A Singular-value-based Map to Highlight Abnormal Regions Associated with Atrial Fibrillation Using High-resolution Electrograms and Multi-lead ECG.

OBJECTIVE: The severity of atrial fibrillation (AF) can be assessed from intra-operative epicardial measurements (high-resolution electrograms), using metrics such as conduction block (CB) and continuous conduction delay and block (cCDCB). These features capture differences in conduction velocity and wavefront propagation, but ignore complementary properties such as the morphology of the action potentials. In this work, we focus on such complementary properties, and derive features to detect variations in the atrial potential waveforms. METHODS: We show that the spatial variation of atrial potential morphology during a single beat may be described by changes in the singular values of the epicardial measurement matrix. The method is non-parametric and requires little preprocessing. A corresponding singular value map points at areas subject to fractionation and block. Further, we developed an experiment where we simultaneously measure electrograms (EGMs) and a multi-lead ECG. RESULTS: The captured data showed that the normalized singular values of the heartbeats during AF are higher than during SR, and that this difference is more pronounced for the (non-invasive) ECG data than for the EGM data, if the electrodes are positioned at favorable locations. CONCLUSION: Overall, the singular value-based features are a useful indicator to detect and evaluate AF. SIGNIFICANCE: The proposed method might be beneficial for identifying electropathological regions in the tissue without estimating the local activation time.

High Dominant Frequencies and Fractionated Potentials Do Not Indicate Focal or Rotational Activation During AF.

BACKGROUND: Dominant frequencies (DFs) or complex fractionated atrial electrograms (CFAEs), indicative of focal sources or rotational activation, are used to identify target sites for atrial fibrillation (AF) ablation in clinical studies, although the relationship among DF, CFAE, and activation patterns remains unclear. OBJECTIVES: This study sought to investigate the relationship between patterns of activation underlying DF and CFAE sites during AF. METHODS: Epicardial high-resolution mapping of the right and left atrium including Bachmann's bundle was performed in 71 participants. We identified the highest dominant frequency (DFmax) and highest degree of CFAE (CFAEmax) with the use of existing clinical criteria and classified patterns of activation as focal or rotational activation and smooth propagation, conduction block (CB), collision and remnant activity, and fibrillation potentials as single, double, or fractionated potentials containing, respectively, 1, 2, or 3 or more negative deflections. Relationships among activation patterns, DFmax, and potential types were investigated. RESULTS: DFmax were primarily located at the left atrioventricular groove and did not harbor focal activation (proportion focal waves: 0% [IQR: 0%-2%]). Compared with non-DFmax sites, DFmax were characterized by more frequent smooth propagation (22% [IQR: 7%-48%] vs 17% [IQR: 11%-24%]; P = 0.001), less frequent conduction block (69% [IQR: 51%-81%] vs 74% [IQR: 69%-78%]; P = 0.006), a higher proportion of single potentials (72% [IQR: 55%-84%] vs 6%1 [IQR: 55%-65%]; P = 0.003), and a lower proportion of fractionated potentials (4% [IQR: 1%-11%] vs 12% [IQR: 9%-15%]; P = 0.004). CFAEmax were mainly found at the pulmonary veins area, and only 1% [IQR: 0%-2%] of all CFAEmax contained focal activation. Compared with non-CFAEmax sites, CFAEmax sites were characterized by less frequent smooth propagation (1% [IQR: 0%-1%] vs 17% [IQR: 12%-24%]; P < 0.001) and more frequent remnant activity (20% [IQR: 12%-29%] vs 8% [IQR: 5%-10%]; P < 0.001), and harbored predominantly fractionated potentials (52% [IQR: 43%-66%] vs 12% [IQR: 9%-14%]; P < 0.001). CONCLUSIONS: Focal or rotational patterns of activation were not consistently detected at DFmax domains and CFAEmax sites. These findings do not support the concept of targeting DFmax or CFAEmax according to existing criteria for AF ablation.

An accurate and efficient method to train classifiers for atrial fibrillation detection in ECGs: Learning by asking better questions.

