Markov Models for Detection of Ventricular Arrhythmia

Area of Science:

• Biomedical Engineering
• Cardiology
• Signal Processing

Background:

• Portable cardiac monitoring enables real-time analysis of cardiac signals.
• Developing algorithms for real-time detection of dangerous heart rhythms, such as ventricular arrhythmias, is crucial.

Purpose of the Study:

• To present a Markov model-based algorithm for the real-time detection of ventricular tachycardia, ventricular flutter, and ventricular fibrillation.
• To evaluate the algorithm's performance without relying on noise removal or peak annotation.

Main Methods:

• Utilized a Markov model for real-time cardiac signal analysis.
• Developed an algorithm for detecting specific ventricular arrhythmias.
• Evaluated performance on ECG signals from three public databases.

Main Results:

• Achieved an Area Under the Curve (AUC) of 0.96 and F1-score of 0.91 for 5-second ECG signals.
• Achieved an AUC of 0.97 and F1-score of 0.93 for 2-second ECG signals.
• Demonstrated effective detection without pre-processing or peak annotation.

Conclusions:

• The proposed Markov model algorithm accurately detects dangerous heart rhythms in real-time.
• The algorithm shows high performance on short ECG signal segments.
• This approach offers a robust method for arrhythmia detection using portable monitoring devices.