10.22489/CinC.2022.047","DOIUrl":"https://doi.org/10.22489/CinC.2022.047","url":null,"abstract":"Cardiac auscultation is an effective method to screen hemodynamic abnormalities. As part of the George B. Moody PhysioNet Challenge 2022, this paper aims to propose an automated algorithm to identify the presence of murmurs in heart sounds from multiple auscultation locations and to determine whether the heart sounds signal is normal. Two methods are explored. In method one, we perform a series of pre-processing such as denoising and segmentation of the heart sounds signal, extract Log Mel-Spectrogram as features, and use fastai's built-in xResNet 18 pre-trained model for classification. In method two, we extract Mel-frequency cepstral coefficients (MFCCs) as features without any pre-processing and build a customized model based on deep residual networks using one-dimensional convolutional neural layers. Our team, USST_ Med, received a challenging score of weighted accuracy of 0.642 (ranked 26th out of 40 teams) and cost of 14529 (ranked 30th out of 39 teams) on the final hidden test set.","PeriodicalId":117840,"journal":{"name":"2022 Computing in Cardiology (CinC)","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133199282","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
N. Biasi, Matteo Mercati, Paolo Seghetti, A. Tognetti
In this work, we developed a closed loop model of the interaction between the heart and a cardiac pacemaker. The main novelty of our framework is the employment of a reaction-diffusion heart model, which could enhance the assessment of cardiac pacing. Additionally, we provided a specific hardware setup for the deployment of our frame-work. Our results show that the heart model reproduces the healthy activation sequence and is feasible for closed loop simulations. Furthermore, we successfully simulated the interaction between heart and pacemaker models during the insurgence of endless loop tachycardia. Finally, we believe that our closed loop system could be an effective supporting tool to evaluate the safety and efficacy of the therapeutic effect of cardiac pacemakers.
{"title":"Electrophysiological Closed Loop Model of the Heart as Supporting Tool for Cardiac Pacing","authors":"N. Biasi, Matteo Mercati, Paolo Seghetti, A. Tognetti","doi":"10.22489/CinC.2022.090","DOIUrl":"https://doi.org/10.22489/CinC.2022.090","url":null,"abstract":"In this work, we developed a closed loop model of the interaction between the heart and a cardiac pacemaker. The main novelty of our framework is the employment of a reaction-diffusion heart model, which could enhance the assessment of cardiac pacing. Additionally, we provided a specific hardware setup for the deployment of our frame-work. Our results show that the heart model reproduces the healthy activation sequence and is feasible for closed loop simulations. Furthermore, we successfully simulated the interaction between heart and pacemaker models during the insurgence of endless loop tachycardia. Finally, we believe that our closed loop system could be an effective supporting tool to evaluate the safety and efficacy of the therapeutic effect of cardiac pacemakers.","PeriodicalId":117840,"journal":{"name":"2022 Computing in Cardiology (CinC)","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133239992","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Mikel Landajuela, R. Anirudh, Joe Loscazo, R. Blake
Current state-of-the-art techniques for non-invasive imaging of cardiac electrical phenomena require voltage recordings from dozens of different torso locations and anatomical models built from expensive medical diagnostic imaging procedures. This study aimed to assess if recent machine learning advances could alternatively reconstruct electroanatomical maps at clinically relevant resolutions using only the standard 12-lead electrocardiogram (ECG) as input. To that end, a computational study was conducted to generate a dataset of over 16000 detailed cardiac simulations, which was then used to train neural network (NN) architectures designed to exploit both spatial and temporal correlations in the ECG signal. Analysis over a validation set showed average errors in activation map reconstruction below 1.7 msec over 75 intracardiac locations. Furthermore, phenotypical patterns of activation and the morphology of the activation potential were correctly reconstructed. The approach offers opportunities to stratify patients non-invasively, both retrospectively and prospectively, using metrics otherwise only available through invasive clinical procedures.
