Date of Award

2004

Document Type

Dissertation

Degree Name

Doctor of Engineering (DEng)

Department

Biological & Agricultural Engineering

First Advisor

Dr. Roger M. Hawk

Abstract

Objective: The objectives were to implement signal-processing techniques to separate the fetal magnetocardiogram (FMCG), as collected by SQUID (Superconducting QUantum Interference Device) gradiometers, from mixtures of maternal and fetal biomagnetic signals recorded during the antepartum period and to develop a database of normal FMCG values of the fetal cardiac time intervals and heart rate variability measures. Materials and Methods: The research was performed in three phases: 1) Independent component analysis (ICA), nonlinear filtering techniques, and orthogonal projection methods were optimized enabling the FMCG to be processed using traditional signal analysis techniques; 2) The cardiac time intervals and heart rate variability measures were computed to create the largest database of normal FMCG recordings performed on a single SQUID system; and 3) FMCG recordings performed on fetuses with cardiac abnormalities were used to investigate the feasibility of performing FMCG recordings for clinically relevant applications. Results: Orthogonal projection proved superior to ICA and nonlinear filtering by enabling the FMCG to be investigated in 98% of the tested datasets without affecting the amplitude of the FMCG. The results of linear regression analysis supported the fact that the cardiac time intervals and heart rate variability measures were dependent on the gestational age of the fetus. FMCGs recorded from fetuses with cardiac abnormalities proved to be successful at monitoring twin pregnancies as well as recordings with irregular fetal cardiac rhythms.

Download

Share

COinS