PROJECT SUMMARY Progressive supranuclear palsy (PSP) is a rapidly progressive disabling neurodegerative tauopathy with a universally fatal outcome. Thus far no disease-modifying therapy exists. Initially thought to be associated with a single syndrome known as Richardson syndrome, we now know that several clinical syndromes can result from PSP, such as progressive gait freezing, postural instability, parkinsonism, frontal behavioral syndrome, corticobasal syndrome, apraxia of speech, and progressive aphasia. Often multiple and overlapping symptoms exist in the same patient, making clinical diagnosis and prognostication challenging. Currently there are no disease specific biomarkers. Accurate identification of patients is also integral for development of effective tau targeted treatments and understanding the natural history of this condition. In this R03 we propose data driven independent component analysis of 18F fluorodeoxyglucose positron emission tomography (FDG-PET) across the clinical spectrum of PSP with the main aim of identifying independent metabolic patterns and thereby distinct clinical profiles. We hypothesize that these metabolic patterns will correlate with unique clinical profiles. Secondly, this project will explore whether the FDG-PET metabolic patterns generated from Aim 1 can predict PSP pathology and distinguish PSP pathology from other causes of Parkinsonism, especially highly overlapping conditions like corticobasal degeneration and Lewy body disease. We will utilize 2 independent cohorts, a clinical cohort for aim 1 and an autopsy-confirmed cohort for aim 2. Patterns of brain hypometabolism on FDG-PET have the potential of serving as a biomarker for diagnosis and longitudinal monitoring of PSP subtypes. Classification of PSP according to brain metabolic patterns, as opposed to purely clinical symptoms, will overcome the challenge in clinical characterization of patients with multiple or evolving phenotypes and potentially provide longitudinal dynamic data on progression, prognosis and predict neuropathology.