Multiscaled Fusion of Deep Convolutional Neural Networks for Screening Atrial Fibrillation From Single Lead Short ECG Recordings

Deep learning based ECG segmentation for delineation of diverse arrhythmias

Accurate delineation of key waveforms in an ECG is a critical step in extracting relevant features to support the diagnosis and treatment of heart conditions. Although deep learning based methods using segmentation models to locate P, QRS, and T waves have shown promising results, their ability to handle arrhythmias has not been studied in any detail. In this paper we investigate the effect of arrhythmias on delineation quality and develop strategies to improve performance in such cases. We introduce a U-Net-like segmentation model for ECG delineation with a particular focus on diverse arrhythmias. This is followed by a post-processing algorithm which removes noise and automatically determines the boundaries of P, QRS, and T waves. Our model has been trained on a diverse dataset and evaluated against the LUDB and QTDB datasets to show strong performance, with F1-scores exceeding 99% for QRS and T waves, and over 97% for P waves in the LUDB dataset. Furthermore, we assess various models across a wide array of arrhythmias and observe that models with a strong performance on standard benchmarks may still perform poorly on arrhythmias that are underrepresented in these benchmarks, such as tachycardias. We propose solutions to address this discrepancy.

Using artificial intelligence and deep learning to optimise the selection of adult congenital heart disease patients in S-ICD screening

INTRODUCTION

The risk of complications associated with transvenous ICDs make the subcutaneous implantable cardiac defibrillator (S-ICD) a valuable alternative in patients with adult congenital heart disease (ACHD). However, higher S-ICD ineligibility and higher inappropriate shock rates-mostly caused by T wave oversensing (TWO)- are observed in this population. We report a novel application of deep learning methods to screen patients for S-ICD eligibility over a longer period than conventional screening.

METHODS

Adult patients with ACHD and a control group of normal subjects were fitted with a 24-h Holters to record their S-ICD vectors. Their T:R ratio was analysed utilising phase space reconstruction matrices and a deep learning-based model to provide an in-depth description of the T: R variation plot for each vector. T: R variation was compared statistically using t-test.

RESULTS

13 patients (age 37.4 ± 7.89 years, 61.5 % male, 6 ACHD and 7 control subjects) were enrolled. A significant difference was observed in the mean and median T: R values between the two groups (p < 0.001). There was also a significant difference in the standard deviation of T: R between both groups (p = 0.04).

CONCLUSIONS

T:R ratio, a main determinant for S-ICD eligibility, is significantly higher with more tendency to fluctuate in ACHD patients when compared to a population with normal hearts. We hypothesise that our novel model could be used to select S-ICD eligible patients by better characterisation of T:R ratio, reducing the risk of TWO and inappropriate shocks in the ACHD patient cohort.

A lightweight deep learning approach for detecting electrocardiographic lead misplacement

Objective. Electrocardiographic (ECG) lead misplacement can result in distorted waveforms and amplitudes, significantly impacting accurate interpretation. Although lead misplacement is a relatively low-probability event, with an incidence ranging from 0.4% to 4%, the large number of ECG records in clinical practice necessitates the development of an effective detection method. This paper aimed to address this gap by presenting a novel lead misplacement detection method based on deep learning models.Approach. We developed two novel lightweight deep learning model for limb and chest lead misplacement detection, respectively. For limb lead misplacement detection, two limb leads and V6 were used as inputs, while for chest lead misplacement detection, six chest leads were used as inputs. Our models were trained and validated using the Chapman database, with an 8:2 train-validation split, and evaluated on the PTB-XL, PTB, and LUDB databases. Additionally, we examined the model interpretability on the LUDB databases. Limb lead misplacement simulations were performed using mathematical transformations, while chest lead misplacement scenarios were simulated by interchanging pairs of leads. The detection performance was assessed using metrics such as accuracy, precision, sensitivity, specificity, and Macro F1-score.Main results. Our experiments simulated three scenarios of limb lead misplacement and nine scenarios of chest lead misplacement. The proposed two models achieved Macro F1-scores ranging from 93.42% to 99.61% on two heterogeneous test sets, demonstrating their effectiveness in accurately detecting lead misplacement across various arrhythmias.Significance. The significance of this study lies in providing a reliable open-source algorithm for lead misplacement detection in ECG recordings. The source code is available athttps://github.com/wjcai/ECG_lead_check.

Automated identification of atrial fibrillation from single-lead ECGs using multi-branching ResNet

Introduction: Atrial fibrillation (AF) is the most common cardiac arrhythmia, which is clinically identified with irregular and rapid heartbeat rhythm. AF puts a patient at risk of forming blood clots, which can eventually lead to heart failure, stroke, or even sudden death. Electrocardiography (ECG), which involves acquiring bioelectrical signals from the body surface to reflect heart activity, is a standard procedure for detecting AF. However, the occurrence of AF is often intermittent, costing a significant amount of time and effort from medical doctors to identify AF episodes. Moreover, human error is inevitable, as even experienced medical professionals can overlook or misinterpret subtle signs of AF. As such, it is of critical importance to develop an advanced analytical model that can automatically interpret ECG signals and provide decision support for AF diagnostics. Methods: In this paper, we propose an innovative deep-learning method for automated AF identification using single-lead ECGs. We first extract time-frequency features from ECG signals using continuous wavelet transform (CWT). Second, the convolutional neural networks enhanced with residual learning (ReNet) are employed as the functional approximator to interpret the time-frequency features extracted by CWT. Third, we propose to incorporate a multi-branching structure into the ResNet to address the issue of class imbalance, where normal ECGs significantly outnumber instances of AF in ECG datasets. Results and Discussion: We evaluate the proposed Multi-branching Resnet with CWT (CWT-MB-Resnet) with two ECG datasets, i.e., PhysioNet/CinC challenge 2017 and ECGs obtained from the University of Oklahoma Health Sciences Center (OUHSC). The proposed CWT-MB-Resnet demonstrates robust prediction performance, achieving an F1 score of 0.8865 for the PhysioNet dataset and 0.7369 for the OUHSC dataset. The experimental results signify the model's superior capability in balancing precision and recall, which is a desired attribute for ensuring reliable medical diagnoses.

Pre-Processing techniques and artificial intelligence algorithms for electrocardiogram (ECG) signals analysis: A comprehensive review

Electrocardiogram (ECG) are the physiological signals and a standard test to measure the heart's electrical activity that depicts the movement of cardiac muscles. A review study has been conducted on ECG signals analysis with the help of artificial intelligence (AI) methods over the last ten years i.e., 2012-22. Primarily, the method of ECG analysis by software systems was divided into classical signal processing (e.g. spectrograms or filters), machine learning (ML) and deep learning (DL), including recursive models, transformers and hybrid. Secondly, the data sources and benchmark datasets were depicted. Authors grouped resources by ECG acquisition methods into hospital-based portable machines and wearable devices. Authors also included new trends like advanced pre-processing, data augmentation, simulations and agent-based modeling. The study found improvement in ECG examination perfection made each year through ML, DL, hybrid models, and transformers. Convolutional neural networks and hybrid models were more targeted and proved efficient. The transformer model extended the accuracy from 90% to 98%. The Physio-Net library helps acquire ECG signals, including the popular benchmark databases such as MIT-BIH, PTB, and challenging datasets. Similarly, wearable devices have been established as a appropriate option for monitoring patient health without the time and place limitations and are also helpful for AI model calibration with so far accuracy of 82%-83% on Samsung smartwatch. In the pre-processing signals, spectrogram generation through Fourier and wavelet transformations are erected leading approaches promoting on average accuracy of 90%-95%. Likewise, data enhancement using geometrical techniques is well-considered; however, extraction and concatenation-based methods need attention. As the what-if analysis in healthcare or cardiac issues can be performed using a complex simulation, the study reviews agent-based modeling and simulation approaches for cardiovascular risk event assessment.

