Parametric Modeling of Electrocardiograms using Particle Swarm optimization.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Computational Cardiology

Background:

• Electrocardiograms (ECG) are crucial for diagnosing heart conditions.
• Parametric ECG models can generate synthetic ECG signals for research and development.
• Existing models may have limitations in capturing diverse cardiac rhythms and morphologies.

Purpose of the Study:

• To develop a framework for fitting parametric ECG generator models to reference signals.
• To enable the generation of realistic ECG-like signals, particularly for disease states.
• To validate the utility of parametric ECG models for device validation and algorithm evaluation.

Main Methods:

• Designed a parametric ECG generator with minimal assumptions on single beat waveform morphology.
• Employed Particle Swarm Optimization (PSO) to determine optimal model parameters.
• Minimized the percent root mean square difference (PRD) between reference and generated signals.

Main Results:

• Achieved high fidelity in generating ECG waveforms for normal, idioventricular, and ventricular flutter rhythms.
• Attained Pearson correlation coefficients greater than 0.9 between generated and recorded signals from the MIT-BIH database.
• Demonstrated the model's capability to capture complex waveform morphologies.

Conclusions:

• The proposed framework effectively fits parametric ECG generator models to reference signals.
• The method provides a valuable tool for creating synthetic ECG data for various applications.
• This approach supports advancements in ECG compression, modeling, and cardiac device validation.