An Algorithm for Heart Rate Extraction From Acoustic Recordings at the Neck.

Area of Science:

• Biomedical Engineering
• Cardiovascular Physiology
• Wearable Technology

Background:

• Traditional electrocardiogram (ECG) methods for heart rate monitoring are complex for home use.
• Wearable devices offer potential for continuous, unobtrusive physiological monitoring.
• Acoustic signals from the suprasternal notch contain cardiac and respiratory information but are prone to artifacts.

Purpose of the Study:

• To develop a novel algorithm for accurate heart rate extraction from wearable acoustic recordings.
• To address challenges of noise and artifacts in wearable cardiac sound monitoring.
• To enable feasible long-term, at-home cardiac condition assessment.

Main Methods:

• A small, wearable device placed on the suprasternal notch records acoustic signals.
• The algorithm constructs the Hilbert energy envelope to analyze instantaneous signal characteristics.
• Cardiac cycles are segmented and classified into S1 and S2 sounds based on timing.

Main Results:

• The algorithm achieved 94.34% accuracy in heart rate extraction.
• Root-mean-square error was 3.96 bpm, indicating high precision.
• A correlation coefficient of 0.93 was observed against a commercial monitoring device.

Conclusions:

• The proposed algorithm reliably extracts heart rate from wearable acoustic recordings.
• This method offers a feasible solution for long-term, at-home cardiac monitoring.
• The technology demonstrates potential for improved patient self-management and clinical assessment.