A comprehensive review on heart disease prognostication using different artificial intelligence algorithms

Prediction of heart diseases on time is significant in order to preserve life. Many conventional methods have taken efforts on earlier prediction but faced with challenges of higher prediction cost, extended time for computation and complexities with larger volume of data which reduced prediction accuracy. In order to overcome such pitfalls, AI (Artificial Intelligence) technology has been evolved in diagnosing heart diseases through deployment of several ML (Machine Learning) and DL (Deep Learning) algorithms. It improves detection by influencing with its capacity of learning from the massive data containing age, obesity, hypertension and other risk factors of patients and extract it accordingly to differentiate on the circumstances. Moreover, storage of larger data with AI greatly assists in analysing the occurrence of the disease from past historical data. Hence, this paper intends to provide a review on different AI based algorithms used in the heart disease prognostication and delivers its benefits through researching on various existing works. It performs comparative analysis and critical assessment as encompassing accuracies and maximum utilization of algorithms focussed by traditional studies in this area. The major findings of the paper emphasized on the evolution and continuous explorations of AI techniques for heart disease prediction and the future researchers aims in determining the dimensions that have attained high and low prediction accuracies on which appropriate research works can be performed. Finally, future research is included to offer new stimulus for further investigation of AI in cardiac disease diagnosis.

Applications of Machine Learning (ML) and Mathematical Modeling (MM) in Healthcare with Special Focus on Cancer Prognosis and Anticancer Therapy: Current Status and Challenges

The use of data-driven high-throughput analytical techniques, which has given rise to computational oncology, is undisputed. The widespread use of machine learning (ML) and mathematical modeling (MM)-based techniques is widely acknowledged. These two approaches have fueled the advancement in cancer research and eventually led to the uptake of telemedicine in cancer care. For diagnostic, prognostic, and treatment purposes concerning different types of cancer research, vast databases of varied information with manifold dimensions are required, and indeed, all this information can only be managed by an automated system developed utilizing ML and MM. In addition, MM is being used to probe the relationship between the pharmacokinetics and pharmacodynamics (PK/PD interactions) of anti-cancer substances to improve cancer treatment, and also to refine the quality of existing treatment models by being incorporated at all steps of research and development related to cancer and in routine patient care. This review will serve as a consolidation of the advancement and benefits of ML and MM techniques with a special focus on the area of cancer prognosis and anticancer therapy, leading to the identification of challenges (data quantity, ethical consideration, and data privacy) which are yet to be fully addressed in current studies.

Efficient Classification of ECG Images Using a Lightweight CNN with Attention Module and IoT

Cardiac disorders are a leading cause of global casualties, emphasizing the need for the initial diagnosis and prevention of cardiovascular diseases (CVDs). Electrocardiogram (ECG) procedures are highly recommended as they provide crucial cardiology information. Telemedicine offers an opportunity to provide low-cost tools and widespread availability for CVD management. In this research, we proposed an IoT-based monitoring and detection system for cardiac patients, employing a two-stage approach. In the initial stage, we used a routing protocol that combines routing by energy and link quality (REL) with dynamic source routing (DSR) to efficiently collect data on an IoT healthcare platform. The second stage involves the classification of ECG images using hybrid-based deep features. Our classification system utilizes the "ECG Images dataset of Cardiac Patients", comprising 12-lead ECG images with four distinct categories: abnormal heartbeat, myocardial infarction (MI), previous history of MI, and normal ECG. For feature extraction, we employed a lightweight CNN, which automatically extracts relevant ECG features. These features were further optimized through an attention module, which is the method's main focus. The model achieved a remarkable accuracy of 98.39%. Our findings suggest that this system can effectively aid in the identification of cardiac disorders. The proposed approach combines IoT, deep learning, and efficient routing protocols, showcasing its potential for improving CVD diagnosis and management.

Self-Attention LSTM-FCN model for arrhythmia classification and uncertainty assessment

This paper presents ArrhyMon, a self-attention-based LSTM-FCN model for arrhythmia classification from ECG signal inputs. ArrhyMon targets to detect and classify six different types of arrhythmia apart from normal ECG patterns. To the best of our knowledge, ArrhyMon is the first end-to-end classification model that successfully targets the classification of six detailed arrhythmia types and compared to previous work does not require additional preprocessing and/or feature extraction operations separate from the classification model. ArrhyMon's deep learning model is designed to capture and exploit both global and local features embedded in ECG sequences by integrating fully convolutional network (FCN) layers and a self-attention-based long and short-term memory (LSTM) architecture. Moreover, to enhance its practicality, ArrhyMon incorporates a deep ensemble-based uncertainty model that generates a confidence-level measure for each classification result. We evaluate ArrhyMon's effectiveness using three publicly available arrhythmia datasets (i.e., MIT-BIH, Physionet Cardiology Challenge 2017 and 2020/2021) to show that ArrhyMon achieves state-of-the-art classification performance (average accuracy 99.63%), and that confidence measures show close correlation with subjective diagnosis made from practitioners.

Deep Learning-Based Non-Contact IPPG Signal Blood Pressure Measurement Research

In this paper, a multi-stage deep learning blood pressure prediction model based on imaging photoplethysmography (IPPG) signals is proposed to achieve accurate and convenient monitoring of human blood pressure. A camera-based non-contact human IPPG signal acquisition system is designed. The system can perform experimental acquisition under ambient light, effectively reducing the cost of non-contact pulse wave signal acquisition while simplifying the operation process. The first open-source dataset IPPG-BP for IPPG signal and blood pressure data is constructed by this system, and a multi-stage blood pressure estimation model combining a convolutional neural network and bidirectional gated recurrent neural network is designed. The results of the model conform to both BHS and AAMI international standards. Compared with other blood pressure estimation methods, the multi-stage model automatically extracts features through a deep learning network and combines different morphological features of diastolic and systolic waveforms, which reduces the workload while improving accuracy.

Electrocardiogram Monitoring Wearable Devices and Artificial-Intelligence-Enabled Diagnostic Capabilities: A Review

Worldwide, population aging and unhealthy lifestyles have increased the incidence of high-risk health conditions such as cardiovascular diseases, sleep apnea, and other conditions. Recently, to facilitate early identification and diagnosis, efforts have been made in the research and development of new wearable devices to make them smaller, more comfortable, more accurate, and increasingly compatible with artificial intelligence technologies. These efforts can pave the way to the longer and continuous health monitoring of different biosignals, including the real-time detection of diseases, thus providing more timely and accurate predictions of health events that can drastically improve the healthcare management of patients. Most recent reviews focus on a specific category of disease, the use of artificial intelligence in 12-lead electrocardiograms, or on wearable technology. However, we present recent advances in the use of electrocardiogram signals acquired with wearable devices or from publicly available databases and the analysis of such signals with artificial intelligence methods to detect and predict diseases. As expected, most of the available research focuses on heart diseases, sleep apnea, and other emerging areas, such as mental stress. From a methodological point of view, although traditional statistical methods and machine learning are still widely used, we observe an increasing use of more advanced deep learning methods, specifically architectures that can handle the complexity of biosignal data. These deep learning methods typically include convolutional and recurrent neural networks. Moreover, when proposing new artificial intelligence methods, we observe that the prevalent choice is to use publicly available databases rather than collecting new data.

Automatic screening of patients with atrial fibrillation from 24-h Holter recording using deep learning

Aims

As the demand for atrial fibrillation (AF) screening increases, clinicians spend a significant amount of time identifying AF signals from massive amounts of data obtained during long-term dynamic electrocardiogram (ECG) monitoring. The identification of AF signals is subjective and depends on the experience of clinicians. However, experienced cardiologists are scarce. This study aimed to apply a deep learning-based algorithm to fully automate primary screening of patients with AF using 24-h Holter monitoring.

