Abstract

Sudden cardiac death(SCD) is a significant cause of death that accounts for more than 180,000 deaths in the US and 4 million deaths worldwide annually. SCDs are mainly caused by irregular heartbeats called arrhythmias, which are caused by abnormal electrical activity within the heart. Precision medicine, in the form of personalized computational cardiac models of a patient’s heart, can suggest optimal therapies for cardiac arrhythmias. Prior work has used finite element meshes derived from cardiac MRIs to simulate cardiac electrical activity. In this study, I sought to augment this approach by developing a neural network to learn the parameters that describe patient-specific cellular activity. The learned parameters can be combined with a 3-D cardiac mesh to achieve more accurate simulations. I found that all parameters except one could be predicted with low error using the current approach. Thus, there exists a parameter subset that is well-predicted and could be used for initial approaches to patient-specific modeling.

Committee Chair

Jonathan Silva

Committee Members

Michael Brent Neal Patwari

Degree

Master of Science (MS)

Author's Department

Computer Science & Engineering

Author's School

McKelvey School of Engineering

Document Type

Thesis

Date of Award

Spring 5-20-2022

Language

English (en)

Author's ORCID

https://orcid.org/0000-0001-9405-5624

Share

COinS