UC San Diego

Magnetic resonance imaging (MRI) is a powerful and noninvasive tool for measuring brain physiology and functions, including mapping perfusion and oxygen metabolism. Arterial spin labeling (ASL) uses magnetically labeled arterial blood water as an endogenous contrast agent to map regional perfusion. Velocity-selective ASL (VSASL) labels arterial water based on the flow velocity instead of location, eliminating the need for a spatial gap and associated heterogeneous transit delays (TDs) encountered in conventional ASL techniques. However, the velocity-selective (VS) gradients used under the label condition may generate eddy currents (ECs) and result in artificial perfusion signal from erroneously labeled static tissue. To address this, a VS preparation based on symmetric arrangement of eight-segment B1-insensitive rotation (BIR-8) RF and gradient pulses (sym-BIR-8) has been developed to effectively minimize the EC effects across a wide range of EC time constants with robust B0 and B1 insensitivity. Because VSASL uses saturation instead of inversion, its labeling efficiency is compromised. To improve the signal-to-noise ratio (SNR) efficiency of VSASL, a strategy using multiple VSS modules (mm-VSASL) has been developed, resulting in an SNR efficiency improvement of more than 20% compared to conventional VSASL. In the process of developing mm-VSASL, the TDs of VSASL were measured, providing the first direct evidence of VSASL being insensitive to TD effects in contrast to other ASL techniques. In an application of mapping the venous oxygenation in the brain, the VS preparation was modified to excite the moving arterial and venous spins. With arterial nulling, the intravascular signals in venules were extracted for oxygenation measurement, providing information on the oxygen metabolism of brain tissue in situ. In vascular territory imaging, the spins from each supplying artery are labeled uniquely and decoded later to yield the territory information. The planning typically requires prior knowledge of the location of each supplying vessel and may be time consuming or inefficient for detecting collateral flows. To address this, a labeling strategy based on random-encoding the vessels has been developed to label arbitrary number of vessels without planning