Methods and results

A deep learning model was developed to automatically detect AF episodes using RR intervals and was trained and evaluated on 23 621 (2297 AF and 21 324 non-AF) 24-h Holter recordings from 23 452 patients. Based on the AF episode detection results, patients with AF were automatically identified using the criterion of at least one AF episode lasting 6 min or longer. Performance was assessed on an independent real-world hospital-scenario test set (19 227 recordings) and a community-scenario test set (1299 recordings). For the two test sets, the model obtained high performance for the identification of patients with AF (sensitivity: 0.995 and 1.000; specificity: 0.985 and 0.997, respectively). Moreover, it obtained good and consistent performance (sensitivity: 1.000; specificity: 0.972) for an external public data set.

Conclusion

Using the criterion of at least one AF episode of 6 min or longer, the deep learning model can fully automatically screen patients for AF with high accuracy from long-term Holter monitoring data. This method may serve as a powerful and cost-effective tool for primary screening for AF.

Convolutional neural network optimized by differential evolution for electrocardiogram classification

The Coronavirus disease 2019, or COVID-19, has shifted the medical paradigm from face-to-face to telehealth. Telehealth has become a vital resource to contain the virus spread and ensure the continued care of patients. In terms of preventing cardiovascular diseases, automating electrocardiogram (ECG) classification is a promising telehealth intervention. The healthcare service ensures that patient care is appropriate, comfortable, and accessible. Convolutional neural networks (CNNs) have demonstrated promising results in ECG categorization, which require high accuracy and short training time to ensure healthcare quality. This study proposes a one-dimensional-CNN (1D-CNN) arrhythmia classification based on the differential evolution (DE) algorithm to optimize the accuracy of ECG classification and training time. The performance of 1D-CNNs of different activation functions are optimized based on the standard DE algorithm. Finally, based on MIT-BIH and SCDH arrhythmia databases, the performances of optimized and unoptimized 1D-CNN are compared and analysed. Results show that the 1D-CNN optimized by the DE has higher accuracy in heartbeats classification. The optimized 1D-CNN improves from 97.6% to 99.5% on MIT-BIH and from 80.2% to 88.5% on SCDH. Therefore, the optimized 1D-CNN shows improvements of 1.9% and 8.3% in the two datasets, respectively. In addition, compared with the unoptimized 1D-CNN based on the same parameter settings, the optimized 1D-CNN has less training time. Under the conditions of ReLU function and 10 epochs, the training takes 9.22 s on MIT-BIH and 10.35 s on SCDH, reducing training time by 67.2% and 64.2%, respectively.

Correlation analysis of deep learning methods in S-ICD screening

BACKGROUND

Machine learning methods are used in the classification of various cardiovascular diseases through ECG data analysis. The concept of varying subcutaneous implantable cardiac defibrillator (S-ICD) eligibility, owing to the dynamicity of ECG signals, has been introduced before. There are practical limitations to acquiring longer durations of ECG signals for S-ICD screening. This study explored the potential use of deep learning methods in S-ICD screening.

METHODS

This was a retrospective study. A deep learning tool was used to provide descriptive analysis of the T:R ratios over 24 h recordings of S-ICD vectors. Spearman's rank correlation test was used to compare the results statistically to those of a "gold standard" S-ICD simulator.

RESULTS

A total of 14 patients (mean age: 63.7 ± 5.2 years, 71.4% male) were recruited and 28 vectors were analyzed. Mean T:R, standard deviation of T:R, and favorable ratio time (FVR)-a new concept introduced in this study-for all vectors combined were 0.21 ± 0.11, 0.08 ± 0.04, and 79 ± 30%, respectively. There were statistically significant strong correlations between the outcomes of our novel tool and the S-ICD simulator (p < .001).

CONCLUSION

Deep learning methods could provide a practical software solution to analyze data acquired for longer durations than current S-ICD screening practices. This could help select patients better suited for S-ICD therapy as well as guide vector selection in S-ICD eligible patients. Further work is needed before this could be translated into clinical practice.

Morphological Autoencoders for Beat-by-Beat Atrial Fibrillation Detection Using Single-Lead ECG

Engineered feature extraction can compromise the ability of Atrial Fibrillation (AFib) detection algorithms to deliver near real-time results. Autoencoders (AEs) can be used as an automatic feature extraction tool, tailoring the resulting features to a specific classification task. By coupling an encoder to a classifier, it is possible to reduce the dimension of the Electrocardiogram (ECG) heartbeat waveforms and classify them. In this work we show that morphological features extracted using a Sparse AE are sufficient to distinguish AFib from Normal Sinus Rhythm (NSR) beats. In addition to the morphological features, rhythm information was included in the model using a proposed short-term feature called Local Change of Successive Differences (LCSD). Using single-lead ECG recordings from two referenced public databases, and with features from the AE, the model was able to achieve an F1-score of 88.8%. These results show that morphological features appear to be a distinct and sufficient factor for detecting AFib in ECG recordings, especially when designed for patient-specific applications. This is an advantage over state-of-the-art algorithms that need longer acquisition times to extract engineered rhythm features, which also requires careful preprocessing steps. To the best of our knowledge, this is the first work that presents a near real-time morphological approach for AFib detection under naturalistic ECG acquisition with a mobile device.

Method for Solving Difficulties in Rhythm Classification Caused by Few Samples and Similar Characteristics in Electrocardiograms

A method for accurately analyzing electrocardiograms (ECGs), which are obtained from electrical signals generated by cardiac activity, is essential in heart disease diagnosis. However, rhythms are typically obtained with relatively few data samples and similar characteristics, making them difficult to classify. To solve these issues, we proposed a novel method that distinguishes a given ECG rhythm using a beat score map (BSM) image. Through the proposed method, the associations between beats and previously used features, such as the R-R interval, were considered. Rhythm classification was implemented by training a convolutional neural network model and using transfer learning with the created BSM image. As a result, the proposed method for ECG rhythms with small data samples showed significant results. It also showed good performance in differentiating atrial fibrillation (AFIB) and atrial flutter (AFL) rhythms, which are difficult to distinguish due to their similar characteristics. The performance for rhythms with a small number of samples of the proposed method is 20% better than an existing method. In addition, the performance based on the F-1 score for classifying AFIB and AFL of the proposed method is 30% better than the existing method. This study solved the previous limitations caused by small sample numbers and similar rhythms.

Atrial Fibrillation Detection with Low Signal-to-Noise Ratio Data Using Artificial Features and Abstract Features

Detecting atrial fibrillation (AF) of short single-lead electrocardiogram (ECG) with low signal-to-noise ratio (SNR) is a key of the wearable heart monitoring system. This study proposed an AF detection method based on feature fusion to identify AF rhythm (A) from other three categories of ECG recordings, that is, normal rhythm (N), other rhythm (O), and noisy (∼) ECG recordings. So, the four categories, that is, N, A, O, and ∼ were identified from the database provided by PhysioNet/CinC Challenge 2017. The proposed method first unified the 9 to 60 seconds unbalanced ECG recordings into 30 s segments by copying, cutting, and symmetry. Then, 24 artificial features including waveform features, interval features, frequency-domain features, and nonlinear feature were extracted relying on prior knowledge. Meanwhile, a 13-layer one-dimensional convolutional neural network (1-D CNN) was constructed to yield 38 abstract features. Finally, 24 artificial features and 38 abstract features were fused to yield the feature matrix. Random forest was employed to classify the ECG recordings. In this study, the mean accuracy (Acc) of the four categories reached 0.857. The F ₁ of N, A, and O reached 0.837. The results exhibited the proposed method had relatively satisfactory performance for identifying AF from short single-lead ECG recordings with low SNR.

