A Bayesian time-course model for functional magnetic resonance imaging data

Functional magnetic resonance imaging (fMRI) is a new technique for studyin g the workings of the active human brain. During an fMRI experiment, a sequ ence of magnetic resonance images is acquired while the subject performs sp ecific behavioral tasks. Changes in the measured signal can be used to iden tify and characterize the brain activity resulting from task performance. T he data obtained from an fMRI experiment are a realization of a complex spa tiotemporal process with many sources of variation, both biological and tec hnological. This article describes a nonlinear Bayesian hierarchical model for fMRI data and presents inferential methods that enable investigators to directly target their scientific questions of interest, many of which are inaccessible to current methods. The article describes optimization and pos terior sampling techniques to fit the model, both of which must be applied many thousands of times for a single dataset. The model is used to analyze data from a psychological experiment and to test a specific prediction of a cognitive theory.