3D Heart Simulation And Recognition Of Various Events

Area of Science:

• Biomedical Engineering
• Computational Electrophysiology
• Medical Imaging

Background:

• The inverse problem of electrocardiography (ECG) is crucial for non-invasively determining cardiac electrical activity.
• Existing methods face challenges with complex arrhythmias and computational demands.
• Accurate heart model parameter estimation is essential for reliable ECG inverse solutions.

Purpose of the Study:

• To develop and validate a robust method for solving the ECG inverse problem using heart model parameter optimization.
• To enhance the efficiency of the inverse solution by employing an artificial neural network (ANN) for preliminary ECG analysis.
• To assess the accuracy of the proposed method in identifying various cardiac conditions, including arrhythmias.

Main Methods:

• An optimization system for heart model parameters was developed for inverse ECG problem-solving.
• An artificial neural network (ANN)-based preliminary ECG analyzer was created to reduce the search space for the optimization algorithm.
• Optimal model parameters were determined by minimizing objective functions comparing observed and model-generated body surface ECGs.

Main Results:

• The developed method achieved approximately 86% accuracy in final evaluations, as validated by physicians.
• The approach demonstrated robustness in handling various malfunction cases, including Wolff-Parkinson-White (WPW) syndrome, atrial fibrillation, and ventricular fibrillation.
• The ANN-based analyzer effectively reduced the computational search space for the optimization algorithm.

Conclusions:

• The proposed event estimation and recognition method provides a robust inverse solution for electrocardiography.
• This approach effectively circumvents many difficulties associated with the traditional ECG inverse problem.
• The high accuracy, even with complex arrhythmias, suggests significant potential for clinical application.