Role of deep learning methods in screening for subcutaneous implantable cardioverter defibrillator in heart failure

INTRODUCTION

S-ICD eligibility is assessed at pre-implant screening where surface ECG traces are used as surrogates for S-ICD vectors. In heart failure (HF) patients undergoing diuresis, electrolytes and fluid shifts can cause changes in R and T waves. Subsequently, T:R ratio, a major predictor of S-ICD eligibility, can be dynamic.

METHODS

This is a prospective study of patients with structurally normal hearts and HF patients undergoing diuresis. All patients were fitted with Holters® to record their S-ICD vectors. Our deep learning model was used to analyze the T:R ratios across the recordings. Welch two sample t-test and Mann-Whitney U were used to compare the data between the two groups.

RESULTS

Twenty-one patients (age 58.43 ± 18.92, 62% male, 14 HF, 7 normal hearts) were enrolled. There was a significant difference in the T:R ratios between both groups. Mean T: R was higher in the HF group (0.18 ± 0.08 vs 0.10 ± 0.05, p < .001). Standard deviation of T: R was also higher in the HF group (0.09 ± 0.05 vs 0.07 ± 0.04, p = .024). There was no difference between leads within the same group.

CONCLUSIONS

T:R ratio, a main determinant for S-ICD eligibility, is higher and has more tendency to fluctuate in HF patients undergoing diuresis. We hypothesize that our novel neural network model could be used to select HF patients eligible for S-ICD by better characterization of T:R ratio reducing the risk of T-wave over-sensing (TWO) and inappropriate shocks. Further work is required to consolidate our findings before applying to clinical practice.

Changes of oscillogram envelope maximum with blood pressure and aging: a quantitative observation

Objective.The oscillometric blood pressure (BP) measurement technique estimates BPs from analyzing the oscillometric cuff pressure waveform (oscillogram) envelope. The oscillogram envelope maximum is associated with physiological changes and influences BP measurement accuracy. We aim to quantitatively investigate the effect of BP and aging on the changes of oscillogram envelope maximum.Approach.Four hundred and sixty-two subjects (214 female, 248 male) were recruited. The cuff pressure was digitally recorded during linear cuff deflation to derive oscillogram envelopes and their maximums. Moderation analysis was performed to investigate whether the relationship between BP and envelope maximum was moderated by age. Subjects were divided into five age categories and three BP groups. The envelope maximums were compared between different BP and age categories to qualify their changes with increased BP and aging.Main results.Age has a significant moderating effect on the relationship between BP and envelope maximum (P < 0.05). The oscillogram envelope maximums increased significantly with increased BPs (P < 0.05 between each BP groups) and aging (P < 0.05 for > 60 years old groups in comparison with younger groups).Significance.This study experientially and theoretically concluded the BPs and aging are two important factors that influence the maximum value of the oscillogram envelope.

Efficient Lightweight Multimodel Deep Fusion Based on ECG for Arrhythmia Classification

An arrhythmia happens when the electrical signals that organize the heartbeat do not work accurately. Most cases of arrhythmias may increase the risk of stroke or cardiac arrest. As a result, early detection of arrhythmia reduces fatality rates. This research aims to provide a lightweight multimodel based on convolutional neural networks (CNNs) that can transfer knowledge from many lightweight deep learning models and decant it into one model to aid in the diagnosis of arrhythmia by using electrocardiogram (ECG) signals. Thus, we gained a multimodel able to classify arrhythmia from ECG signals. Our system's effectiveness is examined by using a publicly accessible database and a comparison to the current methodologies for arrhythmia classification. The results we achieved by using our multimodel are better than those obtained by using a single model and better than most of the previous detection methods. It is worth mentioning that this model produced accurate classification results on small collection of data. Experts in this field can use this model as a guide to help them make decisions and save time.

Atrial fibrillation prediction by combining ECG markers and CMR radiomics

Atrial fibrillation (AF) is the most common cardiac arrhythmia. It is associated with a higher risk of important adverse health outcomes such as stroke and death. AF is linked to distinct electro-anatomic alterations. The main tool for AF diagnosis is the Electrocardiogram (ECG). However, an ECG recorded at a single time point may not detect individuals with paroxysmal AF. In this study, we developed machine learning models for discrimination of prevalent AF using a combination of image-derived radiomics phenotypes and ECG features. Thus, we characterize the phenotypes of prevalent AF in terms of ECG and imaging alterations. Moreover, we explore sex-differential remodelling by building sex-specific models. Our integrative model including radiomics and ECG together resulted in a better performance than ECG alone, particularly in women. ECG had a lower performance in women than men (AUC: 0.77 vs 0.88, p < 0.05) but adding radiomics features, the accuracy of the model was able to improve significantly. The sensitivity also increased considerably in women by adding the radiomics (0.68 vs 0.79, p < 0.05) having a higher detection of AF events. Our findings provide novel insights into AF-related electro-anatomic remodelling and its variations by sex. The integrative radiomics-ECG model also presents a potential novel approach for earlier detection of AF.

Automated diagnosis of atrial fibrillation using ECG component-aware transformer

Atrial fibrillation (AF) is the most common sustained arrhythmia worldwide and imposes a substantial economic burden on the public healthcare system due to its high morbidity and mortality. Early detection of AF is crucial in providing timely treatment and preventing complications such as stroke and other thromboembolism. For AF diagnosis, the 12-lead electrocardiogram (ECG) has been established as the gold standard. However, it requires the clinical experiences of cardiologists and may be vulnerable to inter-observer variability. Although automated AF diagnostic techniques based on deep neural networks (DNN) have been proposed, most studies were conducted using small-scale datasets, resulting in the over-fitting problem. Furthermore, they have not fully exploited ECG components such as P-wave, QRS-complex, and T-wave contrary to the approach adopted by cardiologists who interpret ECG by considering its components. To overcome these limitations, this study presents the component-aware transformer (CAT), which segments the ECG waveform into each component, vectorizes them with length and types information into one vector, and used it as the input of the transformer. We conducted extensive experiments to evaluate the CAT using a large-scale dataset called Shaoxing Hospital Zhejiang University School of Medicine database (AF: 1,780 cases, non-AF: 8,866 cases). The quantitative evaluations demonstrate that the CAT outperforms the conventional deep learning techniques on both single- and 12-lead ECG signals. Moreover, the CAT trained on single-lead ECG is comparable to that of a 12-lead analysis, while conventional methods degraded significantly in performance. Consequently, the CAT is applicable to various single-channel signals such as airway pressure, photoplethysmogram, and blood pressure.

Detection of arrhythmia in 12-lead varied-length ECG using multi-branch signal fusion network

Objective.Automatic detection of arrhythmia based on electrocardiogram (ECG) plays a critical role in early prevention and diagnosis of cardiovascular diseases. With the increase in widely available digital ECG data and the development of deep learning, multi-class arrhythmia classification based on automatic feature extraction of ECG has become increasingly attractive. However, the majority of studies cannot accept varied-length ECG signals and have limited performance in detecting multi-class arrhythmias.Approach.In this study, we propose a multi-branch signal fusion network (MBSF-Net) for multi-label classification of arrhythmia in 12-lead varied-length ECG. Our model utilizes the complementary power between different structures, which include Inception with depthwise separable convolution (DWS-Inception), spatial pyramid pooling (SPP) Layer, and multi-scale fusion Resnet (MSF-Resnet). The proposed method can extract features from each lead of 12-lead ECG recordings separately and then effectively fuse the features of each lead by integrating multiple convolution kernels with different receptive fields, which can achieve the information of complementation between different angles of the ECG signal. In particular, our model can accept 12-lead ECG signals of arbitrary length.Main results.The experimental results show that our model achieved an overall classification F1 score of 83.8% in the 12-lead ECG data of CPSC-2018. In addition, the F1 score of the MBSF-Net performed best among the MBF-Nets which are removed the SPP layer from MBSF-Net. In comparison with the latest ECG classification algorithms, the proposed model can be applied in varied-length signals and has an excellent performance, which not only can fully retain the integrity of the original signals, but also eliminates the cropping/padding signal beforehand when dealing with varied-length signal database.Significance.MBSF-Net provides an end-to-end multi-label classification model with outperfom performance, which allows detection of disease in varied-length signals without any additional cropping/padding. Moreover, our research is beneficial to the development of computer-aided diagnosis.