{"title":"Intracardiac Electrical Imaging Using the 12-Lead ECG: A Machine Learning Approach Using Synthetic Data","authors":"Mikel Landajuela, R. Anirudh, Joe Loscazo, R. Blake","doi":"10.22489/CinC.2022.026","DOIUrl":"https://doi.org/10.22489/CinC.2022.026","url":null,"abstract":"Current state-of-the-art techniques for non-invasive imaging of cardiac electrical phenomena require voltage recordings from dozens of different torso locations and anatomical models built from expensive medical diagnostic imaging procedures. This study aimed to assess if recent machine learning advances could alternatively reconstruct electroanatomical maps at clinically relevant resolutions using only the standard 12-lead electrocardiogram (ECG) as input. To that end, a computational study was conducted to generate a dataset of over 16000 detailed cardiac simulations, which was then used to train neural network (NN) architectures designed to exploit both spatial and temporal correlations in the ECG signal. Analysis over a validation set showed average errors in activation map reconstruction below 1.7 msec over 75 intracardiac locations. Furthermore, phenotypical patterns of activation and the morphology of the activation potential were correctly reconstructed. The approach offers opportunities to stratify patients non-invasively, both retrospectively and prospectively, using metrics otherwise only available through invasive clinical procedures.","PeriodicalId":117840,"journal":{"name":"2022 Computing in Cardiology (CinC)","volume":"40 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131937151","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Myocardial Ischemia (MI) is a fatal heart condition due to insufficient blood flow in the heart muscles, which may cause unexpected heart attacks. Exercise Stress Test (EST) Electrocardiogram (ECG) is a non-invasive diagnostic procedure that can help identify various disease conditions, including MI. This study aims to classify the ischemic and non-ischemic EST ECG using Machine Learning (ML) algorithms. EST ECGs for 152 patients (n=53 female) of mean age ($50 pm 11.92$ years) were used in this study. ST morphology changes, measured during pre-load, load, and recovery at $J+(40$, 60, and 80 ms) were utilized as input to 14 ML classifiers. 70% of the input data to the ML classifiers were considered as train data, and 30% of the input data as test. Random Forest (RF) was selected based on the most suitable output and was used to classify between ischemic and non-ischemic by considering the clinical features such as ST variations, Blood Pressure (BP), Metabolic equivalent (Mets), and Rate Pressure Product (RPP) as input for both lead-II and V5. The model accuracy, sensitivity, precision, and F1 score for lead-II were 93%, 89.17%, 93%, and 89.63%, respectively. For V5, the performance matrices were 91%, 80%, 95%, and 86.14%, respectively.
{"title":"Machine Learning-based Classification of Ischemic and Non-Ischemic Exercise Stress Test ECG","authors":"Dibya Chowdhury, B. Neelapu, K. Pal, J. Sivaraman","doi":"10.22489/CinC.2022.276","DOIUrl":"https://doi.org/10.22489/CinC.2022.276","url":null,"abstract":"Myocardial Ischemia (MI) is a fatal heart condition due to insufficient blood flow in the heart muscles, which may cause unexpected heart attacks. Exercise Stress Test (EST) Electrocardiogram (ECG) is a non-invasive diagnostic procedure that can help identify various disease conditions, including MI. This study aims to classify the ischemic and non-ischemic EST ECG using Machine Learning (ML) algorithms. EST ECGs for 152 patients (n=53 female) of mean age ($50 pm 11.92$ years) were used in this study. ST morphology changes, measured during pre-load, load, and recovery at $J+(40$, 60, and 80 ms) were utilized as input to 14 ML classifiers. 70% of the input data to the ML classifiers were considered as train data, and 30% of the input data as test. Random Forest (RF) was selected based on the most suitable output and was used to classify between ischemic and non-ischemic by considering the clinical features such as ST variations, Blood Pressure (BP), Metabolic equivalent (Mets), and Rate Pressure Product (RPP) as input for both lead-II and V5. The model accuracy, sensitivity, precision, and F1 score for lead-II were 93%, 89.17%, 93%, and 89.63%, respectively. For V5, the performance matrices were 91%, 80%, 95%, and 86.14%, respectively.","PeriodicalId":117840,"journal":{"name":"2022 Computing in Cardiology (CinC)","volume":"56 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133048047","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Atrial fibrillation (AF) is the most frequent atrial rhythm disorder with an incidence increasing with age. Genetic mutations impairing the normal functioning of IKr and Ito channels are implicated in AF outbreaks in healthy patients. The higher susceptibility to AF in presence of KCNH2 T436M, KNCH2 T895M and KCNE3-V17M mutations was previously studied by simulating their effects on atrial electrophysiology in single-cell and tissue. This work aims at extending the previous study to a 3D hiatrial model to assess vulnerability to AF initiation and maintenance on a complex geometry. A realistic model of human atria was used to run 3D simulations and study temporal vulnerability. After stabilization, a train of stimuli was applied to the coronary sinus region to simulate an ectopic stimulus and to induce arrhythmia. The results show a higher susceptibility of the mutant atria to develop arrhythmias in a mutation-dependent fashion. The KCNE3-V17M variant was the most proarrhythmogenic with a 24ms-wide vulnerable window(VW) and instable arrhythmic patterns. The KCNH2 T895M and KCNH2 T436M mutations presented a VW of 7ms and 10ms, respectively, with mainly macro re-entries. These findings highlight the different effects of the genetic mutations and the importance of a patient-specific approach.
