Arterial Spin Labeling (ASL) techniques, commonly used for cerebral perfusion imaging, can also be used for the visualization of intracranial arteries. These techniques have been modified in recent years to perform selective imaging of individual brain feeding arteries, which are mostly the internal carotid arteries (ICA) as well as the vertebral arteries (VAs). In general, ASL relies on two acquisitions, namely a “label” (inversion) and a “control” (without inversion) image, which are subsequently subtracted to remove the static background tissue. Usually, multiple corresponding label and control images are required, which can prolong the scan time and make ASL prone to motion artifacts. In this study, super-selective pCASL was used to tag individual arteries of interest [1]. In a first approach, only one non-selective control acquisition is performed to be shared by three label images from the ICAs and the VAs (“cycled super-selective ASL”). Similar approaches for perfusion territory imaging with pulsed ASL methods have already been proposed [2]. In addition, in this study, a new approach is explored taking advantage of the fact that the blood spin magnetization of contralateral (non-tagged) arteries is in a different (pseudo-randomized) state in superselective ASL. We hypothesize that this pseudo-randomized magnetization state can be used as an intrinsic control condition for subsequent image processing without the need of acquiring a control image (“self-control ASL”).Thirteen healthy volunteers (8 females, mean age 28.2 years) underwent MR scanning under the general protocol for sequence development, approved by the local ethics committee. Imaging was performed on a Philips 3T Achieva (Philips, Best, The Netherlands) scanner using a standard 32 channel SENSE Head coil. Super-selective ASL parameters: were: 400ms labeling and 50ms labeling delay. Image acquisition consisted of a 3D T1-TFE readout with 10° Flip Angle, 0.9*0.9*0.9mm³ voxel size, 120 slices and 6 acquisition time-points with a temporal resolution of 150ms after labeling. The calculations of the final angiographic images were performed using Matlab R2013a (The Mathworks, Natick, MA). Selective labeling was achieved using extra gradient moments in the Gx and Gy direction of 1.08 mT/m [1]. To selectively label the arteries, a low-resolution time-of-flight scan was performed covering the neck. The location of the labeling focus was planned manually. In cycled super-selective ASL and self-control ASL, the labeling focus positions for each of the labeled arteries were entered before the scan. During scanning the individual labeling focus positions for each vessel were cycled periodically so that all data acquisition was performed in a single measurement. Total acquisition times were 5:08 min for non-selective ASL, 15:24 min for super-selective ASL (three individual acquisitions of 5:08min), 10:17 min for cycled super-selective ASL with only a single non-selective control condition, and 7:43 min for self-control ASL, for which no control condition was acquired at all. The resulting images were compared with non-selective ASL and conventional super-selective ASL in terms of signal-to-noise ratio (SNR) measured at specific locations in each artery.Image acquisition was successfully performed in all volunteers (Fig. 1). All measurements present similar SNR compared with non-selective ASL angiography (not shown). In super-selective ASL, the additionally employed gradients perpendicular to the blood flow direction cause the magnetization in the label and control experiments to oscillate in antiphase, resulting in fluctuations of the labeling efficiency outside the labeling focus [1]. The magnetization in a non-selective control experiment does not perfectly correspond to the super-selective label experiment, thus, might result in decreased image quality. However, the acquisition of a single non-selective control image (without additional transversal gradients) appears sufficient to process super-selectively labeled images of different arteries. Furthermore, it was also possible to only use the label images for the generation of angiograms as the magnetization in contralateral arteries can be used as intrinsic control condition. Compared with non-selective ASL acquisitions, some selective ASL methods increase the total imaging time with respect to the number of tagged arteries. This can also increase the appearance of potential artifacts, e.g. when a patient moves in-between different scans or during the image acquisitions. For super-selective ASL, the presented approaches demonstrate a promising alternative for scan time reduction without the loss of image quality, especially, when the aim is to label several small intracranial arteries.Super-selective ASL can be performed using a shared control condition as well as without acquiring a control image and without notable loss of SNR. This can accelerate the total image acquisition time compared with conventional super-selective ASL.