Transformer-based heart language model with electrocardiogram annotations

This paper explores the potential of transformer-based foundation models to detect Atrial Fibrillation (AFIB) in electrocardiogram (ECG) processing, an arrhythmia specified as an irregular heart rhythm without patterns. We construct a language with tokens from heartbeat locations to detect irregular heart rhythms by applying a transformers-based neural network architecture previously used only for building natural language models. Our experiments include 41, 128, 256, and 512 tokens, representing parts of ECG recordings after tokenization. The method consists of training the foundation model with annotated benchmark databases, then finetuning on a much smaller dataset and evaluating different ECG datasets from those used in the finetuning. The best-performing model achieved an F1 score of 93.33 % to detect AFIB in an ECG segment composed of 41 heartbeats by evaluating different training and testing ECG benchmark datasets. The results showed that a foundation model trained on a large data corpus could be finetuned using a much smaller annotated dataset to detect and classify arrhythmia in ECGs. This work paves the way for the transformation of foundation models into invaluable cardiologist assistants soon, opening the possibility of training foundation models with even more data to achieve even better performance scores.

Design and validation of Withings ECG Software 2, a tiny neural network based algorithm for detection of atrial fibrillation

BACKGROUND

Atrial Fibrillation (AF) is the most common form of arrhythmia in the world with a prevalence of 1%-2%. AF is also associated with an increased risk of cardiovascular diseases (CVD), such as stroke, heart failure, and coronary artery diseases, making it a leading cause of death. Asymptomatic patients are a common case (30%-40%). This highlights the importance of early diagnosis or screening. Wearable and home devices offer new opportunities in this regard.

METHODS

We present WECG-SW2, a lightweight algorithm that classifies 30-second lead I ECG strips as 'NSR', 'AF', 'Other' or 'Noise'. By detecting the location of QRS complexes in the signal, the information can be organized into a low dimensionality input which is fed to a tiny Convolutional Neural Network (CNN) with only 3,633 parameters. This approach allows for the algorithm to run directly on the ECG acquisition devices, and improves accuracy by making the most out of the training set.

RESULTS

WECG-SW2 was evaluated on a database which combines three clinical studies sponsored by Withings with three hardware devices, as well as the MIT-BIH Arrhythmia Database. On the proprietary clinical database, the sensitivity and specificity of AF detection were 99.63% (95% CI: 99.15-99.84) and 99.85% (95% CI: 99.61-99.94), respectively, based on 4646 strips taken from 1441 participants. On the MIT-BIH Arrhythmia Database, the sensitivity and specificity were 99.87% (95% CI: 99.53, 99.98) and 100% (95% CI: 98.31, 100.0), respectively, across 2624 analyzed segments.

CONCLUSION

WECG-SW2 demonstrates high sensitivity and specificity in the detection of AF using a wide variety of ECG recording hardware. The binary of WECG-SW2 is available upon request to the corresponding author for research purposes.

Early warning of atrial fibrillation using deep learning

Atrial fibrillation (AF), the most prevalent cardiac rhythm disorder, significantly increases hospitalization and health risks. Reverting from AF to sinus rhythm (SR) often requires intensive interventions. This study presents a deep-learning model capable of predicting the transition from SR to AF on average 30.8 min before the onset appears, with an accuracy of 83% and an F1 score of 85% on the test data. This performance was obtained from R-to-R interval signals, which can be accessible from wearable technology. Our model, entitled Warning of Atrial Fibrillation (WARN), consists of a deep convolutional neural network trained and validated on 24-h Holter electrocardiogram data from 280 patients, with 70 additional patients used for testing and further evaluation on 33 patients from two external centers. The low computational cost of WARN makes it ideal for integration into wearable technology, allowing for continuous heart monitoring and early AF detection, which can potentially reduce emergency interventions and improve patient outcomes.

Continuous Atrial Fibrillation Monitoring From Photoplethysmography: Comparison Between Supervised Deep Learning and Heuristic Signal Processing

BACKGROUND

Continuous monitoring for atrial fibrillation (AF) using photoplethysmography (PPG) from smartwatches or other wearables is challenging due to periods of poor signal quality during motion or suboptimal wearing. As a result, many consumer wearables sample infrequently and only analyze when the user is at rest, which limits the ability to perform continuous monitoring or to quantify AF.

OBJECTIVES

This study aimed to compare 2 methods of continuous monitoring for AF in free-living patients: a well-validated signal processing (SP) heuristic and a convolutional deep neural network (DNN) trained on raw signal.

METHODS

We collected 4 weeks of continuous PPG and electrocardiography signals in 204 free-living patients. Both SP and DNN models were developed and validated both on holdout patients and an external validation set.

RESULTS

The results show that the SP model demonstrated receiver-operating characteristic area under the curve (AUC) of 0.972 (sensitivity 99.6%, specificity: 94.4%), which was similar to the DNN receiver-operating characteristic AUC of 0.973 (sensitivity 92.2, specificity: 95.5%); however, the DNN classified significantly more data (95% vs 62%), revealing its superior tolerance of tracings prone to motion artifact. Explainability analysis revealed that the DNN automatically suppresses motion artifacts, evaluates irregularity, and learns natural AF interbeat variability. The DNN performed better and analyzed more signal in the external validation cohort using a different population and PPG sensor (AUC, 0.994; 97% analyzed vs AUC, 0.989; 88% analyzed).

CONCLUSIONS

DNNs perform at least as well as SP models, classify more data, and thus may be better for continuous PPG monitoring.

