Optical Mapping-Validated Machine Learning Improves Atrial Fibrillation Driver Detection by Multi-Electrode Mapping

Role of Machine Learning and Artificial Intelligence in Arrhythmias and Electrophysiology

Machine learning (ML), a subset of artificial intelligence (AI) centered on machines learning from extensive datasets, stands at the forefront of a technological revolution shaping various facets of society. Cardiovascular medicine has emerged as a key domain for ML applications, with considerable efforts to integrate these innovations into routine clinical practice. Within cardiac electrophysiology, ML applications, especially in the automated interpretation of electrocardiograms, have garnered substantial attention in existing literature. However, less recognized are the diverse applications of ML in cardiac electrophysiology and arrhythmias, spanning basic science research on arrhythmia mechanisms, both experimental and computational, as well as contributions to enhanced techniques for mapping cardiac electrical function and translational research related to arrhythmia management. This comprehensive review delves into various ML applications within the scope of this journal, organized into 3 parts. The first section provides a fundamental understanding of general ML principles and methodologies, serving as a foundational resource for readers interested in exploring ML applications in arrhythmia research. The second part offers an in-depth review of studies in arrhythmia and electrophysiology that leverage ML methodologies, showcasing the broad potential of ML approaches. Each subject is thoroughly outlined, accompanied by a review of notable ML research advancements. Finally, the review delves into the primary challenges and future perspectives surrounding ML-driven cardiac electrophysiology and arrhythmias research.

Transfer Learning for Improved Classification of Drivers in Atrial Fibrillation

"Drivers" are theorized mechanisms for persistent atrial fibrillation. Machine learning algorithms have been used to identify drivers, but the small size of current driver datasets limits their performance. We hypothesized that pretraining with unsupervised learning on a large dataset of unlabeled electrograms would improve classifier accuracy on a smaller driver dataset. In this study, we used a SimCLR-based framework to pretrain a residual neural network on a dataset of 113K unlabeled 64-electrode measurements and found weighted testing accuracy to improve over a non-pretrained network (78.6±3.9% vs 71.9±3.3%). This lays ground for development of superior driver detection algorithms and supports use of transfer learning for other datasets of endocardial electrograms.

Complexity analysis of electrical activity during endocardial and epicardial biventricular mapping of ventricular fibrillation

BACKGROUND

Ventricular fibrillation (VF) is a lethal cardiac arrhythmia that is a significant cause of sudden cardiac death. Comprehensive studies of spatiotemporal characteristics of VF in situ are difficult to perform with current mapping systems and catheter technology.

OBJECTIVE

The goal of this study was to develop a computational approach to characterize VF using a commercially available technology in a large animal model. Prior data suggests that characterization of spatiotemporal organization of electrical activity during VF can be used to provide better mechanistic understanding and potential ablation targets to modify VF and its substrate. We therefore evaluated intracardiac electrograms during biventricular mapping of the endocardium (ENDO) and epicardium (EPI) in acute canine studies.

METHODS

To develop thresholds for organized and disorganized activity, a linear discriminant analysis (LDA)-based approach was performed to the known organized and disorganized activities recorded in ex vivo Langendorff-perfused rat and rabbit hearts using optical mapping experiments. Several frequency- and time-domain approaches were used as individual and paired features to identify the optimal thresholds for the LDA approach. Subsequently, VF was sequentially mapped in 4 canine hearts, using the CARTO mapping system with a multipolar mapping catheter in the ENDO left and right ventricles and EPI to capture the progression of VF at 3 discrete post-induction time intervals: VF period 1 (just after induction of VF to 15 min), VF period 2 (15 to 30 min), and VF period 3 (30 to 45 min). The developed LDA model, cycle lengths (CL), and regularity indices (RI) were applied to all recorded intracardiac electrograms to quantify the spatiotemporal organization of VF in canine hearts.

RESULTS

We demonstrated the presence of organized activity in the EPI as VF progresses, in contrary to the ENDO, where the activity stays disorganized. The shortest CL always occurred in the ENDO, especially the RV, indicating a faster VF activity. The highest RI was found in the EPI in all hearts for all VF stages, indicating spatiotemporal consistency of RR intervals.

CONCLUSION

We identified electrical organization and spatiotemporal differences throughout VF in canine hearts from induction to asystole. Notably, the RV ENDO is characterized by a high level of disorganization and faster VF frequency. In contrast, EPI has a high spatiotemporal organization of VF and consistently long RR intervals.

