Current Trends in Cardiac Electrophysiology and Imaging

Recent advancements in cardiac electrophysiology and imaging have shown significant progress in the development of more accurate and efficient models for simulating cardiac arrhythmias and detecting atrial fibrillation. The field is moving towards integrating physics-informed neural networks (PINNs) and ensemble learning techniques to enhance the precision and speed of cardiac simulations and diagnostic tools. These innovations aim to address the computational challenges associated with detailed cardiac models while improving the reliability of atrial fibrillation detection systems, particularly those that can be deployed on mobile devices.

In the realm of cardiac simulations, there is a notable shift towards faster yet physiologically accurate models that can handle complex re-entrant dynamics, crucial for understanding and predicting arrhythmias. This includes the development of eikonal models that incorporate tissue re-excitability, enabling real-time simulations that could be clinically applicable.

For atrial fibrillation detection, the focus is on creating more robust and user-friendly systems that can operate on smartphones, leveraging advanced signal processing techniques and deep learning models to achieve high accuracy and reliability. These systems aim to provide early and continuous monitoring, overcoming the limitations of traditional, cost-prohibitive devices.

In the area of cardiac fiber orientation, the use of ensemble learning with PINNs is advancing the ability to estimate fiber fields with quantified uncertainty, which is essential for personalized cardiac digital twins and precision medicine.

Noteworthy Developments

• Physics-Informed Neural Networks (PINNs) for Cerebral Blood Flow Estimation: A novel PINN approach significantly improves the accuracy of multi-parameter perfusion estimation from noisy ASL data in infants, potentially advancing cardio-brain network physiology understanding.
• Mobile Atrial Fibrillation Detection System: A smartphone-based system achieves high accuracy in AF detection using advanced signal processing and a ResNet-based model, offering a portable and user-friendly solution for early monitoring.
• Eikonal Model with Re-Excitability: This model enables fast and accurate simulations of cardiac arrhythmias by incorporating tissue re-excitability, paving the way for real-time clinical applications.
• Ensemble Learning of Atrial Fiber Orientation: An ensemble learning approach with PINNs enhances the estimation of cardiac fiber fields with quantified uncertainty, crucial for personalized cardiac modeling.