Testing Local Differences in Placental Shapes Using Weighted Fourier Analysis

• Activity Number: 171
• Type: Contributed
• Date/Time: Monday, August 3, 2009 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Jia Cao*+ and Shubing Wang and Ian McKeague
• Companies: Columbia University and Merck & Co., Inc. and Columbia University
• Keywords: Placenta ; weighted Fourier analysis ; random field theory

Abstract

Placenta serves for nutrient and oxygen exchange with the mother's bloodstream to fetus. Currently there is limited knowledge of the relationship between the shape of the placenta and information as birth weight and gender. The statistical testing of between-group local difference in the shape of placenta is a challenging problem. Methods such as false discovery rate generate ad-hoc results since they do not consider the correlated structure of curve data. We apply weighted Fourier analysis (WFA) method which is a combination of Fourier series and kernel smoothing to the placental shape modeling. The method reduces the Gibbs phenomenon, improves the signal to noise ratio, and increases normality of the estimated errors in the WFA-based generalized linear models. Using curvature representations, we carry out the statistical test of local difference in the shape by the random field theory.