Using computerized analysis of neuroimages can potentially enhance sensitivity and specificity in the diagnosis of psychiatric diseases, such as Alzheimer's disease (AD). However, quantitative analysis of functional brain images remains a manual, tedious task. To address this problem, we are building an extended neuropsychiatric image database system to advance the study of brain images of psychiatric disease toward more comprehensive analyses. We report on our progress so far, and show the design and implementation of a suite of DBMS extensions, as well as the interface used by clinicians to compare metabolism results for single and multiple patients in specific brain regions, a task that was previously insurmountable. We are working toward a first in clinical applications of neuroimaging that will bear directly upon patient assessment, diagnosis, and treatment.