Modeling and Optimization of Look-Locker Spin Labeling for Measuring Perfusion and Transit Time Changes in Activation Studies Taking into Account Arterial Blood Volume S.T. Francis, * R. Bowtell, and P.A. Gowland This work describes a new compartmental model with step- wise temporal analysis for a Look-Locker (LL)-flow-sensitive alternating inversion-recovery (FAIR) sequence, which com- bines the FAIR arterial spin labeling (ASL) scheme with a LL echo planar imaging (EPI) measurement, using a multireadout EPI sequence for simultaneous perfusion and T * 2 measure- ments. The new model highlights the importance of accounting for the transit time of blood through the arteriolar compartment, , in the quantification of perfusion. The signal expected is calculated in a step-wise manner to avoid discontinuities be- tween different compartments. The optimal LL-FAIR pulse se- quence timings for the measurement of perfusion with high signal-to-noise ratio (SNR), and high temporal resolution at 1.5, 3, and 7T are presented. LL-FAIR is shown to provide better SNR per unit time compared to standard FAIR. The sequence has been used experimentally for simultaneous monitoring of perfusion, transit time, and T * 2 changes in response to a visual stimulus in four subjects. It was found that perfusion increased by 83  4% on brain activation from a resting state value of 94  13 ml/100 g/min, while T * 2 increased by 3.5  0.5%. Magn Reson Med 59:316 –325, 2008. © 2007 Wiley-Liss, Inc. Key words: arterial spin labeling (ASL); perfusion; Look-Locker echo planar imaging; FAIR; arterial blood volume; stepwise compartmental model (SCM) Accurate quantification of perfusion using arterial spin labeling (ASL) (1–3) requires data to be acquired at a range of postlabeling delays so that transit times can be mea- sured (4 – 8). This results in long acquisition times, which are a particular problem in ASL-functional MRI (fMRI), in which a wide range of postlabeling delays is required because of the change in transit times between resting and active states (5,9). Transit time–independent techniques such as quantitative imaging of perfusion using a single subtraction (QUIPSS) II (10) assume that transit times are similar between different brain regions, activation states, and subjects. The combination of FAIR with Look-Locker (LL) (11) sampling allows measurement of perfusion and transit time changes with high temporal resolution. LL-echo pla- nar imaging (EPI) (which uses a series of low flip-angle, EPI readout modules to monitor the recovery of longitudi- nal magnetization following a single inversion pulse (12)) (Fig. 1) combined with FAIR ASL (2), or LL-FAIR (13) is variously known as LL-EPI-FAIR (14,15), inflow turbo sampling (ITS)-FAIR (16,17), Turbo-TILT (transfer insen- sitive labeling technique) (7) or QUASAR (quantitative STAR labeling of arterial regions) (18). It has been sug- gested that LL-FAIR can measure perfusion with higher SNR per unit time than conventional FAIR (17), as the multiple LL readout pulses sample the inflowing blood more efficiently than the single readout pulse used in FAIR. LL-FAIR has previously been used to measure resting perfusion (17), to classify voxels based on their transit time (7), and to measure changes in perfusion on activation (7). These studies have demonstrated the value of LL-FAIR, but have made various simplifications when fitting the data or optimizing the sequence, most importantly often failing to take account of the effect of the LL readout pulses on the arteriolar blood magnetization in the imaging voxel. Typically when measuring perfusion, bipolar gradients are incorporated into ASL sequences to crush this arteriolar magnetization (19). However, to quantify perfusion accu- rately with LL-FAIR, the suppression of this arteriolar blood magnetization by the LL readout pulses prior to its exchange with the tissue magnetization must be modeled. Here we propose a compartmental model with step-wise temporal analysis of the evolution of the arteriolar blood and tissue magnetization (a stepwise compartmental model, SCM) to take account of the time taken for the labeled blood to reach the imaging slice, the effect of the readout pulses on both the arteriolar blood and the tissue magnetization, and spin history effects of incomplete re- covery of blood and tissue magnetization at short repeti- tion times (TR). This model has been used to optimize LL-FAIR for measurement of perfusion with maximum SNR at high temporal resolution and in the resting state. LL-FAIR has been optimized previously, but only for rest- ing perfusion with long TR values, and using the limited model outlined above (17). When measuring perfusion changes in response to brain activation, accompanying changes in T * 2 due to the BOLD effect can cause errors in the fitted value of perfusion. We have combined LL-FAIR with multiecho EPI readout to address this issue (Fig. 1) (15,20,21). Furthermore, the simultaneous measurement of perfusion and T * 2 changes allows models of the relationship between brain oxygen- ation, blood volume, and blood flow to be tested. Sir Peter Mansfield Magnetic Resonance Centre, School of Physics and Astronomy, University of Nottingham, UK. Grant sponsor: UK MRC and EPRSC Programme; Grant number: G9900259. *Correspondence to: Dr. S.T. Francis, Sir Peter Mansfield Magnetic Reso- nance Centre, School of Physics and Astronomy, University of Nottingham, Nottingham, UK. E-mail: Susan.Francis@nottingham.ac.uk Received 23 April, 2007; revised 21 August 2007; accepted 15 September 2007. DOI 10.1002/mrm.21442 Published online 6 December 2007 in Wiley InterScience (www.interscience. wiley.com). Magnetic Resonance in Medicine 59:316 –325 (2008) © 2007 Wiley-Liss, Inc. 316