Article ID Journal Published Year Pages File Type
6040666 NeuroImage 2007 13 Pages PDF
Abstract
We propose a method for the analysis of functional magnetic-resonance (fMR) data, based on a Bayesian-network representation. Our method identifies multivariate linear/nonlinear voxel-activation pattern differences across groups, which may provide information complementary to that resulting from a general linear model (GLM)-based analysis. In addition, we describe a model-stabilization method based on data resampling, which may be helpful in the presence of small numbers of subjects, or when data are noisy.
Related Topics
Life Sciences Neuroscience Cognitive Neuroscience
Authors
, ,