{"title":"Computational Study of the Effects of AF-related Genetic Mutations in 3D Human Atrial Model","authors":"Rebecca Belletti, L. Romero, J. Saiz","doi":"10.22489/CinC.2022.070","DOIUrl":"https://doi.org/10.22489/CinC.2022.070","url":null,"abstract":"Atrial fibrillation (AF) is the most frequent atrial rhythm disorder with an incidence increasing with age. Genetic mutations impairing the normal functioning of IKr and Ito channels are implicated in AF outbreaks in healthy patients. The higher susceptibility to AF in presence of KCNH2 T436M, KNCH2 T895M and KCNE3-V17M mutations was previously studied by simulating their effects on atrial electrophysiology in single-cell and tissue. This work aims at extending the previous study to a 3D hiatrial model to assess vulnerability to AF initiation and maintenance on a complex geometry. A realistic model of human atria was used to run 3D simulations and study temporal vulnerability. After stabilization, a train of stimuli was applied to the coronary sinus region to simulate an ectopic stimulus and to induce arrhythmia. The results show a higher susceptibility of the mutant atria to develop arrhythmias in a mutation-dependent fashion. The KCNE3-V17M variant was the most proarrhythmogenic with a 24ms-wide vulnerable window(VW) and instable arrhythmic patterns. The KCNH2 T895M and KCNH2 T436M mutations presented a VW of 7ms and 10ms, respectively, with mainly macro re-entries. These findings highlight the different effects of the genetic mutations and the importance of a patient-specific approach.","PeriodicalId":117840,"journal":{"name":"2022 Computing in Cardiology (CinC)","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133804312","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Alex Gaudio, Miguel Coimbra, A. Campilho, A. Smailagic, S. Schmidt, F. Renna
Late diagnoses of patients affected by pulmonary artery hypertension (PH) have a poor outcome. This observation has led to a call for earlier, non-invasive PH detection. Cardiac auscultation offers a non-invasive and cost-effective alternative to both right heart catheterization and doppler analysis in analysis of PH. We propose to detect PH via analysis of digital heart sound recordings with over-parameterized deep neural networks. In contrast with previous approaches in the literature, we assess the impact of a pre-processing step aiming to separate S2 sound into the aortic (A2) and pulmonary (P2) components. We obtain an area under the ROC curve of. 95, improving over our adaptation of a state-of-the-art Gaussian mixture model PH detector by +.17. Post-hoc explanations and analysis show that the availability of separated A2 and P2 components contributes significantly to prediction. Analysis of stethoscope heart sound recordings with deep networks is an effective, low-cost and non-invasive solution for the detection of pulmonary hypertension.
{"title":"Explainable Deep Learning for Non-Invasive Detection of Pulmonary Artery Hypertension from Heart Sounds","authors":"Alex Gaudio, Miguel Coimbra, A. Campilho, A. Smailagic, S. Schmidt, F. Renna","doi":"10.22489/CinC.2022.295","DOIUrl":"https://doi.org/10.22489/CinC.2022.295","url":null,"abstract":"Late diagnoses of patients affected by pulmonary artery hypertension (PH) have a poor outcome. This observation has led to a call for earlier, non-invasive PH detection. Cardiac auscultation offers a non-invasive and cost-effective alternative to both right heart catheterization and doppler analysis in analysis of PH. We propose to detect PH via analysis of digital heart sound recordings with over-parameterized deep neural networks. In contrast with previous approaches in the literature, we assess the impact of a pre-processing step aiming to separate S2 sound into the aortic (A2) and pulmonary (P2) components. We obtain an area under the ROC curve of. 95, improving over our adaptation of a state-of-the-art Gaussian mixture model PH detector by +.17. Post-hoc explanations and analysis show that the availability of separated A2 and P2 components contributes significantly to prediction. Analysis of stethoscope heart sound recordings with deep networks is an effective, low-cost and non-invasive solution for the detection of pulmonary hypertension.","PeriodicalId":117840,"journal":{"name":"2022 Computing in Cardiology (CinC)","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115326896","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：