Reproducibility of automatic MRI brain tissue segmentation

Introduction Automatic brain tissue segmentation is of great importance to study the aging brain in epidemiological studies, to better understand how diseases affect the brain and to support diagnosis in clinical practice. The following methods are compared: (1) FAST v4.1 (Zhang et al., 2001), part of FSL; (2) SPM5 (Ashburner and Friston, 2005); (3) SPM_segment, an extension to a previous SPM version (Ashburner and Friston, 2000); (4) atlas-based k-nearest neighbor (kNN) classification (Cocosco et al., 2003) using non-rigid registration; and (5) training set based kNN classification (Vrooman et al., 2007).