Project Summary In proton magnetic resonance spectroscopy (MRS) of the human brain, signals arise both upfield (UF) and downfield (DF) from the water resonance. While UF MRS and MR spectroscopic imaging (MRSI) have been extensively studied in humans over the last 30 years, there have been very few downfield studies, and all of them used single voxel spatial localization. Recently, our group developed the first single slice approach for in vivo DF-MRSI at 3T5. Subsequently, I have further implemented the first three-dimensional (3D) DF-MRSI methods in the human brain with whole brain coverage, on both clinical high-field (3T) and research ultra-high- field (7T) MR systems. Currently, there are two significant technical challenges for DF-MRSI, namely (a) the lack of pulse sequences to acquire 3D DF-MRSI with optimum sensitivity in the shortest possible scan time, and (b) specific software for the accurate quantification and visualization of the broad and significantly overlapping DF signals. In addition, the clinical and neuroscience applications of DF-MRSI have yet to be explored. To address these issues, I propose to develop optimized 3D DF-MRSI pulse sequences for both 3T and 7T, and also to develop an open- source software package for improved quantification, analysis, and visualization of DF resonances. DF spectra contain signals from both exchangeable and non-exchangeable protons, and the information content of DF- MRSI may therefore be complementary to chemical exchange saturation transfer (CEST) MRI. In particular, amide-proton transfer (APT) CEST has proven quite successful for the evaluation of human brain tumors; in the R00 phase of this proposal, after establishing normative values and reproducibility, a comparison of the value of 3D DF-MRSI vs. APT-CEST in patients with glioma will be performed. In particular, I will focus on the ability to distinguish recurrent tumor from radiation necrosis in patients treated for high grade glioma; this is an important diagnostic question that directly effects choice of treatment, and which is often difficult to answer using conventional MRI. Developing these novel techniques requires substantial expertise both in MRSI sequence development and in data analysis. This proposal builds upon my unique record in biomedical imaging with new training from a mentoring team of globally recognized experts in the fields of MRSI, clinical multimodal spectroscopic imaging, and development of post-processing and analysis software at the Johns Hopkins University with outstanding career development resources to successfully train me during this Pathway to Independence Award. This project will generate novel tools to study metabolic processes in neurological and neuropathological processes and leverage their potential to advance the understanding of brain tumors, potentially indicating new routes toward improved diagnosis and efficient therapy strategies.