28 – Subpopulation Brain Atlases

This chapter describes a mathematical framework to create population-based brain atlases. This emerging field of medical imaging is already uncovering fundamental features of brain structure and function in health and disease. Brain data are so complex and variable that it is essential to rely on brain atlases, templates, and anatomical models in large-scale investigations. Deformable and probabilistic atlases can warehouse population-based data in a common three-dimensional reference frame. They capture anatomic variability using a variety of mathematical approaches. The interest in cortical anatomy, in particular, has motivated specialized approaches to analyze its structure. Dynamic and genetic brain maps, among other new techniques, are beginning to reveal how the brain develops, how diseases progress, and how genes affect complex patterns of brain structure. A key advantage of a population-based brain atlas is that it can be stratified, according to genetic, demographic, or therapeutic criteria, to reflect a more constrained subset of the population. Differences in a diseased population, or one with known genetic risk, can be visualized by reference to a normative standard. Normative atlases based on young normals can store a rich variety of structural and functional data.