Animal Disease Models and Patient-iPS-Cell-Derived In Vitro Disease Models for Cardiovascular Biology-How Close to Disease?

Currently, zebrafish, rodents, canines, and pigs are the primary disease models used in cardiovascular research. In general, larger animals have more physiological similarities to humans, making better disease models. However, they can have restricted or limited use because they are difficult to handle and maintain. Moreover, animal welfare laws regulate the use of experimental animals. Different species have different mechanisms of disease onset. Organs in each animal species have different characteristics depending on their evolutionary history and living environment. For example, mice have higher heart rates than humans. Nonetheless, preclinical studies have used animals to evaluate the safety and efficacy of human drugs because no other complementary method exists. Hence, we need to evaluate the similarities and differences in disease mechanisms between humans and experimental animals. The translation of animal data to humans contributes to eliminating the gap between these two. In vitro disease models have been used as another alternative for human disease models since the discovery of induced pluripotent stem cells (iPSCs). Human cardiomyocytes have been generated from patient-derived iPSCs, which are genetically identical to the derived patients. Researchers have attempted to develop in vivo mimicking 3D culture systems. In this review, we explore the possible uses of animal disease models, iPSC-derived in vitro disease models, humanized animals, and the recent challenges of machine learning. The combination of these methods will make disease models more similar to human disease.

Exploring interpretability in deep learning prediction of successful ablation therapy for atrial fibrillation

Background: Radiofrequency catheter ablation (RFCA) therapy is the first-line treatment for atrial fibrillation (AF), the most common type of cardiac arrhythmia globally. However, the procedure currently has low success rates in dealing with persistent AF, with a reoccurrence rate of ∼50% post-ablation. Therefore, deep learning (DL) has increasingly been applied to improve RFCA treatment for AF. However, for a clinician to trust the prediction of a DL model, its decision process needs to be interpretable and have biomedical relevance.

Aim: This study explores interpretability in DL prediction of successful RFCA therapy for AF and evaluates if pro-arrhythmogenic regions in the left atrium (LA) were used in its decision process.

Methods: AF and its termination by RFCA have been simulated in MRI-derived 2D LA tissue models with segmented fibrotic regions (n = 187). Three ablation strategies were applied for each LA model: pulmonary vein isolation (PVI), fibrosis-based ablation (FIBRO) and a rotor-based ablation (ROTOR). The DL model was trained to predict the success of each RFCA strategy for each LA model. Three feature attribution (FA) map methods were then used to investigate interpretability of the DL model: GradCAM, Occlusions and LIME.

Results: The developed DL model had an AUC (area under the receiver operating characteristic curve) of 0.78 ± 0.04 for predicting the success of the PVI strategy, 0.92 ± 0.02 for FIBRO and 0.77 ± 0.02 for ROTOR. GradCAM had the highest percentage of informative regions in the FA maps (62% for FIBRO and 71% for ROTOR) that coincided with the successful RFCA lesions known from the 2D LA simulations, but unseen by the DL model. Moreover, GradCAM had the smallest coincidence of informative regions of the FA maps with non-arrhythmogenic regions (25% for FIBRO and 27% for ROTOR).

Conclusion: The most informative regions of the FA maps coincided with pro-arrhythmogenic regions, suggesting that the DL model leveraged structural features of MRI images to identify such regions and make its prediction. In the future, this technique could provide a clinician with a trustworthy decision support tool.

Utilizing machine learning algorithms to predict subject genetic mutation class from in silico models of neuronal networks

Background

Epilepsy is the fourth-most common neurological disorder, affecting an estimated 50 million patients globally. Nearly 40% of patients have uncontrolled seizures yet incur 80% of the cost. Anti-epileptic drugs commonly result in resistance and reversion to uncontrolled drug-resistant epilepsy and are often associated with significant adverse effects. This has led to a trial-and-error system in which physicians spend months to years attempting to identify the optimal therapeutic approach.

Objective

To investigate the potential clinical utility from the context of optimal therapeutic prediction of characterizing cellular electrophysiology. It is well-established that genomic data alone can sometimes be predictive of effective therapeutic approach. Thus, to assess the predictive power of electrophysiological data, machine learning strategies are implemented to predict a subject’s genetically defined class in an in silico model using brief electrophysiological recordings obtained from simulated neuronal networks.

