Calibration for Ultra High-Dimensional Measurement Error Models with Application to Diffusion Tensor Imaging Tractography Data

• Activity Number: 217
• Type: Invited
• Date/Time: Monday, July 30, 2012 : 2:00 PM to 3:50 PM
• Sponsor: Section on Nonparametric Statistics
• Author(s): Ciprian Crainiceanu*+
• Companies: The Johns Hopkins University

Abstract

An increasing number of studies record replicated measurements of ultra high dimensional data, such as brain images. Thus, each such measurement can be viewed as a proxy of a true unobserved ultra high dimensional measurement. In this talk I will discuss the estimation of the unobserved true underlying signal and its association with outcomes. There is a direct connection between this framework and the measurement error framework; the major difference is that in the case considered here both measurements and true signals are ultra high dimensional. Methods are motivated by a brain imaging study where diffusion tensor images of the brain are recorded at multiple visits in multiple sclerosis patients and healthy controls.