Ageing is associated with an increased prevalence of neurovascular disease¹. Cerebral Blood Flow (CBF) as measured by Arterial Spin Labeling (ASL)-MRI has shown promising results as a biomarker². However, the arterial transit time (ATT) is known to influence CBF-quantification and should be taken into account for unbiased comparison between groups³. Furthermore, ATT itself is interesting, as it may reflect underlying pathologies such increased vessel tortuosity or occlusion⁴. Currently, no MRI sequence exists to measure ATT in mice. Transit times can be mapped using multi-delay ASL, but these scans are time-consuming⁵. Recently, time-encoded labeling schemes have been implemented in rats and men, enabling ATT-mapping with higher SNR and less scan-time^6-8. In this study, we show that time-encoded pCASL (te-pCASL) enables transit times measurements in old and young mice.

Experiments were performed on a 7T pre-clinical MRI scanner (Bruker Pharmascan) using young (6 months) and old (25 months) wild type (WT) C57BL6 mice (n=6 /group). Two ASL labeling schemes were implemented: a standard pCASL labeling scheme to measure CBF and a te-pCASL labeling scheme optimized for ATT-measurements (Fig.1). Both were followed by a single-shot spin-echo EPI acquisition (TE=17ms, in-plane resolution=224um², slice thickness=1.5mm, 3 slices, slice gap=1mm). The label and control pCASL-interpulse phases were optimized during a pre-scan⁹. Labeling pulses were applied in the neck, 1cm from the isocenter with the following sequence-specific parameters:

- For pCASL, the 3s-labeling pulse was followed by 300ms post-labeling delay (PLD). 60 pairs of label/control images were acquired within 7 minutes (TR=3498ms).

- For te-pCASL, a Hadamard-12 matrix was used as labeling scheme. The duration of every sub-bolus was 50ms and a final PLD of 30ms was added at the end of the scheme. This resulted in 11 effective PLDs (30, 80, 130…530ms). The scheme was repeated 45 times (7 minutes;TR=778ms). For CBF-quantification, T1-maps were acquired with an inversion recovery EPI sequence (18 inversion times) and labeling efficiency was measured 3mm downstream of the labeling plane with a flow-compensated, ASL-encoded FLASH. By decoding the Hadamard-encoded ASL-signal, individual sub-bolus images were created from the te-pCASL measurements (Fig.2). The signal in each voxel was fitted to Buxton’s general kinetic perfusion model¹⁰ after which ATT and CBF were estimated voxel-by-voxel (Fig.3). To calculate quantitative CBF maps from standard pCASL, the same model was used, with an assumed value of 300ms for ATT. The tissue T1 used for both CBF quantifications was taken from the IR experiment. ROIs were manually drawn in six brain regions on the raw EPIs using a mouse brain atlas as reference¹¹.

Example ATT- and CBF-maps from te-pCASL and CBF-map from standard pCASL are shown in Figure 4. The contrasts in the two CBF-maps are comparable. However, as can be seen from the scale bar, the quantitative values were significantly different (p<0.001; paired t-test). This difference in CBF may be explained by intra-vascular signal still present for the shorter PLDs when using te-pCASL, as it was optimized to see the inflow-phase of the label. Future studies should evaluate whether crusher gradients can eliminate such vascular contamination. In all six brain regions the mean ATT was smaller than 300ms, demonstrating that the choice of pCASL PLD (300ms) was appropriate to reduce arterial contamination of the ASL signal and to limit the sensitivity of pCASL to ATT heterogeneity. Compared to each other, the different regions showed different transit times (p<0.001; one-way ANOVA). The hippocampus had the lowest transit time, which can be explained by its short supply route via the posterior communicating artery and longitudinal hippocampal artery¹². Consequently, this difference in ATT could explain the observed increased CBF-contrast between the hippocampus and the auditory/visual cortex in te-pCASL CBF maps compared with standard pCASL (p<0.001; paired t-test). CBF and ATT were found to be preserved in old WT mice (Fig.5) and no differences were found between hemispheres (data not shown). Constant CBF values in ageing WT mice has been reported previously in the literature. In models for neurodegeneration, reduced CBF has been found¹³, and future research should evaluate ATT in these mouse models.This study shows the successful implementation of te-pCASL in mice. For the first time, ATT was measured in mice and as a function of age. te-pCASL and standard pCASL CBF values were however different, necessitating further research.