Methods

A dynamic network of isogenic neurons is modeled in silico for 1-s for 228 dynamically modeled patients falling into one of three categories: healthy, general sodium channel gain of function, or inhibitory sodium channel loss of function. Data from previous studies investigating the electrophysiological and cellular properties of neurons in vitro are used to define the parameters governing said models. Ninety-two electrophysiological features defining the nature and consistency of network connectivity, activity, waveform shape, and complexity are extracted for each patient network and t-tests are used for feature selection for the following machine learning algorithms: Neural Network, Support Vector Machine, Gaussian Naïve Bayes Classifier, Decision Tree, and Gradient Boosting Decision Tree. Finally, their performance in accurately predicting which genetic category the subjects fall under is assessed.

Results

Several machine learning algorithms excel in using electrophysiological data from isogenic neurons to accurately predict genetic class with a Gaussian Naïve Bayes Classifier predicting healthy, gain of function, and overall, with the best accuracy, area under the curve, and F1. The Gradient Boosting Decision Tree performs the best for loss of function models indicated by the same metrics.

Conclusions

It is possible for machine learning algorithms to use electrophysiological data to predict clinically valuable metrics such as optimal therapeutic approach, especially when combining several models.

Genetic and non-genetic risk factors associated with atrial fibrillation

Atrial fibrillation (AF) is the most common arrhythmic disorder and its prevalence in the United States is projected to increase to more than twelve million cases in 2030. AF increases the risk of other forms of cardiovascular disease, including stroke. As the incidence of atrial fibrillation increases dramatically with age, it is paramount to elucidate risk factors underlying AF pathogenesis. Here, we review tissue and cellular pathways underlying AF, as well as critical components that impact AF susceptibility including genetic and environmental risk factors. Finally, we provide the latest information on potential links between SARS-CoV-2 and human AF. Improved understanding of mechanistic pathways holds promise in preventative care and early diagnostics, and also introduces novel targeted forms of therapy that might attenuate AF progression and maintenance.

Convolutional Neural Networks for Mechanistic Driver Detection in Atrial Fibrillation

The maintaining and initiating mechanisms of atrial fibrillation (AF) remain controversial. Deep learning is emerging as a powerful tool to better understand AF and improve its treatment, which remains suboptimal. This paper aims to provide a solution to automatically identify rotational activity drivers in endocardial electrograms (EGMs) with convolutional recurrent neural networks (CRNNs). The CRNN model was compared with two other state-of-the-art methods (SimpleCNN and attention-based time-incremental convolutional neural network (ATI-CNN)) for different input signals (unipolar EGMs, bipolar EGMs, and unipolar local activation times), sampling frequencies, and signal lengths. The proposed CRNN obtained a detection score based on the Matthews correlation coefficient of 0.680, an ATI-CNN score of 0.401, and a SimpleCNN score of 0.118, with bipolar EGMs as input signals exhibiting better overall performance. In terms of signal length and sampling frequency, no significant differences were found. The proposed architecture opens the way for new ablation strategies and driver detection methods to better understand the AF problem and its treatment.

Big Data in electrophysiology

The quantity of data produced and captured in medicine today is unprecedented. Technological improvements and automation have expanded the traditional statistical methods and enabled the analysis of Big Data. This has permitted the discovery of new associations with a granularity that was previously hidden to human eyes. In the first part of this review, the authors would like to provide an overview of basic Machine Learning (ML) principles and techniques in order to better understand their application in recent publications about cardiac arrhythmias. In the second part, ML-enabled advances in disease detection and diagnosis, outcome prediction, and novel disease characterization in topics like electrocardiography, atrial fibrillation, ventricular arrhythmias, and cardiac devices are presented. Finally, the limitations and challenges of applying ML in clinical practice, such as validation, replication, generalizability, and regulatory issues, are discussed. More carefully designed studies and collaborations are needed for ML to become feasible, trustworthy, accurate, and reproducible and to reach its full potential for patient-oriented precision medicine.

Atrial fibrillation driver identification through regional mutual information networks: a modeling perspective

Purpose

Effective identification of electrical drivers within remodeled tissue is a key for improving ablation treatment for atrial fibrillation. We have developed a mutual information, graph-based approach to identify and propose fault tolerance metric of local efficiency as a distinguishing feature of rotational activation and remodeled atrial tissue.

Methods

Voltage data were extracted from atrial tissue simulations (2D Karma, 3D physiological, and the Multiscale Cardiac Simulation Framework (MSCSF)) using multi-spline open and parallel regional mapping catheter geometries. Graphs were generated based on varied mutual information thresholds between electrode pairs and the local efficiency for each graph was calculated.

