Center for Functional MRI In the Department of Radiology

Arterial Spin Labeling


    At CFMRI, we strive to provide the state-of-the-art arterial spin labeling (ASL) protocols for a quick and robust measure of whole brain cerebral blood flow (CBF).

    As of May 1, 2015, we have updated our existing ASL protocols to ensure that the scan parameters are in compliance with those recommended by the ISMRM Perfusion Study Group1.

    In addition to the standard acquisition parameters, we now provide two additional subsets of parameters optimized for pediatric and geriatric populations.

    The scans for each protocol are also more streamlined, making them easier to execute and less prone to operator error.

    For questions about the ASL protocols, please contact David Shin (


Healthy brain function is critically dependent on well-regulated CBF for the delivery of oxygen and glucose. Regional alterations in CBF have been observed in a wide range of health conditions, including acute and chronic cerebrovascular disease (e.g. stroke, transient ischemic attacks), Alzheimer’s disease, mild cognitive impairment, epilepsy, HIV-related cognitive impairment, multiple sclerosis, depression, schizophrenia, post-traumatic stress disorder, traumatic brain injury, obsessive-compulsive disorder, and vascular dementia. Over the past decade, ASL has emerged as a robust and non-invasive method for acquiring regional CBF maps. Because of its non-invasive nature and ease of use, a growing number of research and clinical sites are now using ASL.

For additional informaiton on ASL, please visit:

We provide both FAIR2 and Pseudocontinuous ASL1 (PCASL) protocols.

1. In addition to the existing CFMRI 2D PCASL, we now support GE 3DASL.
2. The CSF scan is now part of the ASL scan and thus is no longer needed as a separate acquisition (no need to remember to run manual prescan!).
3. Introduced a 10-minute 2D PCASL protocol with background suppression optimized for measurements of whole brain white matter CBF3.
4. Introduced a single-click multi-TI transit delay mapping protocol (PCASL) that provides a whole brain T1 map, transit delay map, and a quantified CBF map in 15 minutes4.
5. Excluding the transit delay protocol, all data acquired from the protocols will be processed by the CBFBIRN, including field map correction and CBF quantification with a wide range of post processing options. Transit delay processing through the CBFBIRN is planned in the near future.
6. After the ASL/CSF scan, a MATLAB script is automatically invoked and the quantified CBF map is displayed on screen to provide immediate feedback on data quality.
Back to Top

The latest ASL protocols are available both on the 3TW and 3TE scanners under fmri3tw/head and fmri3te/Head, respectively. 
They are located at the top of the protocol list when sorted alphabetically (see Figure below).

Note that individual scans under each protocol can be selected and imported into your own protocol.

For each of the protocols selected, users will find a note associated with the ASL scan (see red box in the figure above) that describes how to prescribe the imaging volume

How to see the Protocol Note:
A. Set up a patient and select an ASL protocol
B. Select a desired ASL series in the Work Flow Manger List Window by a single mouse click on the series and then click on Set Up.
C. The note will appear directly on the scan prescription page at the right bottom corner (see Figure below).

Below are the screenshots of the note for each protocol.




Note that the prescription instruction for the FAIR protocol is different from that of the PCASL protocols!
N.B. If you erase the default graphic prescription preloaded with the protocol, make sure to position the first line of the graphic tool from the bottom and add slices toward the top. For ASL, the slices must be acquired from bottom to top for proper quantification. Do not place the first line of the graphic tool at the top and extend slices toward the bottom of the head!

Back to Top

For all protocols other than GE 3DPCASL, the scanner outputs images in GE Pfile format (e.g. P34816.7).

To locate and transfer Pfiles to another server, see example below:

If you have several Pfiles to transfer, it may be easier to use the script that allows you to transfer files in batch.
Example of how the script can be used is shown below.

For 3DPCASL, T1 structural, and field map scans, data are stored in DICOM format.
Refer to this link for instructions on how to move DICOM files from the scanner to another server.

Back to Top

Data acquired from the Transit Delay Mapping protocol can be processed with a MATLAB-based processing program.
The package can be downloaded from here.

For data acquired from all other CFMRI protocols, users are advised to use the CBFBIRN5 for data processing and data storage.
Detailed video tutorials are available on the CBFBIRN project site for data upload, post processing, and reviewing the derived CBF maps.

PCASL is a newer technique that provides several advantages over FAIR. Unless you have an ongoing ASL study using the FAIR protocol, we currently recommend all new users to choose the PCASL protocols.

