An architecture for a brain-image database

The widespread availability of methods for noninvasive assessment of brain structure has enabled researchers to investigate neuroimaging correlates of normal aging, cerebrovascular disease, and other processes; we designate s uch studies as image-based clinical trials (IBCTs). We propose an architect ure for a brain-image database, which integrates image processing and stati stical operators, and thus supports the implementation and analysis of IBCT s. The implementation of this architecture is described and results from th e analysis of image and clinical data from two IBCTs are presented. We expe ct that systems such as this will play a central role in the management and analysis of complex research data sets.