Michael A. Chappell1,2, Bradley J. MacIntosh2,3, Mark W. Woolrich2, Peter Jezzard2, Stephen J. Payne1
1Institute of Biomedical Engineering, University of Oxford, Oxford, United Kingdom; 2FMRIB Centre, University of Oxford, Oxford, United Kingdom; 3Department of Medical Biophysics, University of Toronto, Toronto, Ontario, Canada
The dispersion of the labelled bolus in Arterial Spin Labeling (ASL) has a substantial effect on the quantification of cerebral blood flow. A number of models exist to account for dispersion effects. However, the comparison and validation of these models is difficult using the conventional perfusion signal. Here data from an ASL preparation coupled with a dynamic angiographic readout is used to capture signal whilst blood is still within the arteries. A Goodness-of-fit metric for the different models is compared and models based on a vascular transport function with an asymmetric kernel appear to be most accurate.