Between 0.2-1% of the population is epileptic. Approximately one quarter of epileptics are refractory to medical treatment and most have temporal lobe epilepsy. These patients are candidates for temporal lobectomy. Unfortunately, after surgery, many suffer a loss in naming ability. This is one example where if the sensitivity of a PET activation study were improved so that significant results for a single individual could be attained, management of the patient might be improved. The purpose of this project is to enhance PET activation imaging until this is possible. To date, PET activation studies have been used to form conclusions concerning groups of patients. This work intends to enhance activation studies so that significant conclusions can be drawn on individual patients. This work is an investigation into five areas of the PET activation technique. First, side-by-side tests comparing (15-0)water and (15- 0)butanol will be performed. (15-O)Butanol has been proposed (but not validated) as a potentially superior radiotracer because it is more freely diffusible across the blood brain barrier. The second area is an optimization of the acquisition protocol to maximize the information in a PET study. In this subproject, the relative timing between the task onset and the data acquisition, the duration of the acquisition, and the task presentation will be investigated. Included is an investigation into the effects of habituation. The confrontational naming stimulus task will be used but our intention is to produce a set of general principles for optimizing any activation task. Third, an objective method for identifying activated regions of the brain will be developed. The innovation in the proposed work is to directly manipulate the primary data (2D or 3D sinograms), where the statistical properties are known, to produce images with equal variance in every pixel (Z-images). From the Z-images, the probability that any particular region of the brain has been activated can be obtained. Given an acceptable false positive rate, all significantly activated regions are determined objectively. Fourth, recent PET scanner developments have provided the ability to collect 3D data. With 3D data acquisition, the interplane septa are removed to increase (by an order of magnitude) the overall sensitivity of the scanner but the increase is not uniform across the imaging field of view. Also, the image contrast is reduced because of a greatly increased scatter fraction. The Z-image methods will be used in side-by-side tests of 2D versus 3D PET data acquisition to determine the relative advantages of each. We anticipate that the cumulative improvement from each of these subprojects will produce a PET activation technique that is able to provide significant conclusions concerning individual patients. This work is a proof of principle project. The techniques developed here, specifically for studying epilepsy, are equally applicable to transferring other PET activation research to the clinical setting. Our long term goals are to develop a range of diagnostic activation tests that will allow a physician to treat neuro illnesses based on the neurobiology of their disease rather than relying solely on the presenting symptoms.