Improved event interval reconstruction in synthetic electrocardiograms.

Area of Science:

• Biomedical Signal Processing
• Computational Physiology
• Medical Informatics

Background:

• Mathematical models for synthesizing electrocardiograms (ECGs) are crucial for reproducible research in biomedical signal processing.
• Current state-of-the-art models like ECGsyn capture ECG morphology, heart rate variability, and QT-interval adaptation.

Purpose of the Study:

• To address the time-domain limitations of the ECGsyn model in reconstructing ECG time series from wave annotations.
• To present an alternative algorithm, ECGfm, for improved fidelity in recreating ECG event intervals.

Main Methods:

• Demonstrated the time-domain limitations of the ECGsyn algorithm.
• Developed and applied the ECGfm algorithm, utilizing wave annotations as a phase constraint for signal reconstruction.
• Evaluated algorithm performance on data from the Physionet Physobank QT database.

Main Results:

• Identified significant time-domain limitations in the ECGsyn model for ECG time series reconstruction.
• The proposed ECGfm algorithm demonstrated superior performance in recreating event intervals with greater fidelity.
• Results confirm the improved accuracy of ECGfm compared to existing methods.

Conclusions:

• The ECGfm algorithm offers a more accurate method for synthesizing ECG signals based on wave annotations.
• This advancement is expected to benefit biomedical signal processing research and technology development.
• ECGfm provides a more reliable tool for generating standard ECG signals for research purposes.