An automated detection of atrial fibrillation from single‑lead ECG using HRV features and machine learning

BACKGROUND

Atrial fibrillation (AF) is a disorder of the heart rhythm where irregular and rapid heartbeats are observed. This supraventricular arrhythmia may increase the risk of blood clots, stroke, heart failure, and other serious heart complications. Automatic analysis of AF that is based on machine learning (ML) plays an important role in detecting this heart disease.

METHODS

A new approach for automated AF detection is presented using heart rate variability (HRV) features and machine learning. A set of time-domain, frequency-domain and nonlinear features are extracted from the R-R intervals. A new method for frequency-domain analysis of R-R intervals using the Fourier Decomposition Method is presented, which provides promising results as compared to the usual method of power spectral density estimation. We train the algorithm on the Massachusetts Institute of Technology-Beth Israel Hospital (MIT-BIH) atrial fibrillation database and perform a comprehensive analysis using statistical tests to obtain the results without any intra-patient bias.

RESULTS

The proposed method is able to achieve average result of 95.16% sensitivity, 92.46% specificity and 94.43% accuracy and its performance is better than the existing approaches. Furthermore, the efficacy of the proposed algorithm is tested on eight records from a previously unseen MIT-BIH Arrhythmia Database.

CONCLUSION

This work shows that the proposed HRV features and ML approach can be effectively used for the analysis, detection, and classification of AF.

Visual Reassessment with Flux-Interval Plot Configuration after Automatic Classification for Accurate Atrial Fibrillation Detection by Photoplethysmography

Atrial fibrillation (AFib) is a common type of arrhythmia that is often clinically asymptomatic, which increases the risk of stroke significantly but can be prevented with anticoagulation. The photoplethysmogram (PPG) has recently attracted a lot of attention as a surrogate for electrocardiography (ECG) on atrial fibrillation (AFib) detection, with its out-of-hospital usability for rapid screening or long-term monitoring. Previous studies on AFib detection via PPG signals have achieved good results, but were short of intuitive criteria like ECG p-wave absence or not, especially while using interval randomness to detect AFib suffering from conjunction with premature contractions (PAC/PVC). In this study, we newly developed a PPG flux (pulse amplitude) and interval plots-based methodology, simply comprising an irregularity index threshold of 20 and regression error threshold of 0.06 for the precise automatic detection of AFib. The proposed method with automated detection on AFib shows a combined sensitivity, specificity, accuracy, and precision of 1, 0.995, 0.995, and 0.952 across the 460 samples. Furthermore, the flux-interval plot configuration also acts as a very intuitive tool for visual reassessment to confirm the automatic detection of AFib by its distinctive plot pattern compared to other cardiac rhythms. The study demonstrated that exclusive 2 false-positive cases could be corrected after the reassessment. With the methodology’s background theory well established, the detection process automated and visualized, and the PPG sensors already extensively used, this technology is very user-friendly and convincing for promoted to in-house AFib diagnostics.

Real-time realizable mobile imaging photoplethysmography

Photoplethysmography imaging (PPGI) sensors have attracted a significant amount of attention as they enable the remote monitoring of heart rates (HRs) and thus do not require any additional devices to be worn on fingers or wrists. In this study, we mounted PPGI sensors on a robot for active and autonomous HR (R-AAH) estimation. We proposed an algorithm that provides accurate HR estimation, which can be performed in real time using vision and robot manipulation algorithms. By simplifying the extraction of facial skin images using saturation (S) values in the HSV color space, and selecting pixels based on the most frequent S value within the face image, we achieved a reliable HR assessment. The results of the proposed algorithm using the R-AAH method were evaluated by rigorous comparison with the results of existing algorithms on the UBFC-RPPG dataset (n = 42). The proposed algorithm yielded an average absolute error (AAE) of 0.71 beats per minute (bpm). The developed algorithm is simple, with a processing time of less than 1 s (275 ms for an 8-s window). The algorithm was further validated on our own dataset (BAMI-RPPG dataset [n = 14]) with an AAE of 0.82 bpm.

Simultaneous Whole-Chamber Non-contact Mapping of Highest Dominant Frequency Sites During Persistent Atrial Fibrillation: A Prospective Ablation Study

Purpose

Sites of highest dominant frequency (HDF) are implicated by many proposed mechanisms underlying persistent atrial fibrillation (persAF). We hypothesized that prospectively identifying and ablating dynamic left atrial HDF sites would favorably impact the electrophysiological substrate of persAF. We aim to assess the feasibility of prospectively identifying HDF sites by global simultaneous left atrial mapping.

Methods

PersAF patients with no prior ablation history underwent global simultaneous left atrial non-contact mapping. 30 s of electrograms recorded during AF were exported into a bespoke MATLAB interface to identify HDF regions, which were then targeted for ablation, prior to pulmonary vein isolation. Following ablation of each region, change in AF cycle length (AFCL) was documented (≥ 10 ms considered significant). Baseline isopotential maps of ablated regions were retrospectively analyzed looking for rotors and focal activation or extinction events.

Results

A total of 51 HDF regions were identified and ablated in 10 patients (median DF 5.8Hz, range 4.4-7.1Hz). An increase in AFCL of was seen in 20 of the 51 regions (39%), including AF termination in 4 patients. 5 out of 10 patients (including the 4 patients where AF termination occurred with HDF-guided ablation) were free from AF recurrence at 1 year. The proportion of HDF occurrences in an ablated region was not associated with change in AFCL (τ = 0.11, p = 0.24). Regions where AFCL decreased by 10 ms or more (i.e., AF disorganization) after ablation also showed lowest baseline spectral organization (p < 0.033 for any comparison). Considering all ablated regions, the average proportion of HDF events which were also HRI events was 8.0 ± 13%. Focal activations predominated (537/1253 events) in the ablated regions on isopotential maps, were modestly associated with the proportion of HDF occurrences represented by the ablated region (Kendall's τ = 0.40, p < 0.0001), and very strongly associated with focal extinction events (τ = 0.79, p < 0.0001). Rotors were rare (4/1253 events).

