B0655
Title: FDP control in multivariate linear models using the bootstrap
Authors: Samuel Davenport - University of California, San Diego (United States) [presenting]
Abstract: The approach to performing post hoc inference is discussed over multiple contrasts of interest in the multivariate linear model. To do so, the framework of a prior study is extended to work asymptotically and thus provide simultaneous control of the FDP over all subsets of hypotheses. It is shown that the approach is typically more powerful than existing, state-of-the-art, parametric methods. This is illustrated on a real dataset consisting of fMRI data from the Human Connectome project and on a transcriptomic dataset of chronic obstructive pulmonary disease.
