An understanding of disease processes in the human brain must ultimately be based on a knowledge of the underlying regional hemodynamic, metabolic, and biochemical changes. Although some such information is currently available from various animal models, the conflicting nature of these data often leaves many important questions unanswered and emphasizes the immense difficulty of developing and studying laboratory models of human disease. One obvious alternative is to develop a means by which the hemodynamic, biochemical and metabolic bases of cerebral disease can be safely studied sequentially in humans by using externally detected radiolabeled tracers.