Conclusion

Targeting dynamic HDF sites is feasible and can be efficacious, but lacks specificity in identifying relevant human persAF substrate. Spectral organization may have an adjunctive role in preventing unnecessary substrate ablation. Dynamic HDF sites are not associated with observable rotational activity on isopotential mapping, but epi-endocardial breakthroughs could be contributory.

Efficient on-site confirmatory testing for atrial fibrillation with derived 12-lead ECG in a wireless body area network

Smartphones that can support and assist the screening of various cardiovascular diseases are gaining popularity in recent years. The timely detection, diagnosis, and treatment of atrial fibrillation (AF) are critical, especially for those who are at risk of stroke. AF detection via screening with wearable devices should always be confirmed by a standard 12-lead electrocardiogram (ECG). However, the inability to perform on-site AF confirmatory testing results in increased patient anxiety, followed by unnecessary diagnostic procedures and treatments. Also, the delay in confirmation procedure may conclude the condition as non-AF while it was indeed present at the time of screening. To overcome these challenges, we propose an efficient on-site confirmatory testing for AF with 12-lead ECG derived from the reduced lead set (RLS) in a wireless body area network (WBAN) environment. The reduction in the number of leads enhances the comfort level of patients as well as minimizes the hurdles associated with continuous telemonitoring applications such as data transmission, storage, and bandwidth of the overall system. The proposed method is characterized by segment-wise regression and a lead selection algorithm, facilitating improved P-wave reconstruction. Further, an efficient AF detection algorithm is proposed by incorporating a novel three-level P-wave evidence score with an RR irregularity evidence score. The proposed on-site AF confirmation test reduces false positives and false negatives by 88% and 53% respectively, compared to single lead screening. In addition, the proposed lead derivation method improves accuracy, F 1 -score, and Matthews correlation coefficient (MCC) for the on-site AF detection compared to existing related methods.

Considerations on Performance Evaluation of Atrial Fibrillation Detectors

OBJECTIVE

A large number of atrial fibrillation (AF) detectors have been published in recent years, signifying that the comparison of detector performance plays a central role, though not always consistent. The aim of this study is to shed needed light on aspects crucial to the evaluation of detection performance.

METHODS

Three types of AF detector, using either information on rhythm, rhythm and morphology, or segments of ECG samples, are implemented and studied on both real and simulated ECG signals. The properties of different performance measures are investigated, for example, in relation to dataset imbalance.

RESULTS

The results show that performance can differ considerably depending on the way detector output is compared to database annotations, i.e., beat-to-beat, segment-to-segment, or episode-to-episode comparison. Moreover, depending on the type of detector, the results substantiate that physiological and technical factors, e.g., changes in ECG morphology, rate of atrial premature beats, and noise level, can have a considerable influence on performance.

CONCLUSION

The present study demonstrates overall strengths and weaknesses of different types of detector, highlights challenges in AF detection, and proposes five recommendations on how to handle data and characterize performance.

Detection of Brief Episodes of Atrial Fibrillation Based on Electrocardiomatrix and Convolutional Neural Network

Background: Brief episodes of atrial fibrillation (AF) may evolve into longer AF episodes increasing the chances of thrombus formation, stroke, and death. Classical methods for AF detection investigate rhythm irregularity or P-wave absence in the ECG, while deep learning approaches profit from the availability of annotated ECG databases to learn discriminatory features linked to different diagnosis. However, some deep learning approaches do not provide analysis of the features used for classification. This paper introduces a convolutional neural network (CNN) approach for automatic detection of brief AF episodes based on electrocardiomatrix-images (ECM-images) aiming to link deep learning to features with clinical meaning. Materials and Methods: The CNN is trained using two databases: the Long-Term Atrial Fibrillation and the MIT-BIH Normal Sinus Rhythm, and tested on three databases: the MIT-BIH Atrial Fibrillation, the MIT-BIH Arrhythmia, and the Monzino-AF. Detection of AF is done using a sliding window of 10 beats plus 3 s. Performance is quantified using both standard classification metrics and the EC57 standard for arrhythmia detection. Layer-wise relevance propagation analysis was applied to link the decisions made by the CNN to clinical characteristics in the ECG. Results: For all three testing databases, episode sensitivity was greater than 80.22, 89.66, and 97.45% for AF episodes shorter than 15, 30 s, and for all episodes, respectively. Conclusions: Rhythm and morphological characteristics of the electrocardiogram can be learned by a CNN from ECM-images for the detection of brief episodes of AF.