BACKGROUND: An increasing number of wearables are capable of measuring electrocardiograms (ECGs), which may help in early detection of atrial fibrillation (AF). Therefore, many studies focus on automated detection of AF in ECGs. A major obstacle is the required amount of manually labelled data. This study aimed to provide an efficient and reliable method to train a classifier for AF detection using large datasets of real-life ECGs. METHOD: Human-controlled semi-supervised learning was applied, consisting of two phases: the pre-training phase and the semi-automated training phase. During pre-training, an initial classifier was trained, which was used to predict the classes of new ECG segments in the semi-automated training phase. Based on the degree of certainty, segments were added to the training dataset automatically or after human validation. Thereafter, the classifier was retrained and this procedure was repeated. To test the model performance, a real-life telemetry dataset containing 3,846,564 30-s ECG segments of hospitalized patients (n = 476) and the CinC Challenge 2017 database were used. RESULTS: After pre-training, the average F1-score on a hidden testing dataset was 89.0%. Furthermore, after the pre-training phase 68.0% of all segments in the hidden test set could be classified with an estimated probability of successful classification of 99%, providing an F1-score of 97.9% for these segments. During the semi-automated training phase, this F1-score showed little variation (97.3%-97.9% in the hidden test set), whilst the number of segments which could be automatically classified increased from 68.0% to 75.8% due to the enhanced training dataset. At the same time, the overall F1-score increased from 89.0% to 91.4%. CONCLUSIONS: Human-validated semi-supervised learning makes training a classifier more time efficient without compromising on accuracy, hence this method might be valuable in the automated detection of AF in real-life ECGs.

Joint cardiac tissue conductivity and activation time estimation using confirmatory factor analysis.

Mathematical models of the electrophysiology of cardiac tissue play an important role when studying heart rhythm disorders like atrial fibrillation. Model parameters such as conductivity, activation time, and anisotropy ratio are useful parameters to determine the arrhythmogenic substrate that causes abnormalities in the atrial tissue. Existing methods often estimate the model parameters separately and assume some of the parameters to be known as a priori knowledge. In this work, we propose an efficient method to jointly estimate the parameters of interest from the cross power spectral density matrix (CPSDM) model of the electrograms. By applying confirmatory factor analysis (CFA) to the CPSDMs of multi-electrode electrograms, we can make use of the spatial information of the data and analyze the relationship between the desired resolution and the required amount of data. With the reasonable assumptions that the conductivity parameters and the anisotropy parameters are constant across different frequencies and heart beats, we estimate these parameters using multiple frequencies and multiple heart beats simultaneously to easier satisfy the identifiability conditions in the CFA problem. Results on the simulated data show that using multiple heart beats decreases the estimation errors of the conductivity and the estimated activation time parameters. The experimental results on clinical data show that using multiple heart beats for parameter estimation can reduce the reconstruction errors of the clinical electrograms, which further demonstrates the robustness of the proposed method.

Classification of De novo post-operative and persistent atrial fibrillation using multi-channel ECG recordings.

Atrial fibrillation (AF) is the most sustained arrhythmia in the heart and also the most common complication developed after cardiac surgery. Due to its progressive nature, timely detection of AF is important. Currently, physicians use a surface electrocardiogram (ECG) for AF diagnosis. However, when the patient develops AF, its various development stages are not distinguishable for cardiologists based on visual inspection of the surface ECG signals. Therefore, severity detection of AF could start from differentiating between short-lasting AF and long-lasting AF. Here, de novo post-operative AF (POAF) is a good model for short-lasting AF while long-lasting AF can be represented by persistent AF. Therefore, we address in this paper a binary severity detection of AF for two specific types of AF. We focus on the differentiation of these two types as de novo POAF is the first time that a patient develops AF. Hence, comparing its development to a more severe stage of AF (e.g., persistent AF) could be beneficial in unveiling the electrical changes in the atrium. To the best of our knowledge, this is the first paper that aims to differentiate these different AF stages. We propose a method that consists of three sets of discriminative features based on fundamentally different aspects of the multi-channel ECG data, namely based on the analysis of RR intervals, a greyscale image representation of the vectorcardiogram, and the frequency domain representation of the ECG. Due to the nature of AF, these features are able to capture both morphological and rhythmic changes in the ECGs. Our classification system consists of a random forest classifier, after a feature selection stage using the ReliefF method. The detection efficiency is tested on 151 patients using 5-fold cross-validation. We achieved 89.07% accuracy in the classification of de novo POAF and persistent AF. The results show that the features are discriminative to reveal the severity of AF. Moreover, inspection of the most important features sheds light on the different characteristics of de novo post-operative and persistent AF.

Cardiac tissue conductivity estimation using confirmatory factor analysis.

Impaired electrical conduction has been shown to play an important role in the development of heart rhythm disorders. Being able to determine the conductivity is important to localize the arrhythmogenic substrate that causes abnormalities in atrial tissue. In this work, we present an algorithm to estimate the conductivity from epicardial electrograms (EGMs) using a high-resolution electrode array. With these arrays, it is possible to measure the propagation of the extracellular potential of the cardiac tissue at multiple positions simultaneously. Given this data, it is in principle possible to estimate the tissue conductivity. However, this is an ill-posed problem due to the large number of unknown parameters in the electrophysiological data model. In this paper, we make use of an effective method called confirmatory factor analysis (CFA), which we apply to the cross correlation matrix of the data to estimate the tissue conductivity. CFA comes with identifiability conditions that need to be satisfied to solve the problem, which is, in this case, estimation of the tissue conductivity. These identifiability conditions can be used to find the relationship between the desired resolution and the required amount of data. Numerical experiments on the simulated data demonstrate that the proposed method can localize the conduction blocks in the tissue and can also estimate the smoother variation in the conductivities. The conductivity values estimated from the clinical data are in line with the values reported in literature and the EGMs reconstructed based on the estimated parameters match well with the clinical EGMs.