Novel deep learning architectures for haemodialysis time series classification

Classifying haemodialysis sessions, on the basis of the evolution of specific clinical variables over time, allows the physician to identify patients that are being treated inefficiently, and that may need additional monitoring or corrective interventions. In this paper, we propose a deep learning approach to clinical time series classification, in the haemodialysis domain. In particular, we have defined two novel architectures, able to take advantage of the strengths of Convolutional Neural Networks and of Recurrent Networks. The novel architectures we introduced and tested outperformed classical mathematical classification techniques, as well as simpler deep learning approaches. In particular, combining Recurrent Networks with convolutional structures in different ways, allowed us to obtain accuracies above 81%, coupled with high values of the Matthews Correlation Coefficient (MCC), a parameter particularly suitable to assess the quality of classification when dealing with unbalanced classes-as it was our case. In the future we will test an extension of the approach to additional monitoring time series, aiming at an overall optimization of patient care.

Multiscale Encoding of Electrocardiogram Signals with a Residual Network for the Detection of Atrial Fibrillation

Atrial fibrillation (AF) is one of the most common cardiac arrhythmias, and it is an indication of high-risk factors for stroke, myocardial ischemia, and other malignant cardiovascular diseases. Most of the existing AF detection methods typically convert one-dimensional time-series electrocardiogram (ECG) signals into two-dimensional representations to train a deep and complex AF detection system, which results in heavy training computation and high implementation costs. In this paper, a multiscale signal encoding scheme is proposed to improve feature representation and detection performance without the need for using any transformation or handcrafted feature engineering techniques. The proposed scheme uses different kernel sizes to produce the encoded signal by using multiple streams that are passed into a one-dimensional sequence of blocks of a residual convolutional neural network (ResNet) to extract representative features from the input ECG signal. This also allows networks to grow in breadth rather than in depth, thus reducing the computing time by using the parallel processing capability of deep learning networks. We investigated the effects of the use of a different number of streams with different kernel sizes on the performance. Experiments were carried out for a performance evaluation using the publicly available PhysioNet CinC Challenge 2017 dataset. The proposed multiscale encoding scheme outperformed existing deep learning-based methods with an average F₁ score of 98.54%, but with a lower network complexity.

Prediction of Wellness Condition for Community-Dwelling Elderly via ECG Signals Data-Based Feature Construction and Modeling

The accelerated growth of elderly populations in many countries and regions worldwide is creating a major burden to the healthcare system. Intelligent approaches for continuous health monitoring have the potential to promote the transition to more proactive and affordable healthcare. Electrocardiograms (ECGs), collected from portable devices, with noninvasive and cost-effective merits, have been widely used to monitor various health conditions. However, the dynamic and heterogeneous pattern of ECG signals makes relevant feature construction and predictive model development a challenging task. In this study, we aim to develop an integrated approach for one-day-forward wellness prediction in the community-dwelling elderly using single-lead short ECG signal data via multiple-features construction and predictive model implementation. Vital signs data from the elderly were collected via station-based equipment on a daily basis. After data preprocessing, a set of features were constructed from ECG signals based on the integration of various models, including time and frequency domain analysis, a wavelet transform-based model, ensemble empirical mode decomposition (EEMD), and the refined composite multiscale sample entropy (RCMSE) model. Then, a machine learning based predictive model was established to map the l-day lagged features to wellness condition. The results showed that the approach developed in this study achieved the best performance for wellness prediction in the community-dwelling elderly. In practice, the proposed approach could be useful in the timely identification of elderly people who might have health risks, and could facilitating decision-making to take appropriate interventions.

Identifying Visual Attention Features Accurately Discerning Between Autism and Typically Developing: a Deep Learning Framework

Atypical visual attention is a hallmark of autism spectrum disorder (ASD). Identifying the attention features accurately discerning between people with ASD and typically developing (TD) at the individual level remains a challenge. In this study, we developed a new systematic framework combining high accuracy deep learning classification, deep learning segmentation, image ablation and a direct measurement of classification ability to identify the discriminative features for autism identification. Our two-stream model achieved the state-of-the-art performance with a classification accuracy of 0.95. Using this framework, two new categories of features, Food & drink and Outdoor-objects, were identified as discriminative attention features, in addition to the previously reported features including Center-object and Human-faces, etc. Altered attention to the new categories helps to understand related atypical behaviors in ASD. Importantly, the area under curve (AUC) based on the combined top-9 features identified in this study was 0.92, allowing an accurate classification at the individual level. We also obtained a small but informative dataset of 12 images with an AUC of 0.86, suggesting a potentially efficient approach for the clinical diagnosis of ASD. Together, our deep learning framework based on VGG-16 provides a novel and powerful tool to recognize and understand abnormal visual attention in ASD, which will, in turn, facilitate the identification of biomarkers for ASD.

Atrioventricular Synchronization for Detection of Atrial Fibrillation and Flutter in One to Twelve ECG Leads Using a Dense Neural Network Classifier

This study investigates the use of atrioventricular (AV) synchronization as an important diagnostic criterion for atrial fibrillation and flutter (AF) using one to twelve ECG leads. Heart rate, lead-specific AV conduction time, and P-/f-wave amplitude were evaluated by three representative ECG metrics (mean value, standard deviation), namely RR-interval (RRi-mean, RRi-std), PQ-interval (PQi-mean, PQI-std), and PQ-amplitude (PQa-mean, PQa-std), in 71,545 standard 12-lead ECG records from the six largest PhysioNet CinC Challenge 2021 databases. Two rhythm classes were considered (AF, non-AF), randomly assigning records into training (70%), validation (20%), and test (10%) datasets. In a grid search of 19, 55, and 83 dense neural network (DenseNet) architectures and five independent training runs, we optimized models for one-lead, six-lead (chest or limb), and twelve-lead input features. Lead-set performance and SHapley Additive exPlanations (SHAP) input feature importance were evaluated on the test set. Optimal DenseNet architectures with the number of neurons in sequential [1st, 2nd, 3rd] hidden layers were assessed for sensitivity and specificity: DenseNet [16,16,0] with primary leads (I or II) had 87.9-88.3 and 90.5-91.5%; DenseNet [32,32,32] with six limb leads had 90.7 and 94.2%; DenseNet [32,32,4] with six chest leads had 92.1 and 93.2%; and DenseNet [128,8,8] with all 12 leads had 91.8 and 95.8%, indicating sensitivity and specificity values, respectively. Mean SHAP values on the entire test set highlighted the importance of RRi-mean (100%), RR-std (84%), and atrial synchronization (40-60%) for the PQa-mean (aVR, I), PQi-std (V2, aVF, II), and PQi-mean (aVL, aVR). Our focus on finding the strongest AV synchronization predictors of AF in 12-lead ECGs would lead to a comprehensive understanding of the decision-making process in advanced neural network classifiers. DenseNet self-learned to rely on a few ECG behavioral characteristics: first, characteristics usually associated with AF conduction such as rapid heart rate, enhanced heart rate variability, and large PQ-interval deviation in V2 and inferior leads (aVF, II); second, characteristics related to a typical P-wave pattern in sinus rhythm, which is best distinguished from AF by the earliest negative P-peak deflection of the right atrium in the lead (aVR) and late positive left atrial deflection in lateral leads (I, aVL). Our results on lead-selection and feature-selection practices for AF detection should be considered for one- to twelve-lead ECG signal processing settings, particularly those measuring heart rate, AV conduction times, and P-/f-wave amplitudes. Performances are limited to the AF diagnostic potential of these three metrics. SHAP value importance can be used in combination with a human expert's ECG interpretation to change the focus from a broad observation of 12-lead ECG morphology to focusing on the few AV synchronization findings strongly predictive of AF or non-AF arrhythmias. Our results are representative of AV synchronization findings across a broad taxonomy of cardiac arrhythmias in large 12-lead ECG databases.