Poincaré Plot Image and Rhythm-Specific Atlas for Atrial Bigeminy and Atrial Fibrillation Detection

A detector based only on RR intervals capable of classifying other tachyarrhythmias in addition to atrial fibrillation (AF) could improve cardiac monitoring. In this paper a new classification method based in a 2D non-linear RRI dynamics representation is presented. For this aim, the concepts of Poincaré Images and Atlases are introduced. Three cardiac rhythms were targeted: Normal sinus rhythm (NSR), AF and atrial bigeminy (AB). Three Physionet open source databases were used. Poincaré Images were generated for all signals using different Poincaré plot configurations: RR, dRR and RRdRR. The study was computed for different time window lengths and bin sizes. For each rhythm, the Poincaré Images of the 80% of that rhythm's patients were used to create a reference image, a Poincaré Atlas. The remaining 20% were used as test set and classified into one of the three rhythms using normalized mutual information and 2D correlation. The process was iterated in a tenfold cross-validation and patient-wise dataset division. Sensitivity results obtained for RRdRR configuration and bin size 40 ms, for a 60 s time window were 94.35% ±3.68, 82.07% ±9.18 and 88.86% ±12.79 with a specificity of 85.52% ±7.46, 95.91% ±3.14, 96.10% ±2.25 for AF, NSR and AB respectively. Results suggest that a rhythms general RRI pattern may be captured using Poincaré Atlases and that these can be used to classify other signal segments using Poincaré Images. In contrast with other studies, the former method could be generalized to more cardiac rhythms and does not depend on rhythm-specific thresholds.

Variable step dynamic threshold local binary pattern for classification of atrial fibrillation

OBJECTIVE

In this paper, we proposed new methods for feature extraction in machine learning-based classification of atrial fibrillation from ECG signal.

METHODS

Our proposed methods improved conventional 1-dimensional local binary pattern method in two ways. First, we proposed a dynamic threshold LBP code generation method for use with 1-dimensional signals, enabling the generated LBP codes to have a more detailed representation of the signal morphological pattern. Second, we introduced a variable step value into the LBP code generation algorithm to better cope with a high sampling frequency input signal without a downsampling process. The proposed methods do not employ computationally expensive processes such as filtering, wavelet transform, up/downsampling, or beat detection, and can be implemented using only simple addition, division, and compare operations.

RESULTS

Combining these two approaches, our proposed variable step dynamic threshold local binary pattern method achieved 99.11% sensitivity and 99.29% specificity when used as a feature generation algorithm in support vector machine classification of atrial fibrillation from MIT-BIH Atrial Fibrillation Database dataset. When applied on signals from MIT-BIH Arrhythmia Database, our proposed method achieved similarly good 99.38% sensitivity and 98.97% specificity.

CONCLUSION

Our proposed methods achieved one of the best results among published works in atrial fibrillation classification using the same dataset while using less computationally expensive calculations, without significant performance degradation when applied on signals from multiple databases with different sampling frequencies.

Emerging Technologies for Identifying Atrial Fibrillation

Atrial fibrillation (AF) is a major cause of morbidity and mortality globally, and much of this is driven by challenges in its timely diagnosis and treatment. Existing and emerging mobile technologies have been used to successfully identify AF in a variety of clinical and community settings, and while these technologies offer great promise for revolutionizing AF detection and screening, several major barriers may impede their effectiveness. The unclear clinical significance of device-detected AF, potential challenges in integrating patient-generated data into existing healthcare systems and clinical workflows, harm resulting from potential false positives, and identifying the appropriate scope of population-based screening efforts are all potential concerns that warrant further investigation. It is crucial for stakeholders such as healthcare providers, researchers, funding agencies, insurers, and engineers to actively work together in fulfilling the tremendous potential of mobile technologies to improve AF identification and management on a population level.

Detection of Atrial Fibrillation from Single Lead ECG Signal Using Multirate Cosine Filter Bank and Deep Neural Network

Atrial fibrillation (AF) is a cardiac arrhythmia which is characterized based on the irregsular beating of atria, resulting in, the abnormal atrial patterns that are observed in the electrocardiogram (ECG) signal. The early detection of this pathology is very helpful for minimizing the chances of stroke, other heart-related disorders, and coronary artery diseases. This paper proposes a novel method for the detection of AF pathology based on the analysis of the ECG signal. The method adopts a multi-rate cosine filter bank architecture for the evaluation of coefficients from the ECG signal at different subbands, in turn, the Fractional norm (FN) feature is evaluated from the extracted coefficients at each subband. Then, the AF detection is carried out using a deep learning approach known as the Hierarchical Extreme Learning Machine (H-ELM) from the FN features. The proposed method is evaluated by considering normal and AF pathological ECG signals from public databases. The experimental results reveal that the proposed multi-rate cosine filter bank based on FN features is effective for the detection of AF pathology with an accuracy, sensitivity and specificity values of 99.40%, 98.77%, and 100%, respectively. The performance of the proposed diagnostic features of the ECG signal is compared with other existing features for the detection of AF. The low-frequency subband FN features found to be more significant with a difference of the mean values as 0.69 between normal and AF classes.

A Review of Atrial Fibrillation Detection Methods as a Service

Atrial Fibrillation (AF) is a common heart arrhythmia that often goes undetected, and even if it is detected, managing the condition may be challenging. In this paper, we review how the RR interval and Electrocardiogram (ECG) signals, incorporated into a monitoring system, can be useful to track AF events. Were such an automated system to be implemented, it could be used to help manage AF and thereby reduce patient morbidity and mortality. The main impetus behind the idea of developing a service is that a greater data volume analyzed can lead to better patient outcomes. Based on the literature review, which we present herein, we introduce the methods that can be used to detect AF efficiently and automatically via the RR interval and ECG signals. A cardiovascular disease monitoring service that incorporates one or multiple of these detection methods could extend event observation to all times, and could therefore become useful to establish any AF occurrence. The development of an automated and efficient method that monitors AF in real time would likely become a key component for meeting public health goals regarding the reduction of fatalities caused by the disease. Yet, at present, significant technological and regulatory obstacles remain, which prevent the development of any proposed system. Establishment of the scientific foundation for monitoring is important to provide effective service to patients and healthcare professionals.

