Novel method for superior detection of functional changes in CBF using arterial spin labeling fMRI

The accurate and reliable measurement of cerebral blood flow (CBF) is essential for our understanding of brain metabolism and function. Compared to other imaging methods that can be used to measure CBF, arterial spin labeling (ASL) perfusion MRI offers several important advantages. ASL uses arterial water as an endogenous tracer and as such does not require intravenous administration of costly exogenous contrast agents which may be harmful or uncomfortable to the patient. Furthermore, ASL yields an absolute measurement of CBF in physiologically meaningful units. A unique property of the ASL signal is that it is theoretically insensitive to temporal drift (i.e., 1/f noise). This property of ASL makes the technique potentially well suited for studies tracking CBF changes over time due to recovery or therapeutic intervention. The accuracy of CBF quantification using ASL imaging, however, is confounded by measurement noise, partial volume effects (PVE) and accuracy of assumptions needed to convert measured ASL signal to CBF units, most notably the assumed arterial transit time. The main aims of the research presented in this dissertation were to: (1) minimize the effect of the various confounds on the accuracy of CBF quantification using ASL, (2) optimize the ASL technique to detect local changes in CBF with improved sensitivity, and (3) further investigate and exploit advantages of ASL for the study of brain pathology and function. To improve SNR, we implemented ASL on a 3T scanner. To boost the sensitivity of ASL for detecting focal changes in CBF, a recently developed image analysis method that corrects for PVE was optimized for functional imaging. The benefits of the developed PVEc ASL method are threefold. First, it improved ASL sensitivity for detecting CBF changes due to functional activation and pathology (e.g. stroke). Second, the ASL method allows more accurate quantification of perfusion by accounting for PVEs. Finally, the method allows independent quantification of gray matter (GM) and white matter (WM) perfusion on a voxel basis, and independently from underlying tissue content which normally contributes to intersubject variance in group analysis. To improve the accuracy of CBF quantification, we exploited the ability of the PVE-correction (PVEc) algorithm to isolate the higher SNR GM signal from voxels with significant WM and CSF content and, in turn, obtain voxel-wise maps of GM arterial transit times (δa). Since these maps contain information on regional heterogeneity of δa, they could also be used independently in studies investigating flow dynamics associated with various vascular pathologies (e.g. carotid occlusion). Finally, to investigate the stability and reproducibility of ASL signal, we performed a functional activation experiment in which baseline and task CBF images were acquired one month apart. The ability of ASL to detect CBF changes in longitudinal studies, combined with novel techniques of quantifying CBF independent of tissue content with reduced sensitivity to regional differences in δa, is expected to extend the range of fMRI studies on elderly and patient populations