A novel EEG-based major depressive disorder detection framework with two-stage feature selection

Background

Major depressive disorder (MDD) is a common mental illness, characterized by persistent depression, sadness, despair, etc., troubling people’s daily life and work seriously.

Methods

In this work, we present a novel automatic MDD detection framework based on EEG signals. First of all, we derive highly MDD-correlated features, calculating the ratio of extracted features from EEG signals at frequency bands between \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\beta$$\end{document}β and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\alpha$$\end{document}α. Then, a two-stage feature selection method named PAR is presented with the sequential combination of Pearson correlation coefficient (PCC) and recursive feature elimination (RFE), where the advantages lie in minimizing the feature searching space. Finally, we employ widely used machine learning methods of support vector machine (SVM), logistic regression (LR), and linear regression (LNR) for MDD detection with the merit of feature interpretability.

Results

Experiment results show that our proposed MDD detection framework achieves competitive results. The accuracy and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$F_{1}$$\end{document}F1 score are up to 0.9895 and 0.9846, respectively. Meanwhile, the regression determination coefficient \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$R^2$$\end{document}R2 for MDD severity assessment is up to 0.9479. Compared with existing MDD detection methods with the best accuracy of 0.9840 and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$F_1$$\end{document}F1 score of 0.97, our proposed framework achieves the state-of-the-art MDD detection performance.

Conclusions

Development of this MDD detection framework can be potentially deployed into a medical system to aid physicians to screen out MDD patients.

A spatio-temporal learning-based model for sleep apnea detection using single-lead ECG signals

Sleep apnea (SA) is a common chronic sleep breathing disorder, which would cause stroke, cognitive decline, cardiovascular disease, or even death. The SA symptoms often manifest as frequent breathing interruptions during sleep and most individuals with sleeping disorders are not aware of the SA events. Using a portable device with single-lead ECG signal is an effective way to help an individual to monitor their sleep conditions at home. However, the SA detection performance of ECG-based methods is still difficult to meet the clinical practice requirement. In this study, we propose an end-to-end spatio-temporal learning-based SA detection method, which consists of multiple spatio-temporal blocks. Each block has the identical architecture with a convolutional neural network (CNN) layer, a max-pooling layer, and a bi-gated recurrent unit (BiGRU) layer. This architecture with repeated spatio-temporal blocks can well capture the morphological spatial feature information as well as the temporal feature information from ECG signals. The proposed SA detection model was evaluated on the publicly available datasets of PhysioNet Apnea-ECG dataset (Apnea-ECG) and University College Dublin Sleep Apnea Database (UCDDB). Extensive experimental results show that our proposed SA model on both Apnea-ECG and UCDDB datasets achieves state-of-the-art results, which are obviously superior to existing ECG-based SA detection methods. It means that our proposed method has the potential to be deployed into a healthcare system to provide a sleep monitoring service, which can screen out SA population with high risk and help to take timely interventions to prevent serious consequences.

Automated sleep stage scoring employing a reasoning mechanism and evaluation of its explainability

Scoring sleep stages from biological signals is an essential but labor-intensive inspection for sleep diagnosis. The existing automated scoring methods have achieved high accuracy but are not widely applied in clinical practice. In our understanding, the existing methods have failed to establish the trust of sleep experts (e.g., physicians and clinical technologists) due to a lack of ability to explain the evidences/clues for scoring. In this study, we developed a deep-learning-based scoring model with a reasoning mechanism called class activation mapping (CAM) to solve this problem. This mechanism explicitly shows which portions of the signals support our model’s sleep stage decision, and we verified that these portions overlap with the “characteristic waves,” which are evidences/clues used in the manual scoring process. In exchange for the acquisition of explainability, employing CAM makes it difficult to follow some scoring rules. Although we concerned the negative effect of CAM on the scoring accuracy, we have found that the impact is limited. The evaluation experiment shows that the proposed model achieved a scoring accuracy of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$86.9\%$$\end{document}86.9%. It is superior to those of some existing methods and the inter-rater reliability among the sleep experts. These results suggest that Sleep-CAM achieved both explainability and required scoring accuracy for practical usage.

Non-Contact Atrial Fibrillation Detection From Face Videos by Learning Systolic Peaks

OBJECTIVE

We propose a non-contact approach for atrial fibrillation (AF) detection from face videos.

METHODS

Face videos, electrocardiography (ECG), and contact photoplethysmography (PPG) from 100 healthy subjects and 100 AF patients are recorded. Data recordings from healthy subjects are all labeled as healthy. Two cardiologists evaluated ECG recordings of patients and labeled each recording as AF, sinus rhythm (SR), or atrial flutter (AFL). We use the 3D convolutional neural network for remote PPG monitoring and propose a novel loss function (Wasserstein distance) to use the timing of systolic peaks from contact PPG as the label for our model training. Then a set of heart rate variability (HRV) features are calculated from the inter-beat intervals, and a support vector machine (SVM) classifier is trained with HRV features.

RESULTS

Our proposed method can accurately extract systolic peaks from face videos for AF detection. The proposed method is trained with subject-independent 10-fold cross-validation with 30s video clips and tested on two tasks. 1) Classification of healthy versus AF: the accuracy, sensitivity, and specificity are 96.00%, 95.36%, and 96.12%. 2) Classification of SR versus AF: the accuracy, sensitivity, and specificity are 95.23%, 98.53%, and 91.12%. In addition, we also demonstrate the feasibility of non-contact AFL detection.

CONCLUSION

We achieve good performance of non-contact AF detection by learning systolic peaks.

SIGNIFICANCE

non-contact AF detection can be used for self-screening of AF symptoms for suspectable populations at home or self-monitoring of AF recurrence after treatment for chronic patients.

MGNN: A multiscale grouped convolutional neural network for efficient atrial fibrillation detection

The reliable detection of atrial fibrillation (AF) is of great significance for monitoring disease progression and developing tailored care paths. In this work, we proposed a novel and robust method based on deep learning for the accurate detection of AF. Using RR interval sequences, a multiscale grouped convolutional neural network (MGNN) combined with self-attention was designed for automatic feature extraction, and AF and non-AF classification. An average accuracy of 97.07% was obtained in the 5-fold cross-validation. The generalization ability of the proposed MGNN was further independently tested on four other unseen datasets, and the accuracy was 92.23%, 96.86%, 94.23% and 95.91%. Moreover, comparison of the network structures indicated that the MGNN had not only better detection performance but also lower computational complexity. In conclusion, the proposed model is shown to be an efficient AF detector that has great potential for use in clinical auxiliary diagnosis and long-term home monitoring based on wearable devices.

Federated Learning-Based Detection of Invasive Carcinoma of No Special Type with Histopathological Images

Invasive carcinoma of no special type (IC-NST) is known to be one of the most prevalent kinds of breast cancer, hence the growing research interest in studying automated systems that can detect the presence of breast tumors and appropriately classify them into subtypes. Machine learning (ML) and, more specifically, deep learning (DL) techniques have been used to approach this problem. However, such techniques usually require massive amounts of data to obtain competitive results. This requirement makes their application in specific areas such as health problematic as privacy concerns regarding the release of patients’ data publicly result in a limited number of publicly available datasets for the research community. This paper proposes an approach that leverages federated learning (FL) to securely train mathematical models over multiple clients with local IC-NST images partitioned from the breast histopathology image (BHI) dataset to obtain a global model. First, we used residual neural networks for automatic feature extraction. Then, we proposed a second network consisting of Gabor kernels to extract another set of features from the IC-NST dataset. After that, we performed a late fusion of the two sets of features and passed the output through a custom classifier. Experiments were conducted for the federated learning (FL) and centralized learning (CL) scenarios, and the results were compared. Competitive results were obtained, indicating the positive prospects of adopting FL for IC-NST detection. Additionally, fusing the Gabor features with the residual neural network features resulted in the best performance in terms of accuracy, F1 score, and area under the receiver operation curve (AUC-ROC). The models show good generalization by performing well on another domain dataset, the breast cancer histopathological (BreakHis) image dataset. Our method also outperformed other methods from the literature.