A Wearable Electrocardiogram Telemonitoring System for Atrial Fibrillation Detection

In this paper we proposed a wearable electrocardiogram (ECG) telemonitoring system for atrial fibrillation (AF) detection based on a smartphone and cloud computing. A wearable ECG patch was designed to collect ECG signals and send the signals to an Android smartphone via Bluetooth. An Android APP was developed to display the ECG waveforms in real time and transmit every 30 s ECG data to a remote cloud server. A machine learning (CatBoost)-based ECG classification method was proposed to detect AF in the cloud server. In case of detected AF, the cloud server pushed the ECG data and classification results to the web browser of a doctor. Finally, the Android APP displayed the doctor's diagnosis for the ECG signals. Experimental results showed the proposed CatBoost classifier trained with 17 selected features achieved an overall F1 score of 0.92 on the test set (n = 7,270). The proposed wearable ECG monitoring system may potentially be useful for long-term ECG telemonitoring for AF detection.

Photoplethysmography based atrial fibrillation detection: a review

Atrial fibrillation (AF) is a cardiac rhythm disorder associated with increased morbidity and mortality. It is the leading risk factor for cardioembolic stroke and its early detection is crucial in both primary and secondary stroke prevention. Continuous monitoring of cardiac rhythm is today possible thanks to consumer-grade wearable devices, enabling transformative diagnostic and patient management tools. Such monitoring is possible using low-cost easy-to-implement optical sensors that today equip the majority of wearables. These sensors record blood volume variations-a technology known as photoplethysmography (PPG)-from which the heart rate and other physiological parameters can be extracted to inform about user activity, fitness, sleep, and health. Recently, new wearable devices were introduced as being capable of AF detection, evidenced by large prospective trials in some cases. Such devices would allow for early screening of AF and initiation of therapy to prevent stroke. This review is a summary of a body of work on AF detection using PPG. A thorough account of the signal processing, machine learning, and deep learning approaches used in these studies is presented, followed by a discussion of their limitations and challenges towards clinical applications.

Standardised Framework for Quantitative Analysis of Fibrillation Dynamics

The analysis of complex mechanisms underlying ventricular fibrillation (VF) and atrial fibrillation (AF) requires sophisticated tools for studying spatio-temporal action potential (AP) propagation dynamics. However, fibrillation analysis tools are often custom-made or proprietary, and vary between research groups. With no optimal standardised framework for analysis, results from different studies have led to disparate findings. Given the technical gap, here we present a comprehensive framework and set of principles for quantifying properties of wavefront dynamics in phase-processed data recorded during myocardial fibrillation with potentiometric dyes. Phase transformation of the fibrillatory data is particularly useful for identifying self-perpetuating spiral waves or rotational drivers (RDs) rotating around a phase singularity (PS). RDs have been implicated in sustaining fibrillation, and thus accurate localisation and quantification of RDs is crucial for understanding specific fibrillatory mechanisms. In this work, we assess how variation of analysis parameters and thresholds in the tracking of PSs and quantification of RDs could result in different interpretations of the underlying fibrillation mechanism. These techniques have been described and applied to experimental AF and VF data, and AF simulations, and examples are provided from each of these data sets to demonstrate the range of fibrillatory behaviours and adaptability of these tools. The presented methodologies are available as an open source software and offer an off-the-shelf research toolkit for quantifying and analysing fibrillatory mechanisms.

ECGNET: Learning where to attend for detection of atrial fibrillation with deep visual attention

The complexity of the patterns associated with atrial fibrillation (AF) and the high level of noise affecting these patterns have significantly limited the application of current signal processing and shallow machine learning approaches to accurately detect this condition. Deep neural networks have shown to be very powerful to learn the non-linear patterns in various problems such as computer vision tasks. While deep learning approaches have been utilized to learn complex patterns related to the presence of AF in electrocardiogram (ECG) signals, they can considerably benefit from knowing which parts of the signal is more important to focus on during learning. In this paper, we introduce a two-channel deep neural network to more accurately detect the presence of AF in the ECG signals. The first channel takes in an ECG signal and automatically learns where to attend for detection of AF. The second channel simultaneously takes in the same ECG signal to consider all features of the entire signal. Besides improving detection accuracy, this model can guide the physicians via visualization that what parts of the given ECG signal are important to attend while trying to detect atrial fibrillation. The experimental results confirm that the proposed model significantly improves the performance of AF detection on well-known MIT-BIH AF database with 5-s ECG segments (achieved a sensitivity of 99.53%, specificity of 99.26% and accuracy of 99.40%).

Noise Detection in Electrocardiogram Signals for Intensive Care Unit Patients

Long term electrocardiogram (ECG) signals recorded in an intensive care unit (ICU) are often corrupted by severe motion and noise artifacts (MNA), which may lead to many false alarms including inaccurate detection of atrial fibrillation (AF). We developed an automated method to detect MNA from ECG recordings in the Medical Information Mart for Intensive Care (MIMIC) III database. Since AF detection is often based on identification of irregular RR intervals derived from the QRS complexes, the main design focus of our MNA detection algorithm was to identify corrupted QRS complexes of the ECG signals. The MNA in the MIMIC III database contain not only motion-induced noise, but also a plethora of non-ECG waveforms, which must also be automatically identified. Our algorithm is designed to first discriminate between ECG and non-ECG waveforms using both time and spectral-domain properties. For the segments of data containing ECG waveforms, a time-frequency spectrum and its sub-band decomposition approach were used to identify MNA, and high frequency noise ECG segments, respectively. The algorithm was tested on data from 35 subjects in normal sinus rhythm and 25 AF subjects. The proposed method is shown to accurately discriminate between segments that contained real ECG waveforms and those that did not, even though the latter were numerous in some subjects. In addition, we found a significant reduction (> 94%) in false positive detection of AF in normal subjects when our MNA detection algorithm was used. Without using it, we inaccurately detected AF owing to the MNA.

