Cross-correlation: A Contrast method for studying fetal cardiovascular time series

Congenital cardiac malformations are a serious health concern affecting the development of the fetal cardiovascular system. Identification of a fetus with abnormalities early in pregnancy, 10-20 weeks gestation, could assist in clinical diagnosis and be a great asset in health care. This thesis is a continuation of, and works in concert with, other research efforts conducted at Rochester Institute of Technology (RIT) aimed at characterizing fetal cardiovascular health through applied signal processing. This thesis utilizes cross-correlation as a technique for analyzing the cyclic variations and assessing commonalities of individual parts of fetal umbilical artery time series data. These time series include heart rate variability, peak velocity variability, and mean velocity variability. Here, common frequency components of cross-correlated signals are evident and unique frequency components of an individual signal are attenuated. Specifically, the studies conducted herein contain analysis of two data types. A study of maternal breathing intact versus maternal stop breathing was performed. This study evaluates the cyclic variations associated with fetal umbilical artery data taken during maternal breathing and stop breathing states. Additionally, an analysis of a cross gestational data group was done. The cross gestational study is a cross sectional study containing three gestational ages (10-12, 13-16, 17-20 weeks). This study was done to analyze the commonalities and differences in fetal umbilical artery blood velocity variability data obtained during the different gestational periods. Frequency consistencies and differences were discovered in both studies and are discussed