Results

High-resolution mapping catheter geometries can distinguish between rotational and irregular activation patterns using the derivative of local efficiency as a function of increasing mutual information threshold. The derivative is decreased for rotational activation patterns comparing to irregular activations in both a simplified 2D model (0.0017 ± 1 × 10⁻⁴ vs. 0.0032 ± 1 × 10⁻⁴, p < 0.01) and a more realistic 3D model (0.00092 ± 5 × 10⁻⁵ vs. 0.0014 ± 4 × 10⁻⁵, p < 0.01). Average local efficiency derivative can also distinguish between degrees of remodeling. Simulations using the MSCSF model, with 10 vs. 90% remodeling, display distinct derivatives in the grid design parallel spline catheter configuration (0.0015 ± 5 × 10⁻⁵ vs. 0.0019 ± 6 × 10⁻⁵, p < 0.01) and the flower shaped open spline configuration (0.0011 ± 5 × 10⁻⁵ vs. 0.0016 ± 4 × 10⁻⁵, p < 0.01).

Conclusion

A decreased derivative of local efficiency characterizes rotational activation and varies with atrial remodeling. This suggests a distinct communication pattern in cardiac rotational activation detectable via high-resolution regional mapping and could enable identification of electrical drivers for targeted ablation.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10840-021-01101-z.

Deep Learning Classification of Unipolar Electrograms in Human Atrial Fibrillation: Application in Focal Source Mapping

Focal sources are potential targets for atrial fibrillation (AF) catheter ablation, but they can be time-consuming and challenging to identify when unipolar electrograms (EGM) are numerous and complex. Our aim was to apply deep learning (DL) to raw unipolar EGMs in order to automate putative focal sources detection. We included 78 patients from the Focal Source and Trigger (FaST) randomized controlled trial that evaluated the efficacy of adjunctive FaST ablation compared to pulmonary vein isolation alone in reducing AF recurrence. FaST sites were identified based on manual classification of sustained periodic unipolar QS EGMs over 5-s. All periodic unipolar EGMs were divided into training (n = 10,004) and testing cohorts (n = 3,180). DL was developed using residual convolutional neural network to discriminate between FaST and non-FaST. A gradient-based method was applied to interpret the DL model. DL classified FaST with a receiver operator characteristic area under curve of 0.904 ± 0.010 (cross-validation) and 0.923 ± 0.003 (testing). At a prespecified sensitivity of 90%, the specificity and accuracy were 81.9 and 82.5%, respectively, in detecting FaST. DL had similar performance (sensitivity 78%, specificity 89%) to that of FaST re-classification by cardiologists (sensitivity 78%, specificity 79%). The gradient-based interpretation demonstrated accurate tracking of unipolar QS complexes by select DL convolutional layers. In conclusion, our novel DL model trained on raw unipolar EGMs allowed automated and accurate classification of FaST sites. Performance was similar to FaST re-classification by cardiologists. Future application of DL to classify FaST may improve the efficiency of real-time focal source detection for targeted AF ablation therapy.

The Role of Artificial Intelligence and Machine Learning in Clinical Cardiac Electrophysiology

In recent years, artificial intelligence (AI) has found numerous applications in cardiology due in part to large digitized datasets and the evolution of high performance computing. In the discipline of cardiac electrophysiology (EP), a number of clinical, imaging, and electrical waveform data are considered in the diagnosis, prognostication and management of arrhythmias, which lend themselves well to automation through AI. But equally relevant, AI offers a unique opportunity to discover novel EP concepts and improve clinical care through its inherent, hierarchical tenets of self-learning. This review will focus on the application of AI in clinical EP and summarize state-of-the art, large, clinical studies in the following key domains: (1) ECG-based arrhythmia and disease classification, (2) atrial fibrillation source detection, (3) substrate and risk assessment for atrial fibrillation and ventricular tachyarrhythmias, and (4) predicting outcomes after cardiac resynchronization therapy. Many are small, single-center, proof-of-concept investigations, but they still demonstrate groundbreaking performance of deep learning, a subdomain of AI, which surpasses traditional statistical analysis. Larger studies, for instance classifying arrhythmias from ECG recordings, have further provided external validation of their high accuracy. Ultimately, the performance of AI is dependent on the quality of the input data and the rigor of algorithm development. The field is still nascent and several barriers will need to be overcome, including prospective validation in large, well-labelled datasets and more seamless information technology-based data collection/integration, before AI can be adopted into broader clinical EP practice. This review will conclude with a discussion of these challenges and future work.

