Detection of premature ventricular contractions using the RR-interval signal: a simple algorithm for mobile devices

Area of Science:

• Cardiology
• Biomedical Engineering
• Signal Processing

Background:

• Premature ventricular contractions (PVCs) are common cardiac abnormalities.
• Traditional PVC detection relies on lengthy electrocardiogram (ECG) recordings.
• A mobile device-based approach using heart rate sensors is desirable.

Purpose of the Study:

• Develop an algorithm for PVC detection.
• Utilize RR-interval data from ECGs or mobile heart rate signals.
• Enable simpler and more accessible cardiac monitoring.

Main Methods:

• Extracted timing interval features: prematurity and compensatory pause.
• Applied a linear classifier for beat classification.
• Validated using the MIT-BIH Arrhythmia Database and ROC analysis for unbalanced classes.

Main Results:

• Achieved 90.13% sensitivity and 82.52% specificity.
• Obtained an Area Under the ROC Curve (AUC) of 0.928.
• Demonstrated effective PVC detection with RR-interval data.

Conclusions:

• The developed method efficiently detects PVCs using only RR-interval signals.
• This approach offers an advantage over complex ECG-based methods.
• Results are comparable to more sophisticated techniques, suggesting potential for mobile health applications.