ECG characterization of paroxysmal atrial fibrillation: parameter extraction and automatic diagnosis algorithm

Area of Science:

• Cardiology and Biomedical Engineering
• Medical Diagnostics

Background:

• Paroxysmal atrial fibrillation (PAF) is a prevalent cardiac arrhythmia.
• Detecting PAF is difficult as episodes are transient and may not occur during monitoring.
• Current diagnostic methods often miss PAF due to its intermittent nature.

Purpose of the Study:

• To identify and evaluate low-level electrocardiogram (ECG) parameters for characterizing PAF patients.
• To develop an automated classification algorithm for PAF diagnosis.
• To enable PAF detection even in the absence of an active arrhythmia episode.

Main Methods:

• Extraction of novel low-level parameters from ECG traces.
• Analysis of the ability of these parameters to differentiate PAF patients.
• Development and evaluation of a modular automatic classification algorithm based on the identified parameters.

Main Results:

• Identification of a set of low-level ECG parameters capable of characterizing PAF patients.
• Demonstration of the utility of these parameters for automated PAF diagnosis.
• Successful development and evaluation of a modular classification algorithm.

Conclusions:

• The proposed low-level ECG parameters offer a promising approach for PAF characterization.
• The developed algorithm facilitates automated PAF diagnosis, improving detection rates.
• This method enhances the ability to diagnose PAF, particularly when episodes are not captured during standard ECG.