Automatic Detection of Aortic Valve Opening Using Seismocardiography in Healthy Individuals

Area of Science:

• Biomedical Engineering
• Cardiovascular Physiology
• Signal Processing

Background:

• Accurate detection of seismocardiogram (SCG) fiducial points is crucial for clinical applications but remains challenging.
• Existing methods often rely on electrocardiogram (ECG) as a reference, limiting their standalone utility.
• Identifying the aortic valve opening (AO) instant is a key fiducial point for SCG analysis.

Purpose of the Study:

• To propose an automated method for detecting aortic valve opening (AO) instants using only the seismocardiogram (SCG) signal.
• To develop a robust technique that does not require an electrocardiogram (ECG) as a reference.
• To evaluate the performance of the proposed method on clean and noisy SCG data.

Main Methods:

• Utilized multiscale wavelet decomposition to analyze SCG signal components.
• Developed a dominant-multiscale-kurtosis and central-frequency criterion for selecting relevant wavelet subbands.
• Employed signal reconstruction with derived weights and Shannon energy for systole envelope construction.
• Applied a Gaussian-derivative-filtering scheme for final AO peak detection.

Main Results:

• The proposed method demonstrated high performance in detecting AO instants from SCG signals.
• Achieved an average sensitivity of 94%, prediction rate of 90%, and detection accuracy of 86%.
• Validated robustness across clean and noisy SCG signals from a combined physiological database.

Conclusions:

• The developed automated method effectively detects aortic valve opening instants from SCG signals independently of ECG.
• This technique offers a promising, non-reference-dependent approach for SCG fiducial point detection.
• The high accuracy suggests significant potential for clinical integration of SCG analysis.