Preclinical studies of radiopharmaceuticals for positron emission tomography (PET) are critical for their translation from the radiochemistry laboratory to their use to perform physiologic measurements in humans. Research areas include the development and validation of tracer models that describe the in vivo behavior of the compounds, and dosimetry studies to determine the radiation exposure subjects will receive. ? ? Studies were performed with [F-18]FPWAY, which binds to brain serotonin receptors. Paroxitine, which increases serotonin concentration in the synapse, was administered intravenously to see if it decreases the amount of [F-18]FPWAY binding, in order to determine if the binding level was sensitive to serotonin released by brain cells. If so, it could be used to assess serotonin levels in patients with neuropsychiatric disease. Binding in the brainstem raphe decreased, attributable due to paroxitine-induced increase in serotonin. Cerebral cortex showed increased binding (i.e., serotonin decrease), presumably due to the greater dampening effect of raphe serotonin autoreceptors causing a decrease in serotonin release in the cortex by serotonin neurons that travel from raphe to cortex. Thus, [F-18]FPWAY binding reflected transient changes in synaptic brain serotonin.? ? Imaging characteristics of a new PET scanner, the High Resolution Research Tomograph (HRRT) were studied, specifically a method for reconstructing the images (a motion-compensation OSEM list-mode algorithm for resolution recovery reconstruction). Optimal use of this algorithm requires an assessment of the resolution and noise of the images that are obtained from actual scans. PET studies were performed with several PET radiopharmaceuticals: [F 18]FCWAY, [C-11]raclopride, [C-11]DASB, and [F-18]fluorodopa. Resolution improvement and noise in the curve of tissue radioactivity versus time were measured after each iteration of the algorithm. With four successive iterations, transaxial image resolution improved by a mean of 3.0, 1.7, 1.3, and 1.2 mm respectively; noise was moderate, increasing from 5% to 9%. Therefore, with sufficient image counting statistics, improvement in image resolution can be obtained using this method with moderate increase in noise.