The current diagnostic system in psychiatry is symptom-based, as described in detail in DSM-III-R. In other branches of medicine the logic of diagnosis is based on knowledge of the normal function of the organ system affected by disease, and direct measures of pathophysiological process play an increasingly important role in clinical practice. New understanding of normal brain function has opened the way for psychiatry to begin to follow this path and develop a brain-based, rather than symptom-based, nosology. Different models of normal brain function, however, suggest quite different theoretical and research approaches. One way in which these models differ is the degree to which they link specific functions (and illnesses) to specific anatomic locations. In a previous paper I discussed the value of adopting a model of brain function that includes important roles for the processes that integrate multiple brain regions into ever-changing functional constellations. In this paper I have proposed the existence of a disease characterized by abnormality in the general physiological mechanism of task-specific regional brain activation. This is not an etiologic diagnosis and therefore presents nothing more than a symptom-based diagnosis about the cause of the disorder. However, I predict that this diagnosis will define a group of patients that is more homogeneous with respect to both clinical course and etiology than do the current symptom-based diagnoses to which these patients would otherwise be assigned. Furthermore, it may be possible to develop new treatments to address this physiological dysfunction.