Date of Award

5-2026

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Biomedical Engineering and Sciences

First Advisor

Peshala Priyadarshana Thibbotuwawa Gamage

Second Advisor

Christopher A. Bashur

Third Advisor

Mehmet Kaya

Fourth Advisor

Linxia Gu

Abstract

Cardiovascular diseases affect millions of people worldwide, emphasizing the need for accessible and noninvasive monitoring methods. Cardiac mechanical activity generates vibrations that propagate through the body and can be measured using wearable sensors. Seismocardiography (SCG) captures chest wall vibrations, while carotid vibration signals reflect arterial pulse propagation. Understanding the relationship between these signals can provide valuable insights into cardiovascular dynamics.

This study investigates the morphological and temporal relationship between SCG signals recorded at the sternum and carotid vibration signals acquired using a neck-mounted inertial measurement unit (IMU). Data were collected from twelve healthy subjects using wearable Delsys sensors, with electrocardiogram (ECG) signals used as a temporal reference for segmenting cardiac cycles.

A signal processing pipeline was developed to preprocess signals, detect ECG R-peaks, segment individual cardiac cycles, and remove noisy beats. Representative waveform templates were generated using a Dynamic Time Warping (DTW)-based medoid approach. Fiducial points corresponding to key cardiac mechanical events were identified in SCG signals and compared with features in carotid vibration waveforms.

The results demonstrate that carotid vibration signals consistently capture cardiac mechanical activity and exhibit a stable temporal relationship with SCG signals. Stronger correlations were observed during later cardiac phases, while early phases showed greater variability. Propagation delays between SCG and carotid features reflect pulse transit dynamics from the heart to the carotid artery.

These findings indicate that carotid sensing provides meaningful physiological information and shows strong potential as a complementary, wearable approach for noninvasive cardiovascular monitoring.

Available for download on Tuesday, May 09, 2028

Share

COinS