Gaussian Process-Based Spatiotemporal Modeling of Electrical Wave Propagation in Human Atrium^.

Area of Science:

• Computational biology
• Cardiac electrophysiology
• Biomedical modeling

Background:

• Cardiac rhythm regularity relies on electrical impulse propagation through the cardiac conduction system.
• Abnormal electrical activity can cause severe cardiac disorders and sudden death.
• Understanding human heart electrical activity is crucial for diagnosing cardiac disorders and designing therapies.

Purpose of the Study:

• To develop a computationally efficient metamodel for simulating cardiac electrical activity.
• To address the computational challenges posed by complex, multi-scale human heart models.
• To accurately reconstruct spatiotemporal variations in cardiac cell membrane potential.

Main Methods:

• Developed a metamodel using Gaussian Process (GP) regression.
• Reconstructed spatiotemporal variations of cell membrane potential in the left atrium.
• Utilized four different covariance functions to infer potential distributions.

Main Results:

• The GP metamodel accurately estimates the spatiotemporal propagation of cardiac electrical waves.
• The model requires a smaller dataset compared to traditional simulation models.
• Demonstrated significant computational advantages over traditional cardiac simulation models.

Conclusions:

• Gaussian Process metamodels offer an accurate and computationally efficient alternative for cardiac electrophysiology simulations.
• This approach can accelerate the study of cardiac function and the development of therapeutic strategies.
• The developed metamodel aids in understanding electrical wave propagation in the human heart.