Vital nutrients, accumulated wastes and therapeutic agents are all transported by diffusion in their journey through brain tissues. 3D computational models of the brain that predict species transport have proven helpful in regional analysis of disease and drug delivery. In our group, we have developed computational models using magnetic resonance diffusion tensor imaging (DTI) data sets that account for heterogeneity and anisotropy of transport. To date, we have used these models to predict spatial depositions following brain infusions.

This dissertation describes the development and implementation of advanced slice-accelerated (SMS) MRI methods for imaging blood perfusion and water diffusion in the human brain. Since its introduction in 1977, Echo-Planar Imaging (EPI) paved the way toward a detailed assessment of the structural and functional properties of the human brain. Currently, EPI is one of the most important MRI techniques for neuroscientific studies and clinical applications. Despite its high prevalence in modern medical imaging, EPI still suffers from sub-optimal time efficiency - especially when high isotropic resolutions are required to adequately resolve sophisticated structures as the human brain. The utilization of novel slice-acceleration methods can help to overcome issues related to low temporal efficiency of EPI acquisitions. The aim of the four studies outlining this thesis is to overcome current limitations of EPI by developing methods for slice-accelerated MRI. The first experimental work of this thesis describes the development of a slice-accelerated MRI sequence for dynamic susceptibility contrast imaging. This method for assessing blood perfusion is commonly employed for brain tumor classifications in clinical practice. Following up, the second project of this thesis aims to extend SMS imaging to diffusion MRI at 7 Tesla. Here, a specialized acquisition method was developed employing various methods to overcome problems related to increased energy deposition and strong image distortion. The increased energy depositions for slice-accelerated diffusion MRI are due to specific radiofrequency (RF) excitation pulses. High energy depositions can limit the acquisition speed of SMS imaging, if high slice-acceleration factors are employed. Therefore, the third project of this thesis aimed at developing a specialized RF pulse to reduce the amount of energy deposition. The increased temporal efficiency of SMS imaging can be employed to acquire higher amounts of imaging data for signal averaging and more stable model fits. This is especially true for diffusion MRI measurements, which suffer from intrinsically low signal-to-noise ratios. However, the typically acquired magnitude MRI data introduce a noise bias in diffusion images with low signal-to-noise ratio. Therefore, the last project of this thesis aimed to resolve the pressing issue of noise bias in diffusion MRI. This was achieved by transforming the diffusion magnitude data into a real-valued data representation without noise bias. In combination, the developed methods enable rapid MRI measurements with high temporal efficiency. The diminished noise bias widens the scope of applications of slice- accelerated MRI with high temporal efficiency by enabling true signal averaging and unbiased model fits. Slice-accelerated imaging for the assessment of water diffusion and blood perfusion represents a major step in the field of neuroimaging. It demonstrates that cur- rent limitations regarding temporal efficiency of EPI can be overcome by utilizing modern data acquisition and reconstruction strategies.