Accuracy and Usability of a Novel Algorithm for Detection of Irregular Pulse Using a Smartwatch Among Older Adults: Observational Study

BACKGROUND

Atrial fibrillation (AF) is often paroxysmal and minimally symptomatic, hindering its diagnosis. Smartwatches may enhance AF care by facilitating long-term, noninvasive monitoring.

OBJECTIVE

This study aimed to examine the accuracy and usability of arrhythmia discrimination using a smartwatch.

METHODS

A total of 40 adults presenting to a cardiology clinic wore a smartwatch and Holter monitor and performed scripted movements to simulate activities of daily living (ADLs). Participants' clinical and sociodemographic characteristics were abstracted from medical records. Participants completed a questionnaire assessing different domains of the device's usability. Pulse recordings were analyzed blindly using a real-time realizable algorithm and compared with gold-standard Holter monitoring.

RESULTS

The average age of participants was 71 (SD 8) years; most participants had AF risk factors and 23% (9/39) were in AF. About half of the participants owned smartphones, but none owned smartwatches. Participants wore the smartwatch for 42 (SD 14) min while generating motion noise to simulate ADLs. The algorithm determined 53 of the 314 30-second noise-free pulse segments as consistent with AF. Compared with the gold standard, the algorithm demonstrated excellent sensitivity (98.2%), specificity (98.1%), and accuracy (98.1%) for identifying irregular pulse. Two-thirds of participants considered the smartwatch highly usable. Younger age and prior cardioversion were associated with greater overall comfort and comfort with data privacy with using a smartwatch for rhythm monitoring, respectively.

CONCLUSIONS

A real-time realizable algorithm analyzing smartwatch pulse recordings demonstrated high accuracy for identifying pulse irregularities among older participants. Despite advanced age, lack of smartwatch familiarity, and high burden of comorbidities, participants found the smartwatch to be highly acceptable.

Automatic Prediction of Atrial Fibrillation Based on Convolutional Neural Network Using a Short-term Normal Electrocardiogram Signal

BACKGROUND

In this study, we propose a method for automatically predicting atrial fibrillation (AF) based on convolutional neural network (CNN) using a short-term normal electrocardiogram (ECG) signal.

METHODS

We designed a CNN model and optimized it by dropout and normalization. One-dimensional convolution, max-pooling, and fully-connected multiple perceptron were used to analyze the short-term normal ECG. The ECG signal was preprocessed and segmented to train and evaluate the proposed CNN model. The training and test sets consisted of the two AF and one normal dataset from the MIT-BIH database.

RESULTS

The proposed CNN model for the automatic prediction of AF achieved a high performance with a sensitivity of 98.6%, a specificity of 98.7%, and an accuracy of 98.7%.

CONCLUSION

The results show the possibility of automatically predicting AF based on the CNN model using a short-term normal ECG signal. The proposed CNN model for the automatic prediction of AF can be a helpful tool for the early diagnosis of AF in healthcare fields.

Atrial Fibrillation Detection by the Combination of Recurrence Complex Network and Convolution Neural Network

In this paper, R wave peak interval independent atrial fibrillation detection algorithm is proposed based on the analysis of the synchronization feature of the electrocardiogram signal by a deep neural network. Firstly, the synchronization feature of each heartbeat of the electrocardiogram signal is constructed by a Recurrence Complex Network. Then, a convolution neural network is used to detect atrial fibrillation by analyzing the eigenvalues of the Recurrence Complex Network. Finally, a voting algorithm is developed to improve the performance of the beat-wise atrial fibrillation detection. The MIT-BIH atrial fibrillation database is used to evaluate the performance of the proposed method. Experimental results show that the sensitivity, specificity, and accuracy of the algorithm can achieve 94.28%, 94.91%, and 94.59%, respectively. Remarkably, the proposed method was more effective than the traditional algorithms to the problem of individual variation in the atrial fibrillation detection.

Electrocardiogram modeling during paroxysmal atrial fibrillation: application to the detection of brief episodes

OBJECTIVE

A model for simulating multi-lead ECG signals during paroxysmal atrial fibrillation (AF) is proposed.

SIGNIFICANCE

The model is of particular significance when evaluating detection performance in the presence of brief AF episodes, especially since annotated databases with such episodes are lacking.

APPROACH

The proposed model accounts for important characteristics such as switching between sinus rhythm and AF, varying P-wave morphology, repetition rate of f-waves, presence of atrial premature beats, and various types of noise.

MAIN RESULTS

Two expert cardiologists assessed the realism of simulated signals relative to real ECG signals, both in sinus rhythm and AF. The cardiologists identified the correct rhythm in all cases, and considered two-thirds of the simulated signals as realistic. The proposed model was also investigated by evaluating the performance of two AF detectors which explored either rhythm only or both rhythm and morphology. The results show that detection performance is strongly dependent on AF episode duration, and, consequently, demonstrate that the model can play a significant role in the investigation of detector properties.