Digital biomarkers and algorithms for detection of atrial fibrillation using surface electrocardiograms: A systematic review.

AIMS: Automated detection of atrial fibrillation (AF) in continuous rhythm registrations is essential in order to prevent complications and optimize treatment of AF. Many algorithms have been developed to detect AF in surface electrocardiograms (ECGs) during the past few years. The aim of this systematic review is to gain more insight into these available classification methods by discussing previously used digital biomarkers and algorithms and make recommendations for future research. METHODS: On the 14th of September 2020, the PubMed database was searched for articles focusing on algorithms for AF detection in ECGs using the MeSH terms Atrial Fibrillation, Electrocardiography and Algorithms. Articles which solely focused on differentiation of types of rhythm disorders or prediction of AF termination were excluded. RESULTS: The search resulted in 451 articles, of which 130 remained after full-text screening. Not only did the amount of research on methods for AF detection increase over the past years, but a trend towards more complex classification methods is observed. Furthermore, three different types of features can be distinguished: atrial features, ventricular features, and signal features. Although AF is an atrial disease, only 22% of the described methods use atrial features. CONCLUSION: More and more studies focus on improving accuracy of classification methods for AF in ECGs. As a result, algorithms become increasingly complex and less well interpretable. Only a few studies focus on detecting atrial activity in the ECG. Developing innovative methods focusing on detection of atrial activity might provide accurate classifiers without compromising on transparency.

Analyzing the effect of electrode size on electrogram and activation map properties.

BACKGROUND: Atrial electrograms recorded from the epicardium provide an important tool for studying the initiation, perpetuation, and treatment of AF. However, the properties of these electrograms depend largely on the properties of the electrode arrays that are used for recording these signals. METHOD: In this study, we use the electrode's transfer function to model and analyze the effect of electrode size on the properties of measured electrograms. To do so, we use both simulated as well as clinical data. To simulate electrogram arrays we use a two-dimensional (2D) electrogram model as well as an action propagation model. For clinical data, however, we first estimate the trans-membrane current for a higher resolution 2D modeled cell grid and later use these values to interpolate and model electrograms with different electrode sizes. RESULTS: We simulate electrogram arrays for 2D tissues with 3 different levels of heterogeneity in the conduction and stimulation pattern to model the inhomogeneous wave propagation observed during atrial fibrillation. Four measures are used to characterize the properties of the simulated electrogram arrays of different electrode sizes. The results show that increasing the electrode size increases the error in LAT estimation and decreases the length of conduction block lines. Moreover, visual inspection also shows that the activation maps generated by larger electrodes are more homogeneous with a lower number of observed wavelets. The increase in electrode size also increases the low voltage areas in the tissue while decreasing the slopes and the number of detected deflections. The effect is more pronounced for a tissue with a higher level of heterogeneity in the conduction pattern. Similar conclusions hold for the measurements performed on clinical data. CONCLUSION: The electrode size affects the properties of recorded electrogram arrays which can respectively complicate our understanding of atrial fibrillation. This needs to be considered while performing any analysis on the electrograms or comparing the results of different electrogram arrays.

Assessing the signal quality of electrocardiograms from varied acquisition sources: A generic machine learning pipeline for model generation.

BACKGROUND AND OBJECTIVE: Long-term electrocardiogram monitoring comes at the expense of signal quality. During unconstrained movements, the electrocardiogram is often corrupted by motion artefacts, which can lead to inaccurate physiological information. In this situation, automated quality assessment methods are useful to increase the reliability of the measurements. A generic machine learning pipeline that generates classification models for electrocardiogram quality assessment is presented in this article. The presented pipeline is tested on signals from varied acquisition sources, towards selecting segments that can be used for heart rate analysis in lifestyle applications. METHODS: Electrocardiogram recordings from traditional, wearable and ubiquitous devices, are segmented in 10 s windows and manually labeled by experienced researchers into two quality classes. To capture the electrocardiogram dynamics, a comprehensive set of 43 features is extracted from each segment, based on the time-domain signal, its Fast Fourier Transform, the Autocorrelation function and the Stationary Wavelet Transform. To select the most relevant features for each acquisition source we employ both a customized hybrid approach and the state-of-the-art Neighborhood Component Analysis method and compare them. Support Vector Machines (SVM), Decision Trees, K-Nearest-Neighbors and supervised ensemble methods are tested as possible binary classifiers. RESULTS: The results for the best performing models on traditional, wearable and ubiquitous electrocardiogram datasets are, respectively: balanced-accuracy: 89%, F1-score: 93% with the Fine Gaussian SVM model and 10 features; balanced-accuracy: 93%, F1-score: 93% with the Fine Gaussian SVM model and 11 features; balanced-accuracy: 95%, F1-score: 86%, with the Fine Gaussian SVM model and 8 features. CONCLUSIONS: According to the results, our generic pipeline can generate classification models tailored to individual acquisition sources, provided that a standard Lead I or Lead II is available. Such models accurately establish whether the electrocardiogram quality is good or bad for heart rate analysis. Furthermore, removing bad quality segments decreases errors in heart rate calculation.

