Evaluation of an automatic cardiac activation detector for bipolar electrograms

Area of Science:

• Biomedical Engineering
• Computational Cardiology
• Signal Processing

Background:

• Accurate identification of local activation times (LAT) in bipolar cardiac electrograms is crucial for isochronal map construction.
• Manual LAT assignment is time-consuming and prone to variability due to complex waveform characteristics.

Purpose of the Study:

• To develop and evaluate a computer program (AP) for automated, reliable identification of local activation times in bipolar cardiac electrograms.
• To assess the performance of the AP program against human expert assignments.

Main Methods:

• A computer program, AP, was designed using a set of empirical rules implemented as a syntactic analyzer.
• Canine epicardial recordings were used for evaluation.
• Receiver Operating Characteristic (ROC) analysis, the Hermes-Cox model, and bootstrap statistics were employed for performance evaluation.

Main Results:

• The AP program demonstrated reliability in event detection and LAT assignment comparable to independent human investigators.
• Discrepancy analysis confirmed the consistency of AP's performance relative to expert evaluations.
• The system's design and evaluation methodologies are broadly applicable to waveform component detection.

Conclusions:

• The AP computer program provides an effective and reliable method for automating the identification of local activation times in cardiac electrograms.
• This automation can significantly improve the efficiency and consistency of cardiac mapping procedures.
• The developed approach offers a transferable methodology for waveform analysis in various signal processing applications.