A0677
Title: Multivariate inference for effect size images
Authors: Simon Vandekar - Vanderbilt University (United States) [presenting]
Abstract: Inference in neuroimaging research has focused on hypothesis testing rather than estimation of effect sizes. As a solution, colleagues in biostatistics have developed procedures to construct spatial confidence sets for images that can be used to identify regions with target effect sizes above a given threshold with a specified probability. These confidence sets represent a paradigm shift in group-level inference for neuroimaging data. However, there is no generalized approach to estimating and constructing confidence regions on a unitless scale. This limits the general applicability of the approach. Nonparametric bootstrapping and recently developed approaches are used to construct confidence sets from simultaneous confidence intervals to establish a confidence set procedure for effect sizes of arbitrary model parameters. Their finite sample is evaluated, and the methods are used to identify regions associated with age and diagnostic differences in two datasets studying autism and psychosis.