Understanding PITX2-Dependent Atrial Fibrillation Mechanisms through Computational Models

Atrial fibrillation (AF) is a common arrhythmia. Better prevention and treatment of AF are needed to reduce AF-associated morbidity and mortality. Several major mechanisms cause AF in patients, including genetic predispositions to AF development. Genome-wide association studies have identified a number of genetic variants in association with AF populations, with the strongest hits clustering on chromosome 4q25, close to the gene for the homeobox transcription PITX2. Because of the inherent complexity of the human heart, experimental and basic research is insufficient for understanding the functional impacts of PITX2 variants on AF. Linking PITX2 properties to ion channels, cells, tissues, atriums and the whole heart, computational models provide a supplementary tool for achieving a quantitative understanding of the functional role of PITX2 in remodelling atrial structure and function to predispose to AF. It is hoped that computational approaches incorporating all we know about PITX2-related structural and electrical remodelling would provide better understanding into its proarrhythmic effects leading to development of improved anti-AF therapies. In the present review, we discuss advances in atrial modelling and focus on the mechanistic links between PITX2 and AF. Challenges in applying models for improving patient health are described, as well as a summary of future perspectives.

Computational models of atrial fibrillation: achievements, challenges, and perspectives for improving clinical care

Despite significant advances in its detection, understanding and management, atrial fibrillation (AF) remains a highly prevalent cardiac arrhythmia with a major impact on morbidity and mortality of millions of patients. AF results from complex, dynamic interactions between risk factors and comorbidities that induce diverse atrial remodelling processes. Atrial remodelling increases AF vulnerability and persistence, while promoting disease progression. The variability in presentation and wide range of mechanisms involved in initiation, maintenance and progression of AF, as well as its associated adverse outcomes, make the early identification of causal factors modifiable with therapeutic interventions challenging, likely contributing to suboptimal efficacy of current AF management. Computational modelling facilitates the multilevel integration of multiple datasets and offers new opportunities for mechanistic understanding, risk prediction and personalized therapy. Mathematical simulations of cardiac electrophysiology have been around for 60 years and are being increasingly used to improve our understanding of AF mechanisms and guide AF therapy. This narrative review focuses on the emerging and future applications of computational modelling in AF management. We summarize clinical challenges that may benefit from computational modelling, provide an overview of the different in silico approaches that are available together with their notable achievements, and discuss the major limitations that hinder the routine clinical application of these approaches. Finally, future perspectives are addressed. With the rapid progress in electronic technologies including computing, clinical applications of computational modelling are advancing rapidly. We expect that their application will progressively increase in prominence, especially if their added value can be demonstrated in clinical trials.

How machine learning is impacting research in atrial fibrillation: implications for risk prediction and future management

There has been an exponential growth of artificial intelligence (AI) and machine learning (ML) publications aimed at advancing our understanding of atrial fibrillation (AF), which has been mainly driven by the confluence of two factors: the advances in deep neural networks (DeepNNs) and the availability of large, open access databases. It is observed that most of the attention has centred on applying ML for dvsetecting AF, particularly using electrocardiograms (ECGs) as the main data modality. Nearly a third of them used DeepNNs to minimize or eliminate the need for transforming the ECGs to extract features prior to ML modelling; however, we did not observe a significant advantage in following this approach. We also found a fraction of studies using other data modalities, and others centred in aims, such as risk prediction, AF management, and others. From the clinical perspective, AI/ML can help expand the utility of AF detection and risk prediction, especially for patients with additional comorbidities. The use of AI/ML for detection and risk prediction into applications and smart mobile health (mHealth) technology would enable ‘real time’ dynamic assessments. AI/ML could also adapt to treatment changes over time, as well as incident risk factors. Incorporation of a dynamic AI/ML model into mHealth technology would facilitate ‘real time’ assessment of stroke risk, facilitating mitigation of modifiable risk factors (e.g. blood pressure control). Overall, this would lead to an improvement in clinical care for patients with AF.

