PROJECT SUMMARY/ABSTRACT Since its inception nearly five decades ago, magnetic resonance imaging (MRI) has continued to evolve and is far from having reached its ultimate potential. MRI is unquestionably the most complex but also the richest and most versatile imaging method, therefore requiring systematic training. Although inherently quantitative, MRI has been used largely as a qualitative imaging technique practiced by radiologists utilizing predominantly qualitative criteria for establishing a diagnosis or excluding disease. This approach is fraught with problems, its main limi- tation being the subjective nature of the result, i.e. sensitivity to reader experience and judgment. An increasing number of problems in medicine require a quantitative assessment of tissue structure, physiology and function. Moreover, for many diagnostic or staging problems quantification of an observation is not merely a better option but the qualitative approach is entirely unsuited. Examples are measurement of blood flow and perfusion, quan- tification of metabolite concentration by spectroscopic imaging and chemical exchange saturation transfer (CEST), the assessment of non-focal systemic disorders such as degenerative neurologic or metabolic bone disease where a quantitative measurement of some structural or functional parameter has to be made. Over the years the modality has become ever more complex with the ongoing emergence of new methodologies, providing increasingly detailed insight into tissue function and metabolism. Recent years saw the development and inte- gration of advanced machine learning approaches for a variety of tasks including segmentation and computer assisted diagnosis. Successful participation in these developments demands in-depth, modality-specific training to enable future scientists to effectively deploy the myriad of mathematical tools for pulse sequence design and data reconstruction. Translation of new methods from the bench to the clinic is equally important and highlighted as one of NIH’s key priorities. The training process therefore needs to be multidisciplinary, requiring close coop- eration among MR physicists, engineers, computer scientists and physicians in the various subspecialties. Basic science trainees often understand the medical problem incompletely and typically have difficulties in translating abstract technical concepts to the practicing physician. The proposed training program builds on the director’s earlier program and its record in terms of achieved training outcome, showing the large majority of former train- ees having attained academic faculty or senior research positions in industry. The new program builds on this successful formula by proposing to train four predoctoral candidates in MRI physics and engineering, with par- ticular focus on structural, physiologic and functional applications, for a period of two years. Training modalities involve a combination of colloquia, structured teaching and hands-on l...