CACHET-CADB: A Contextualized Ambulatory Electrocardiography Arrhythmia Dataset

ECG is a non-invasive tool for arrhythmia detection. In recent years, wearable ECG-based ambulatory arrhythmia monitoring has gained increasing attention. However, arrhythmia detection algorithms trained on existing public arrhythmia databases show higher FPR when applied to such ambulatory ECG recordings. It is primarily because the existing public databases are relatively clean as they are recorded using clinical-grade ECG devices in controlled clinical environments. They may not represent the signal quality and artifacts present in ambulatory patient-operated ECG. To help build and evaluate arrhythmia detection algorithms that can work on wearable ECG from free-living conditions, we present the design and development of the CACHET-CADB, a multi-site contextualized ECG database from free-living conditions. The CACHET-CADB is subpart of the REAFEL study, which aims at reaching the frail elderly patient to optimize the diagnosis of atrial fibrillation. In contrast to the existing databases, along with the ECG, CACHET-CADB also provides the continuous recording of patients' contextual data such as activities, body positions, movement accelerations, symptoms, stress level, and sleep quality. These contextual data can aid in improving the machine/deep learning-based automated arrhythmia detection algorithms on patient-operated wearable ECG. Currently, CACHET-CADB has 259 days of contextualized ECG recordings from 24 patients and 1,602 manually annotated 10 s heart-rhythm samples. The length of the ECG records in the CACHET-CADB varies from 24 h to 3 weeks. The patient's ambulatory context information (activities, movement acceleration, body position, etc.) is extracted for every 10 s interval cumulatively. From the analysis, nearly 11% of the ECG data in the database is found to be noisy. A software toolkit for the use of the CACHET-CADB is also provided.

Multi-Label Attribute Selection of Arrhythmia for Electrocardiogram Signals with Fusion Learning

There are three primary challenges in the automatic diagnosis of arrhythmias by electrocardiogram (ECG): the significant variation among individual patients, the multiple pathologies in the ECG signal and the high cost in annotating clinical ECG with the corresponding labels. Traditional ECG processing approaches rely heavily on prior knowledge, such as those from feature extraction and waveform analysis. The preprocessing for prior knowledge incurs computational overhead. Furthermore, standard deep learning methods do not fully consider the dynamic temporal, spatial and multi-labeling characteristics of ECG data. In clinical ECG waveforms, it is common to see multi-labeling in which a patient is labeled with multiple classes of arrhythmias. However, multiclass approaches in current research mainly solve the multi-label machine learning problem, ignoring the correlation between diseases, resulting in information loss. In this paper, an arrhythmia detection and classification scheme called multi-label fusion deep learning is proposed. The objective is to build a unified system with automatic feature learning which supports effective multi-label classification. First, a multi-label ECG-based feature selection method is combined with a matrix decomposition and sparse learning theory. The optimal feature subset is selected as a preprocessing algorithm for ECG data. A multi-label classifier is then constructed by fusing CNN and RNN networks to fully exploit the interactions and features of the time and space dimensions. The experimental result demonstrates that the proposed method can achieve a state-of-the-art performance compared to other algorithms in multi-label database experiments.

Deepaware: A hybrid deep learning and context-aware heuristics-based model for atrial fibrillation detection

BACKGROUND

State-of-the-art automatic atrial fibrillation (AF) detection models trained on RR-interval (RRI) features generally produce high performance on standard benchmark electrocardiogram (ECG) AF datasets. These models, however, result in a significantly high false positive rates (FPRs) when applied on ECG data collected under free-living ambulatory conditions and in the presence of non-AF arrhythmias.

METHOD

This paper proposes DeepAware, a novel hybrid model combining deep learning (DL) and context-aware heuristics (CAH), which reduces the FPR effectively and improves the AF detection performance on participant-operated ambulatory ECG from free-living conditions. It exploits the RRI and P-wave features, as well as the contextual features from the ambulatory ECG.

RESULTS

DeepAware is shown to be very generalizable and superior to the state-of-the-art models when applied on unseen benchmark ECG AF datasets. Most importantly, the model is able to detect AF efficiently when applied on participant-operated ambulatory ECG recordings from free-living conditions and has achieved a sensitivity (Se), specificity (Sp), and accuracy (Acc) of 97.94%, 98.39%, 98.06%, respectively. Results also demonstrate the effect of atrial activity analysis (via P-waves detection) and CAH in reducing the FPR over the RRI features-based AF detection model.

CONCLUSIONS

The proposed DeepAware model can substantially reduce the physician's workload of manually reviewing the false positives (FPs) and facilitate long-term ambulatory monitoring for early detection of AF.

Deep learning-based insights on T:R ratio behaviour during prolonged screening for S-ICD eligibility

BACKGROUND

A major predictor of eligibility of subcutaneous implantable cardiac defibrillators (S-ICD) is the T:R ratio. The eligibility cut-off of the T:R ratio incorporates a safety margin to accommodate for fluctuations of ECG signal amplitudes. We introduce a deep learning-based tool that accurately measures the degree of T:R ratio fluctuations and explore its role in S-ICD screening.

METHODS

Patients were fitted with Holters for 24 h to record their S-ICD vectors. Our tool was used to assess the T:R ratio over the duration of the recordings. Multiple T:R ratio cut-off values were applied, identifying patients at high risk of T-wave oversensing (TWO) at each of the proposed values. The purpose of our study is to identify the ratio that recognises patients at high risk of TWO while not inappropriately excluding true S-ICD candidates.

RESULTS

Thirty-seven patients (age 54.5 + / - 21.3 years, 64.8% male) were recruited. Fourteen patients had heart-failure, 7 hypertrophic cardiomyopathy, 7 had normal hearts, 6 had congenital heart disease, and 3 had prior inappropriate S-ICD shocks due to TWO. 54% of patients passed the screening at a T: R of 1:3. All patients passed the screening at a T: R of 1:1. The only subgroup to wholly pass the screening utilising all the proposed ratios are the participants with normal hearts.

CONCLUSION

We propose adopting prolonged screening to select patients eligible for S-ICD with low probability of TWO and inappropriate shocks. The appropriate T:R ratio likely lies between 1:3 and 1:1. Further studies are required to identify the optimal screening thresholds.

Semi-supervised learning for automatic atrial fibrillation detection in 24-hour Holter monitoring

Paroxysmal atrial fibrillation (AF) is generally diagnosed by long-term dynamic electrocardiogram (ECG) monitoring. Identifying AF episodes from long-term ECG data can place a heavy burden on clinicians. Many machine-learning-based automatic AF detection methods have been proposed to solve this issue. However, these methods require numerous annotated data to train the model, and the annotation of AF in long-term ECG is extremely time-consuming. Reducing the demand for labeled data can effectively improve the clinical practicability of automatic AF detection methods. In this study, we developed a novel semi-supervised learning method that generated modified low-entropy labels of unlabeled samples for training a deep learning model to automatically detect paroxysmal AF in 24 h Holter monitoring data. Our method employed a 1D CNN-LSTM neural network with RR intervals as input and used few labeled training data with numerous unlabeled data for training the neural network. This method was evaluated using a 24 h Holter monitoring dataset collected from 1000 paroxysmal AF patients. Using labeled samples from only 10 patients for model training, our method achieved a sensitivity of 97.8%, specificity of 97.9%, and accuracy of 97.9% in five-fold cross-validation. Compared to the supervised learning method with complete labeled samples, the detection accuracy of our method was only 0.5% lower, while the workload of data annotation was significantly reduced by more than 98%. In general, this is the first study to apply semi-supervised learning techniques for automatic AF detection using ECG. Our method can effectively reduce the demand for AF data annotations and can improve the clinical practicability of automatic AF detection.