Machine Learning in Arrhythmia and Electrophysiology

Machine learning (ML), a branch of artificial intelligence, where machines learn from big data, is at the crest of a technological wave of change sweeping society. Cardiovascular medicine is at the forefront of many ML applications, and there is a significant effort to bring them into mainstream clinical practice. In the field of cardiac electrophysiology, ML applications have also seen a rapid growth and popularity, particularly the use of ML in the automatic interpretation of ECGs, which has been extensively covered in the literature. Much lesser known are the other aspects of ML application in cardiac electrophysiology and arrhythmias, such as those in basic science research on arrhythmia mechanisms, both experimental and computational; in the development of better techniques for mapping of cardiac electrical function; and in translational research related to arrhythmia management. In the current review, we examine comprehensively such ML applications as they match the scope of this journal. The current review is organized in 3 parts. The first provides an overview of general ML principles and methodologies that will afford readers of the necessary information on the subject, serving as the foundation for inviting further ML applications in arrhythmia research. The basic information we provide can serve as a guide on how one might design and conduct an ML study. The second part is a review of arrhythmia and electrophysiology studies in which ML has been utilized, highlighting the broad potential of ML approaches. For each subject, we outline comprehensively the general topics, while reviewing some of the research advances utilizing ML under the subject. Finally, we discuss the main challenges and the perspectives for ML-driven cardiac electrophysiology and arrhythmia research.

Artificial intelligence for a personalized diagnosis and treatment of atrial fibrillation

Although atrial fibrillation (AF) is the most common cardiac arrhythmia, its early identification, diagnosis, and treatment is still challenging. Due to its heterogeneous mechanisms and risk factors, targeting an individualized treatment of AF demands a large amount of patient data to identify specific patterns. Artificial intelligence (AI) algorithms are particularly well suited for treating high-dimensional data, predicting outcomes, and eventually, optimizing strategies for patient management. The analysis of large patient samples combining different sources of information such as blood biomarkers, electrical signals, and medical images opens a new paradigm for improving diagnostic algorithms. In this review, we summarize suitable AI techniques for this purpose. In particular, we describe potential applications for understanding the structural and functional bases of the disease, as well as for improving early noninvasive diagnosis, developing more efficient therapies, and predicting long-term clinical outcomes of patients with AF.

Detection of Atrial Fibrillation Using a Machine Learning Approach

The atrial fibrillation (AF) is one of the most well-known cardiac arrhythmias in clinical practice, with a prevalence of 1–2% in the community, which can increase the risk of stroke and myocardial infarction. The detection of AF electrocardiogram (ECG) can improve the early detection of diagnosis. In this paper, we have further developed a framework for processing the ECG signal in order to determine the AF episodes. We have implemented machine learning and deep learning algorithms to detect AF. Moreover, the experimental results show that better performance can be achieved with long short-term memory (LSTM) as compared to other algorithms. The initial experimental results illustrate that the deep learning algorithms, such as LSTM and convolutional neural network (CNN), achieved better performance (10%) as compared to machine learning classifiers, such as support vectors, logistic regression, etc. This preliminary work can help clinicians in AF detection with high accuracy and less probability of errors, which can ultimately result in reduction in fatality rate.

Comprehensive evaluation of electrophysiological and 3D structural features of human atrial myocardium with insights on atrial fibrillation maintenance mechanisms

Atrial fibrillation (AF) occurrence and maintenance is associated with progressive remodeling of electrophysiological (repolarization and conduction) and 3D structural (fibrosis, fiber orientations, and wall thickness) features of the human atria. Significant diversity in AF etiology leads to heterogeneous arrhythmogenic electrophysiological and structural substrates within the 3D structure of the human atria. Since current clinical methods have yet to fully resolve the patient-specific arrhythmogenic substrates, mechanism-based AF treatments remain underdeveloped. Here, we review current knowledge from in-vivo, ex-vivo, and in-vitro human heart studies, and discuss how these studies may provide new insights on the synergy of atrial electrophysiological and 3D structural features in AF maintenance. In-vitro studies on surgically acquired human atrial samples provide a great opportunity to study a wide spectrum of AF pathology, including functional changes in single-cell action potentials, ion channels, and gene/protein expression. However, limited size of the samples prevents evaluation of heterogeneous AF substrates and reentrant mechanisms. In contrast, coronary-perfused ex-vivo human hearts can be studied with state-of-the-art functional and structural technologies, such as high-resolution near-infrared optical mapping and contrast-enhanced MRI. These imaging modalities can resolve atrial arrhythmogenic substrates and their role in reentrant mechanisms maintaining AF and validate clinical approaches. Nonetheless, longitudinal studies are not feasible in explanted human hearts. As no approach is perfect, we suggest that combining the strengths of direct human atrial studies with high fidelity approaches available in the laboratory and in realistic patient-specific computer models would elucidate deeper knowledge of AF mechanisms. We propose that a comprehensive translational pipeline from ex-vivo human heart studies to longitudinal clinically relevant AF animal studies and finally to clinical trials is necessary to identify patient-specific arrhythmogenic substrates and develop novel AF treatments.