Abstract
Seismocardiography (SCG) is the measured chest surface vibrations resulting from cardiac activity. Although SCG can contain information that correlate with cardiac health, its utility may be limited by lack of understanding of the signal genesis and a variability that can mask subtle SCG changes. The current research utilized medical imaging reconstruction and finite element method (FEM) to simulate SCG by modeling the propagation of myocardial movements to the chest surface. FEM analysis provided a link between myocardial movements and the SCG signals measured at the chest surface and suggested that myocardial movement is a primary source of SCG. Increased understanding of the sources and propagation of SCG may help increase the utility of SCG as a cardiac monitoring tool. To reduce the variability of SCG measured in human subjects, unsupervised machine learning (ML) was implemented to group SCG beats into clusters with minimal intra-cluster heterogeneity. The clustering helped reduce the SCG variability and unveiled consistent relations with the respiratory phases and SCG morphology. This clustering reduced noise in calculating signal features and provided additional useful features. The study also analyzed longitudinal SCG from heart failure (HF) patients in order to predict HF readmission. Here, many time- and frequency-domain SCG features were extracted. Certain features showed good correlations with readmission. Using supervised ML algorithms, high classification accuracies (up to 100%) were achieved suggesting high SCG utility for monitoring HF patients and possibly other heart conditions. Effective monitoring followed by timely intervention can lead to improved patient management and reduced mortality.
Notes
If this is your thesis or dissertation, and want to learn how to access it or for more information about readership statistics, contact us at STARS@ucf.edu
Graduation Date
2020
Semester
Summer
Advisor
Mansy, Hansen
Degree
Doctor of Philosophy (Ph.D.)
College
College of Engineering and Computer Science
Department
Mechanical and Aerospace Engineering
Degree Program
Mechanical Engineering
Format
application/pdf
Identifier
CFE0008593; DP0024269
URL
https://purls.library.ucf.edu/go/DP0024269
Language
English
Release Date
February 2021
Length of Campus-only Access
None
Access Status
Doctoral Dissertation (Open Access)
STARS Citation
Thibbotuwawa Gamage, Peshala, "Seismocardiography - Genesis, and Utilization of Machine Learning for Variability Reduction and Improved Cardiac Health Monitoring" (2020). Electronic Theses and Dissertations, 2020-2023. 622.
https://stars.library.ucf.edu/etd2020/622