1 2 functions, 3 have 4 spatial resolution and sensitivity. Such specifications impact both neurologic and neuro-oncologic diseases. In 5 the former, they allow detecting, quantitating, and tracking small changes of PET signal in minute brain regions 6 such as brain nuclei that have been implicated in many neurologic diseases. In the latter, they improve the 7 accuracy of tumor target volume definition in radiotherapy and surgical resection, thus treatment outcome. 8 The only commercial brain-dedicated PET is the HRRT, a 2 decade old technology that has been discontinued. 9 Therefore, there is compelling need to develop the next generation brain-dedicated PET with ultra-high 10 specifications to improve diagnostics that can institute therapies earlier in the evolution of the disease. 11 This proposal brings together two highly collaborative teams from Weill Cornell Medicine (WCM) and the 12 Institute for Instrumentation in Molecular Imaging (i3M), with an industrial partner, Oncovision, to build an 13 ultra-high performance brain-dedicated PET, UHB-PET. UHB-PET will exhibit: (i) volumetric spatial 14 resolution of ~0.5mm3 across the gantry, that is >4x better than that of the best brain-dedicated PET being 15 developed; (j) effective sensitivity >26x that of the brain PET with the highest spatial resolution being developed. 16 Our intensive experimental and Monte Carlo simulation results prove that our goals are highly achievable, 17 which we will attain as follows: Specific 18 maximize 19 FOV 20 and 21 Quantitative machine 22 learning to accurately determine the 3D position of 511keV 's interaction within the semi-monolithic slab, infer the 23 attenuation-corrected PET without CT scans, and minimize image noise, thus reduce the administered dose), (d) 24 use a preconditioned fixed-point image reconstruction approach to suppress the noise in sub-millimeter size 25 pixels, (e) adopt motion tracking tool we previously developed to correct for inter- and intra- head motion during 26 dynamic PET imaging, and (f) adopt methods from our previous work to accurately image-derive the input function 27 for kinetic modeling. Specific Aim 3: Assessment of image noise, target lesion visibility and quantitative accuracy 28 attained by the scanner in a characteristic set of specific neurology and neuro-oncology human studies. 29 The ultimate goal is a fully operational ultra-high performance dedicated brain PET scanner with accurate 30 quantitative capabilities for diagnosing and monitoring treatment in brain diseases. . Positron Emission Tomography (PET) is a powerful quantitative tool for studying metabolic and biochemical pharmacology, and pathology in living brains. In the past 3 decades, a myriad of brain PET tracers, been developed.In parallel, PET underwent dramatic advancement in technology that enabled much higher Aim 1 : In 2.5 years we will build UHB-PET with trapezoidal-shape (to sensitivity) semi-monolithic LYSO slabs coupled to high-perform...