Heart rhythm characterization through induced physiological variables

Atrial fibrillation remains a major cause of morbi-mortality, making mass screening desirable and leading industry to actively develop devices devoted to automatic AF detection. Because there is a tendency toward mobile devices, there is a need for an accurate, rapid method for studying short inter-beat interval time series for real-time automatic medical monitoring. We report a new methodology to efficiently select highly discriminative variables between physiological states, here a normal sinus rhythm or atrial fibrillation. We generate induced variables using the first ten time derivatives of an RR interval time series and formally express a new multivariate metric quantifying their discriminative power to drive state variable selection. When combined with a simple classifier, this new methodology results in 99.9% classification accuracy for 1-min RR interval time series (n = 7,400), with heart rate accelerations and jerks being the most discriminant variables. We show that the RR interval time series can be drastically reduced from 60 s to 3 s, with a classification accuracy of 95.0%. We show that heart rhythm characterization is facilitated by induced variables using time derivatives, which is a generic methodology that is particularly suitable to real-time medical monitoring.

Screening strategies for atrial fibrillation: a systematic review and cost-effectiveness analysis

Background

Atrial fibrillation (AF) is a common cardiac arrhythmia that increases the risk of thromboembolic events. Anticoagulation therapy to prevent AF-related stroke has been shown to be cost-effective. A national screening programme for AF may prevent AF-related events, but would involve a substantial investment of NHS resources.

Objectives

To conduct a systematic review of the diagnostic test accuracy (DTA) of screening tests for AF, update a systematic review of comparative studies evaluating screening strategies for AF, develop an economic model to compare the cost-effectiveness of different screening strategies and review observational studies of AF screening to provide inputs to the model.

Design

Systematic review, meta-analysis and cost-effectiveness analysis.

Setting

Primary care.

Participants

Adults.

Intervention

Screening strategies, defined by screening test, age at initial and final screens, screening interval and format of screening {systematic opportunistic screening [individuals offered screening if they consult with their general practitioner (GP)] or systematic population screening (when all eligible individuals are invited to screening)}.

Main outcome measures

Sensitivity, specificity and diagnostic odds ratios; the odds ratio of detecting new AF cases compared with no screening; and the mean incremental net benefit compared with no screening.

Review methods

Two reviewers screened the search results, extracted data and assessed the risk of bias. A DTA meta-analysis was perfomed, and a decision tree and Markov model was used to evaluate the cost-effectiveness of the screening strategies.

Results

Diagnostic test accuracy depended on the screening test and how it was interpreted. In general, the screening tests identified in our review had high sensitivity (> 0.9). Systematic population and systematic opportunistic screening strategies were found to be similarly effective, with an estimated 170 individuals needed to be screened to detect one additional AF case compared with no screening. Systematic opportunistic screening was more likely to be cost-effective than systematic population screening, as long as the uptake of opportunistic screening observed in randomised controlled trials translates to practice. Modified blood pressure monitors, photoplethysmography or nurse pulse palpation were more likely to be cost-effective than other screening tests. A screening strategy with an initial screening age of 65 years and repeated screens every 5 years until age 80 years was likely to be cost-effective, provided that compliance with treatment does not decline with increasing age.

Conclusions

A national screening programme for AF is likely to represent a cost-effective use of resources. Systematic opportunistic screening is more likely to be cost-effective than systematic population screening. Nurse pulse palpation or modified blood pressure monitors would be appropriate screening tests, with confirmation by diagnostic 12-lead electrocardiography interpreted by a trained GP, with referral to a specialist in the case of an unclear diagnosis. Implementation strategies to operationalise uptake of systematic opportunistic screening in primary care should accompany any screening recommendations.

Limitations

Many inputs for the economic model relied on a single trial [the Screening for Atrial Fibrillation in the Elderly (SAFE) study] and DTA results were based on a few studies at high risk of bias/of low applicability.

Future work

Comparative studies measuring long-term outcomes of screening strategies and DTA studies for new, emerging technologies and to replicate the results for photoplethysmography and GP interpretation of 12-lead electrocardiography in a screening population.

Study registration

This study is registered as PROSPERO CRD42014013739.

Funding

The National Institute for Health Research Health Technology Assessment programme.

Dynamic ECG features for atrial fibrillation recognition

BACKGROUND

Atrial fibrillation (AF) can cause the formation of blood clots in the heart. The clots may move to the brain and cause a stroke. Therefore, this study analyzed the ECG features of AF and normal sinus rhythm signals for AF recognition which were extracted by using a second-order dynamic system (SODS) concept.

OBJECTIVE

To find the appropriate windowing length for feature extraction based on SODS and to determine a machine learning method that could provide higher accuracy in recognizing AF.

METHOD

ECG features were extracted based on a dynamic system (DS) that uses a second-order differential equation to describe the short-term behavior of ECG signals according to the natural frequency (ω), damping coefficient, (ξ), and forcing input (u). The extracted features were windowed into 2, 3, 4, 6, 8, and 10 second episodes to find the appropriate windowing size for AF signal processing. ANOVA and t-tests were used to determine the significant features. In addition, pattern recognition machine learning methods (an artificial neural network (ANN) and a support vector machine (SVM)) with k-fold cross validation (k-CV) were used to develop the ECG recognition system.

RESULTS

Significant differences (p < 0.0001) were observed among all ECG groups (NSR, N, AF) using 2, 3, 4 and 6 second episodes for the features ω and u/ω; 4, 6 and 8 second episodes for features ω and u; 4 and 6 second episodes for features ω, u and u/ω, and; 10 second episodes for the feature ξ. The highest accuracy for AF recognition (AF, NSR) using ANN with k-CV was 95.3% using combination of features (ω and u; ω, u and u/ω) and SVM with k-CV was 95.0% using a combination of features ω, u and u/ω.