Automatic Multichannel Electrocardiogram Record Classification Using XGBoost Fusion Model

There is an increasing demand for automatic classification of standard 12-lead electrocardiogram signals in the medical field. Considering that different channels and temporal segments of a feature map extracted from the 12-lead electrocardiogram record contribute differently to cardiac arrhythmia detection, and to the classification performance, we propose a 12-lead electrocardiogram signal automatic classification model based on model fusion (CBi-DF-XGBoost) to focus on representative features along both the spatial and temporal axes. The algorithm extracts local features through a convolutional neural network and then extracts temporal features through bi-directional long short-term memory. Finally, eXtreme Gradient Boosting (XGBoost) is used to fuse the 12-lead models and domain-specific features to obtain the classification results. The 5-fold cross-validation results show that in classifying nine categories of electrocardiogram signals, the macro-average accuracy of the fusion model is 0.968, the macro-average recall rate is 0.814, the macro-average precision is 0.857, the macro-average F1 score is 0.825, and the micro-average area under the curve is 0.919. Similar experiments with some common network structures and other advanced electrocardiogram classification algorithms show that the proposed model performs favourably against other counterparts in F1 score. We also conducted ablation studies to verify the effect of the complementary information from the 12 leads and the auxiliary information of domain-specific features on the classification performance of the model. We demonstrated the feasibility and effectiveness of the XGBoost-based fusion model to classify 12-lead electrocardiogram records into nine common heart rhythms. These findings may have clinical importance for the early diagnosis of arrhythmia and incite further research. In addition, the proposed multichannel feature fusion algorithm can be applied to other similar physiological signal analyses and processing.

Automatic Detection of Atrial Fibrillation from Single-Lead ECG Using Deep Learning of the Cardiac Cycle

Objective and Impact Statement. Atrial fibrillation (AF) is a serious medical condition that requires effective and timely treatment to prevent stroke. We explore deep neural networks (DNNs) for learning cardiac cycles and reliably detecting AF from single-lead electrocardiogram (ECG) signals. Introduction. Electrocardiograms are widely used for diagnosis of various cardiac dysfunctions including AF. The huge amount of collected ECGs and recent algorithmic advances to process time-series data with DNNs substantially improve the accuracy of the AF diagnosis. DNNs, however, are often designed as general purpose black-box models and lack interpretability of their decisions. Methods. We design a three-step pipeline for AF detection from ECGs. First, a recording is split into a sequence of individual heartbeats based on R-peak detection. Individual heartbeats are then encoded using a DNN that extracts interpretable features of a heartbeat by disentangling the duration of a heartbeat from its shape. Second, the sequence of heartbeat codes is passed to a DNN to combine a signal-level representation capturing heart rhythm. Third, the signal representations are passed to a DNN for detecting AF. Results. Our approach demonstrates a superior performance to existing ECG analysis methods on AF detection. Additionally, the method provides interpretations of the features extracted from heartbeats by DNNs and enables cardiologists to study ECGs in terms of the shapes of individual heartbeats and rhythm of the whole signals. Conclusion. By considering ECGs on two levels and employing DNNs for modelling of cardiac cycles, this work presents a method for reliable detection of AF from single-lead ECGs.

Multiclass Arrhythmia Detection and Classification From Photoplethysmography Signals Using a Deep Convolutional Neural Network

Background

Studies have reported the use of photoplethysmography signals to detect atrial fibrillation; however, the use of photoplethysmography signals in classifying multiclass arrhythmias has rarely been reported. Our study investigated the feasibility of using photoplethysmography signals and a deep convolutional neural network to classify multiclass arrhythmia types.

Methods and Results

ECG and photoplethysmography signals were collected simultaneously from a group of patients who underwent radiofrequency ablation for arrhythmias. A deep convolutional neural network was developed to classify multiple rhythms based on 10‐second photoplethysmography waveforms. Classification performance was evaluated by calculating the area under the microaverage receiver operating characteristic curve, overall accuracy, sensitivity, specificity, and positive and negative predictive values against annotations on the rhythm of arrhythmias provided by 2 cardiologists consulting the ECG results. A total of 228 patients were included; 118 217 pairs of 10‐second photoplethysmography and ECG waveforms were used. When validated against an independent test data set (23 384 photoplethysmography waveforms from 45 patients), the DCNN achieved an overall accuracy of 85.0% for 6 rhythm types (sinus rhythm, premature ventricular contraction, premature atrial contraction, ventricular tachycardia, supraventricular tachycardia, and atrial fibrillation); the microaverage area under the microaverage receiver operating characteristic curve was 0.978; the average sensitivity, specificity, and positive and negative predictive values were 75.8%, 96.9%, 75.2%, and 97.0%, respectively.

Conclusions

This study demonstrated the feasibility of classifying multiclass arrhythmias from photoplethysmography signals using deep learning techniques. The approach is attractive for population‐based screening and may hold promise for the long‐term surveillance and management of arrhythmia.

Registration

URL: www.chictr.org.cn. Identifier: ChiCTR2000031170.

Deep Multi-scale Fusion of Convolutional Neural Networks for EMG-based Movement Estimation

EMG-based motion estimation is required for applications such as myoelectric control, where the simultaneous estimation of kinematic information, namely joint angle and velocity, is challenging and critical. We propose a novel method for accurately modelling the generated joint angle and velocity simultaneously under isotonic, isokinetic (quasi-dynamic), and fully dynamic conditions. Our solution uses two streams of CNN, called TS-CNN to learn informative features from raw EMG data using different scales and estimate the generated motion during elbow flexion and extension. The experimental results show the robustness of our approach in comparison to conventional CNN as well as some methods used in the literature. The best obtained R2 values, are 0.81±0.06, 0.71±0.06, and 0.80±0.13 for joint angle estimation and 0.78±0.05, 0.79±0.07, and 0.71±0.13 for the velocity estimation, during isotonic, isokinetic, and dynamic contractions, respectively. Additionally, our results indicate that the experimental condition can have an impact on the model's performance for motion prediction. EMG-based velocity estimation obtains higher performance than joint angle estimation under isokinetic conditions. Under dynamic conditions, joint angle estimation is more accurate than velocity estimation, and there is no difference between joint angle and velocity estimation in the isotonic case.

A novel fetal ecg signal extraction from maternal ecg signal using conditional generative adversarial networks (CGAN)

Fetal Electrocardiogram (ECG) signal extraction from non-invasive abdominal ECG signal is one of the important clinical practices followed to observe the fetal health state. Information about heart growth and health conditions of a fetus can be observed from fetal ECG signals. However, acquiring fetal ECG from abdominal ECG signals is still considered as a challenging task in biomedical analysis. This is mainly due to corrupted high amplitude maternal ECG signals, low signal to noise ratio of fetal ECG signal, difficulties in reduction of QRS (Q wave, R wave, S wave) complexities, fetal ECG signal superimposed characteristics, other motion, and electromyography artifacts. To reduce these conventional challenges, in fetal ECG analysis of a novel Conditional Generative adversarial network (CGAN) is introduced in this research work to extract the fetal ECG signal. The proposed classification model was classified efficiently in fetal ECG signals from non-invasive abdominal ECG signals. The experimental analysis demonstrates that the proposed network model provides better results in terms of sensitivity, specificity, and accuracy compared to the conventional fetal ECG extraction models like singular value decomposition, periodic component analysis, and Adaptive neuro-fuzzy inference system.