Improved local activation time annotation of fractionated atrial electrograms for atrial mapping.

BACKGROUND: Local activation time (LAT) annotation in unipolar electrograms is complicated by interference from nonlocal atrial activities of neighboring tissue. This happens due to the spatial blurring that is inherent to electrogram recordings. In this study, we aim to exploit multi-electrode electrogram recordings to amplify the local activity in each electrogram and subsequently improve the annotation of LATs. METHODS: An electrogram array can be modeled as a spatial convolution of per cell transmembrane currents with an appropriate distance kernel, which depends on the cells' distances to the electrodes. By deconvolving the effect of the distance kernel from the electrogram array, we undo the blurring and estimate the underlying transmembrane currents as our desired local activities. However, deconvolution problems are typically highly ill-posed and result in unstable solutions. To overcome this issue, we propose to use a regularization term that exploits the sparsity of the first-order time derivative of the transmembrane currents. RESULTS: We perform experiments on simulated two-dimensional tissues, as well as clinically recorded electrograms during paroxysmal atrial fibrillation. The results show that the proposed approach for deconvolution can improve the annotation of the true LAT in the electrograms. We also discuss, in summary, the required electrode array specifications for an appropriate recording and subsequent deconvolution. CONCLUSION: By ignoring small but local deflections, algorithms based on steepest descent are prone to generate smoother activation maps. However, by exploiting multi-electrode recordings, we can efficiently amplify small but local deflections and reveal new details in the activation maps that were previously missed.

A compact matrix model for atrial electrograms for tissue conductivity estimation.

Finding the hidden parameters of the cardiac electrophysiological model would help to gain more insight on the mechanisms underlying atrial fibrillation, and subsequently, facilitate the diagnosis and treatment of the disease in later stages. In this work, we aim to estimate tissue conductivity from recorded electrograms as an indication of tissue (mal)functioning. To do so, we first develop a simple but effective forward model to replace the computationally intensive reaction-diffusion equations governing the electrical propagation in tissue. Using the simplified model, we present a compact matrix model for electrograms based on conductivity. Subsequently, we exploit the simplicity of the compact model to solve the ill-posed inverse problem of estimating tissue conductivity. The algorithm is demonstrated on simulated data as well as on clinically recorded data. The results show that the model allows to efficiently estimate the conductivity map. In addition, based on the estimated conductivity, realistic electrograms can be regenerated demonstrating the validity of the model.

Local Activation Time Estimation in Fractionated Electrograms of Cardiac Mappings.

In this study, we propose a novel approach for estimation of local activation times (LATs) in fractionated electrograms. Using an electrophysiological tissue model, we first formulate the electrogram array as a convolution of transmembrane currents with a distance kernel. These currents are more local activities and less affected by the heterogeneity in the tissue compared to electrograms. We then deconvolve the distance kernel with the electrograms to reconstruct the transmembrane current. To stabilize the solution of this ill-posed deconvolution, we use spatio-temporal total variation as a regularization. This helps to preserve sharp spatial and temporal deflections in the currents that are of higher importance in LAT estimation. Finally, the maximum negative slope of the reconstructed transmembrane currents are used to estimate the LATs. Instrumental comparison to two reference approaches shows that the proposed approach performs better in estimating the LATs in fractionated electrograms.

An evaluation of objective measures for intelligibility prediction of time-frequency weighted noisy speech.

Existing objective speech-intelligibility measures are suitable for several types of degradation, however, it turns out that they are less appropriate in cases where noisy speech is processed by a time-frequency weighting. To this end, an extensive evaluation is presented of objective measure for intelligibility prediction of noisy speech processed with a technique called ideal time frequency (TF) segregation. In total 17 measures are evaluated, including four advanced speech-intelligibility measures (CSII, CSTI, NSEC, DAU), the advanced speech-quality measure (PESQ), and several frame-based measures (e.g., SSNR). Furthermore, several additional measures are proposed. The study comprised a total number of 168 different TF-weightings, including unprocessed noisy speech. Out of all measures, the proposed frame-based measure MCC gave the best results (ρ = 0.93). An additional experiment shows that the good performing measures in this study also show high correlation with the intelligibility of single-channel noise reduced speech.