CONCLUSION

This study found that 4 s is the most appropriate windowing length, using two features (ω and u) for AF detection with an accuracy of 95.3%. Moreover, the pattern recognition learning machine uses an ANN with 10-fold cross validation based on DS.

Application of the relative wavelet energy to heart rate independent detection of atrial fibrillation

BACKGROUND AND OBJECTIVES

Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia and a growing healthcare burden worldwide. It is often asymptomatic and may appear as episodes of very short duration; hence, the development of methods for its automatic detection is a challenging requirement to achieve early diagnosis and treatment strategies. The present work introduces a novel method exploiting the relative wavelet energy (RWE) to automatically detect AF episodes of a wide variety in length.

METHODS

The proposed method analyzes the atrial activity of the surface electrocardiogram (ECG), i.e., the TQ interval, thus being independent on the ventricular activity. To improve its performance under noisy recordings, signal averaging techniques were applied. The method's performance has been tested with synthesized recordings under different AF variable conditions, such as the heart rate, its variability, the atrial activity amplitude or the presence of noise. Next, the method was tested with real ECG recordings.

RESULTS

Results proved that the RWE provided a robust automatic detection of AF under wide ranges of heart rates, atrial activity amplitudes as well as noisy recordings. Moreover, the method's detection delay proved to be shorter than most of previous works. A trade-off between detection delay and noise robustness was reached by averaging 15 TQ intervals. Under these conditions, AF was detected in less than 7 beats, with an accuracy higher than 90%, which is comparable to previous works.

CONCLUSIONS

Unlike most of previous works, which were mainly based on quantifying the irregular ventricular response during AF, the proposed metric presents two major advantages. First, it can perform successfully even under heart rates with no variability. Second, it consists of a single metric, thus turning its clinical interpretation and real-time implementation easier than previous methods requiring combined indices under complex classifiers.

A Real-Time Atrial Fibrillation Detection Algorithm Based on the Instantaneous State of Heart Rate

Atrial fibrillation (AF), the most frequent cause of cardioembolic stroke, is increasing in prevalence as the population ages, and presents with a broad spectrum of symptoms and severity. The early identification of AF is an essential part for preventing the possibility of blood clotting and stroke. In this work, a real-time algorithm is proposed for accurately screening AF episodes in electrocardiograms. This method adopts heart rate sequence, and it involves the application of symbolic dynamics and Shannon entropy. Using novel recursive algorithms, a low-computational complexity can be obtained. Four publicly-accessible sets of clinical data (Long-Term AF, MIT-BIH AF, MIT-BIH Arrhythmia, and MIT-BIH Normal Sinus Rhythm Databases) were used for assessment. The first database was selected as a training set; the receiver operating characteristic (ROC) curve was performed, and the best performance was achieved at the threshold of 0.639: the sensitivity (Se), specificity (Sp), positive predictive value (PPV) and overall accuracy (ACC) were 96.14%, 95.73%, 97.03% and 95.97%, respectively. The other three databases were used for independent testing. Using the obtained decision-making threshold (i.e., 0.639), for the second set, the obtained parameters were 97.37%, 98.44%, 97.89% and 97.99%, respectively; for the third database, these parameters were 97.83%, 87.41%, 47.67% and 88.51%, respectively; the Sp was 99.68% for the fourth set. The latest methods were also employed for comparison. Collectively, results presented in this study indicate that the combination of symbolic dynamics and Shannon entropy yields a potent AF detector, and suggest this method could be of practical use in both clinical and out-of-clinical settings.

Study of atrial activities for abnormality detection by phase rectified signal averaging technique

Non-invasive detection of Atrial Fibrillation (AF) and Atrial Flutter (AFL) from ECG at the time of their onset can prevent forthcoming dangers for patients. In most of the previous detection algorithms, one of the steps includes filtering of the signal to remove noise and artefacts present in the signal. In this paper, a method of AF and AFL detection is proposed from ECG without the conventional filtering stage. Here Phase Rectified Signal Average (PRSA) technique is used with a novel optimized windowing method to achieve an averaged signal without quasi-periodicities. Both time domain and statistical features are extracted from a novel SQ concatenated section of the signal for non-linear Support Vector Machine (SVM) based classification. The performance of the proposed algorithm is tested with the MIT-BIH Arrhythmia database and good performance parameters are obtained, as indicated in the result section.

Arrhythmia discrimination using a smart phone

We hypothesize that our smartphone-based arrhythmia discrimination algorithm with data acquisition approach reliably differentiates between normal sinus rhythm (NSR), atrial fibrillation (AF), premature ventricular contractions (PVCs) and premature atrial contraction (PACs) in a diverse group of patients having these common arrhythmias. We combine root mean square of successive RR differences and Shannon entropy with Poincare plot (or turning point ratio method) and pulse rise and fall times to increase the sensitivity of AF discrimination and add new capabilities of PVC and PAC identification. To investigate the capability of the smartphone-based algorithm for arrhythmia discrimination, 99 subjects, including 88 study participants with AF at baseline and in NSR after electrical cardioversion, as well as seven participants with PACs and four with PVCs were recruited. Using a smartphone, we collected 2-min pulsatile time series from each recruited subject. This clinical application results show that the proposed method detects NSR with specificity of 0.9886, and discriminates PVCs and PACs from AF with sensitivities of 0.9684 and 0.9783, respectively.