Discriminative Analysis for Image-Based Studies

Abstract. In this paper, we present a methodology for performing statistical analysis for image-based studies of differences between populations and describe our experience applying the technique in several different population comparison experiments. Unlike traditional analysis tools, we consider all features simultaneously, thus accounting for potential correlations between the features. The result of the analysis is a classifier function that can be used for labeling new examples and a map over the original features indicating the degree to which each feature participates in estimating the label for any given example. Our experiments include shape analysis of subcortical structures in schizophrenia, cortical thinning in healthy aging and Alzheimer's disease and comparisons of fMRI activations in response to different visual stimuli.

• Type: Conference
• Year: 2002
• Where: MICCAI
• Authors: Polina Golland, Bruce Fischl, Mona Spiridon, Nancy Kanwisher, Randy L. Buckner, Martha Elizabeth Shenton, Ron Kikinis, Anders M. Dale, W. Eric L. Grimson
• Keywords: Cortical Thinning, Medical Imaging, MICCAI 2002, Statistical Analysis, Traditional Analysis Tools