Computing Brain Change over Time

This article begins by introducing the applications and advantages of longitudinal brain imaging data; these include the study of development, structure, plasticity, and disease. Methods for modeling and analysis are described, spanning a range of spatial scales from brain to voxel. After outlining fundamental techniques of volumetry, image registration, and segmentation, more specialized concepts are explored, including direct measurement of change and the use of groupwise or series-wise modeling. The article focuses on volume change, but nonvolumetric measures are briefly discussed. Sections on statistical modeling and on potential confounds and pitfalls cover some of the more subtle but important issues.