If you are using oxford_asl then you need to have already performed the tag-control difference (usually using asl_file, but it could be done via FEAT).

Michael

On 13 May 2013, at 21:31, Shan(Hannah) Luo <[log in to unmask]> wrote:

Thanks very much Michael. Is it ok if my input data is feat preprocessed pASL data ( perfusion subtraction turned on)? Or it requires just one volume differenced data ( tag vs. control)? 

On Mon, May 13, 2013 at 11:29 AM, Michael Chappell <[log in to unmask]> wrote:

If you M0 has only one volume (implying that the mean has been taken already of any multiple volumes) then you just need to fool oxford_asl into thinking that you have two volumes and it will discard the first one (I will modify this behaviour in future to avoid confusion). This is easy with fslmerge

fslmerge -t M0 M0 M0

Michael

On 13 May 2013, at 15:53, Shan(Hannah) Luo <[log in to unmask]> wrote:

There's only one volume for M0. 

On Monday, May 13, 2013, Mark Jenkinson wrote:

Hi,

The first problem can be fixed by doing the following in a terminal before running the GUI:

  export FSLDEVDIR=$FSLDIR

The second problem sounds more serious.  How many volumes are in your M0 image?  You can find out by using fslsize and looking at dim4.

All the best,

Mark

On 10 May 2013, at 18:51, Shan(Hannah) Luo <[log in to unmask]> wrote:

Hi Michael,

Thanks for your detailed response. I am also currently analyzing resting state pASL data and when I typed Asl_gui in my terminal, it gave me an error message: 

Error in startup script: can't read "env(FSLDEVDIR)": no such variable

    while executing

"set FSLDEVDIR $env(FSLDEVDIR) "

    (file "/usr/local/fsl/tcl/asl.tcl" line 72)

    invoked from within

"source ${FSLDIR}/tcl/${origname}.tcl"

    (file "/usr/local/fsl/bin/Asl_gui" line 22)

Any ideas what's wrong? I also tried using oxford_asl -i input -s struct -t high2standard.mat -c M0 --tis 0.7,1.8, it gave me this error message" Error: Insufficient volumes in calibration image (to make a mask), please check M0.nii.gz". 

Thanks ahead of time. 

On Fri, May 10, 2013 at 1:54 AM, Michael Chappell <[log in to unmask]> wrote:

If you are trying to quantify resting perfusion from pASL data then the easiest thing to do is use oxford_asl (or the new user interface Asl_gui) - I would recommend that you make sure you are using FSL 5.0.3 for this, since it handles single TI data better.

In essence what you need to do if you want absolute perfusion are three steps (all carried out by the GUI now):
1. Tag-control subtraction (asl_file).
2. Kinetic model inversion (oxford_asl does this)
3. Calibration (estimation of M0a) (oxford_asl or asl_calib)
If you can live with relative perfusion then the first step might be all you want to start with. I think that correction for partial volume effects is a good idea, but this is not widely used as of yet in the community so it can currently be harder to interpret the results since there is less literature to refer to.
Motion correction is generally a good idea and mcflirt can be used as a first step. Motion correction can be problematic with ASL data though.

Michael

On 9 May 2013, at 14:00, Dídac Vidal <[log in to unmask]> wrote:

> Hi everybody!.
>
> I am trying to preprocess fMRI perfusion data with BASIL toolset scripts for first time in a pASL with single TI data and I am getting a little confused.
>
> I think I would use
> a) asl_reg for registration,
> b) asl_file in order to get a CBF quantification,
> and then I woul use c) asl_calib in order to get an absolute quantification of CBF.
>
> My questions are:
> 1) is that correct?
> 2) I am missing something?
> Is PVE recommended /necessary?
> I guess some motion correction should be carried out before asl_file? Probably mcflirt should work fine.
>
> Thanks in advance
>