Functional magnetic resonance imaging (fMRI) of neural function has been in existence only as of about 1992 (1-4), and since that time its use for neuroscience research has expanded rapidly.  e basic method has remained the same: Magnetic resonance images of the brain, brainstem, or spinal cord are acquired repeatedly for several minutes to detect changes in the images over time. Whenever neuronal activity in a region changes as part of a cognitive process, sensory stimulus, motor task, and so forth, the appearance of the tissues in that region changes subtly in the magnetic resonance (MR) images and can be detected if the changes occur consistently each time the same function is performed. Clinical uses of fMRI for diagnosis and monitoring have been proposed since the method was Ÿrst developed, and yet this potential remains largely untapped so far for clinical practice, although fMRI studies of many di˜erent neurological disorders have been carried out. Important questions need to be answered. For example, is fMRI reliable and sensitive enough to guide clinical decisions? Are the results worth the time and e˜ort that could be spent on other tests or patients?