Automatic ballistocardiogram (BCG) beat detection using a template matching approach

Area of Science:

• Biomedical Engineering
• Cardiovascular Physiology
• Signal Processing

Background:

• Ballistocardiogram (BCG) monitoring offers a non-invasive method for assessing cardiac function.
• Accurate beat detection in BCG signals is challenging due to noise and lack of synchronization with Electrocardiogram (ECG).
• The I-J-K complex in BCG correlates with ventricular contraction and the ECG's QRS complex, making J peak detection critical.

Purpose of the Study:

• To develop and validate a robust beat detection method for BCG signals.
• To achieve accurate J peak extraction from unconstrained cardiac monitoring devices.
• To enable BCG analysis without requiring simultaneous ECG synchronization.

Main Methods:

• A two-stage algorithm based on template matching using a correlation function within a local moving-window.
• Stage 1: Expert-defined BCG template construction based on empirical analysis.
• Stage 2: Correlation analysis to match the template against the BCG signal in a sliding window.

Main Results:

• The developed method achieved high performance in J peak detection.
• Sensitivity of 95.16% was recorded.
• Positive predictivity value reached 94.76%.

Conclusions:

• The proposed template matching method provides accurate and reliable J peak detection in BCG signals.
• This technique is suitable for unconstrained cardiac monitoring, offering a viable alternative to ECG-synchronized methods.
• The findings demonstrate the potential of this algorithm for clinical applications in sleep medicine and beyond.