Analysis of Functional MRI Data

Abstract

A major goal of functional MRI (fMRI) measurements is the localization of the neural correlates of sensory, motor, and cognitive functions. This includes the proper comparison of response amplitudes in brain areas with respect to various experimental conditions. In order to reach solid conclusions about the localization and functional properties of specialized brain areas, a number of analysis steps need to be performed. These analysis steps include preprocessing operations that reduce imaging artifacts, alignment of functional and anatomical data sets, and inferential statistics. In order to efficiently generalize statistical analyses to the population level, homologue brain regions need to be identified across individuals. Furthermore, proper steps need to be performed to correct for massive multiple testing when using univariate (voxel-based) statistical approaches for whole-brain analyses. While more sensitive multivariate pattern analysis tools are increasingly used to analyze fMRI data, univariate statistical approaches still prove very useful to characterize local response properties. As a recent example of univariate functional mapping, the analysis of the intrinsic organization of specialized brain areas at a columnar level will be discussed.KeywordsGeneral Linear ModelOrdinary Little SquareSerial CorrelationDesign MatrixfMRI DataThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.