Kevin Koch

Fundamental magnetic resonance (MR) theory assumes the spatial homogeneity of a dominating static magnetic field B = B 0zˆ. When this assumption is violated, a myriad of artifacts and compromising factors are introduced to MR spectra and images. Though in vivo nuclear magnetic resonance (NMR) is one of the most widely used scientific and diagnostic tools in medicine and biology, it remains haunted by the continual and persistant ghost of B0 inhomogeneity. An inclusive list of in vivo NMR applications severely impacted by B0 inhomogeneity could go on ad infinitum. Examples of such applications include neurosurgical utility in functional magnetic resonance imaging (fMRI), cerebral metabolic flux mapping, cerebral diffusion tractography, and abdominal diagnostic imaging. Given this wide impact on in vivo NMR, significant effort has been exerted in developing methods of compensating B0 inhomogeneity. Complicating this task is the sample-specific nature of in vivo B 0 inhomogeneity and its exacerbation with ever increasing B 0 field strengths. State of the art B 0 inhomogeneity compensation is currently at a critical juncture where homogenization demands are overwhelming the outer capabilities of existing technology and methods. This thesis addresses the B 0 inhomogeneity problem in the mammalian brain and presents novel solutions to the homogenization technology stalemate.