FAIR and 2D PCASL CBF maps acquired from the same subject are shown below.

Back to Top

Use the table below to assess the pros and cons of each protocol.




Less sensitive to head motion and can be corrected retrospectively

Provides temporal changes in CBF over time, amenable to conventional fMRI and resting state fMRI studies

Temporal filtering can be used to remove physio noise from perfusion signal

More flexible post processing options and CBF quantification methods available via the CBFBIRN

Single scan

System-built data processing path by GE

Higher SNR

Higher spatial resolution

Post labeling delay fixed across slices


Requires 1:30 minutes of prescan

Field mapping is recommended to correct for blurring/signal dropouts in spiral images, which requires additional 1:15 min of scan time

Longer reconstruction time (if field map correction is used on the CBFBIRN)

Post labeling delay varies across slices

Sensitive to head motion - retrospective motion correction is currently not possible

Suffers from through-plane blurring that may overestimate white matter CBF

Not compatible with fMRI studies, i.e. perfusion changes over time are not available

Background suppression is required that reduces perfusion signal, which may also confound CBF quantification


If your subject population (e.g. young children) is prone to head motion, this protocol is recommended.

If reliability of absolute quantification is important, this protocol is recommended.

3DPCASL is becoming more popular and is a great option if scan time is the primary concern and/or ease of use is an important consideration.

3DPCASL and 2D PCASL CBF maps acquired from the same subject are shown below.
 Also shown are corresponding anatomical images.

2D 3D
TR 2500 ms 4300 ms 4632 ms
TE 3.2 ms 3.2 ms 10.5 ms
FOV 24 cm 24 cm 24 cm
# of Spirals 1 (single shot) 1 (single shot) 8
Points per Spiral 4872 4872 512
Image Matrix 64x64x20 64x64x24 128x128x36
Slice Thickness 6 mm 6 mm 4 mm
Brain Coverage 12 cm 14.4 cm 14.4 cm
Temporal Bolus 800 ms 1800 ms 1500 ms
Post Labeling Delay 1000 ms 1800 ms 1525 ms
Background Suppression Optional Optional On
Scan Time 4:50 + 30-sec calibration 4:18 + 30-sec calibration 4:29

Dual echo OptPCASL provides both BOLD and ASL time series simulateneously6,7,8.
If OptPCASL is used for a series of functional scans, the same scan can be used to acquire whole brain CBF in the same time it takes to use regular 2D PCASL but with improved labeling efficiency.
Please contact David Shin if you are interested in using the OptPCASL protocol for fMRI studies.

We are actively involved in the development and testing of advanced ASL techniques. Some examples include velocity selective ASL9 and vascular territory imaging10.
If you are interested in using these techniques for your study, please contact David Shin (


  1. Alsop et al. Recommended implementation of arterial spin-labeled perfusion MRI for clinical applications: A consensus of the ISMRM perfusion study group and the European consortium for ASL in dementia. MRM 73:102-116, 2014.
  2. Wong et al. Quantifying CBF With Pulsed ASL: Technical and Pulse Sequence Factors. JMRI 22:727-731, 2005.
  3. Jia et al. White Matter Perfusion Measurement using Velocity-Selective Arterial Spin Labeling – A Comparison with Pulsed ASL and Pseudo-Continuous ASL. Abstract #2702. ISMRM 2014.
  4. Jia et al. Comparison of Transit Delay Sensitivity between Pseudo-Continuous ASL, Pulsed ASL and Velocity-Selective ASL. Abstract #2705. ISMRM 2014. ISMRM 2014.
  5. Shin et al. The Cerebral Blood Flow Biomedical Informatics Research Network (CBFBIRN) database and analysis pipeline for arterial spin labeing MRI data. Front Neuroinform 7:21, 2012.
  6. Shin et al. Pseudocontinuous Arterial Spin Labeling with Optimized Tagging Efficiency. MRM 68:1135-44, 2012.
  7. Restom et al. Calibrated fMRI in the medial temporal lobe during a memory-encoding task. NeuroImage 40:1495, 2008.
  8. Restom et al. Cerebral Blood Flow and BOLD Responses to a Memory Encoding Task: A Comparison Between Healthy Young and Elderly Adults. NeuroImage 37:430, 2007.
  9. Wong et al. Velocity-selective arterial spin labeling. MRM 55:1334-1341, 2006.
  10. Wong et al. Vessel-Encoded Arterial Spin-Labeling Using Pseudocontinuous Tagging. MRM 58:1086-1091, 2007.
Back to Top