Monday, February 22nd, 2016
11:00am - 12:00pm, in McCormack 1-420

David Degras-Valabregue

DePaul University

Nonparametric modeling and simultaneous inference in fMRI data analysis

Abstract: Since its inception in the early 1990s, function magnetic resonance imaging (fMRI) has enabled fundamental advances in understanding the brain function. This technology has found important applications in neuropsychology, pharmacology, and clinical practice. From a statistical perspective, the analysis of fMRI data poses unique challenges related to the high spatial resolution of brain scans, their low signal/noise ratio, the complex spatio-temporal structure of brain activity, and the variability of brain anatomy across individuals.

A central task of functional neuroimaging is to study brain specialization, that is, to determine which specific brain regions are active during a given cognitive or sensorimotor task. In this talk I will present a new method for estimating and detecting brain activations in fMRI data. This approach overcomes important limitations of mainstream analytic methods by providing a nonparametric, location-specific model of the hemodynamic response and by enabling the detection of brain activations via simultaneous confidence regions. After describing the large-sample properties of the method, I will illustrate its numerical performance using artificial and real fMRI data.