The overall approach sounds sensible, a few things you might want to correct though:
(1) Yes this sounds consistent with others who have had similar data
(2) T2 (1.8s) will be the inversion time —ti=1.8, T1 (0.7s) is the bolus duration —bolus=0.7. There are specific reasons why Q2TIPS uses this terminology and BASIL slightly different, which is confusing.
(3) Yes, BASIL does not do the structural to standard space transformation - so that you can choose how you would like this done and then use the result. You might simply prefer to let oxford_asl run and produce images in native space (and/or structural space) and then do the final transformation yourself using FLIRT (or even better none-linear with FNRIT).
(4) I would be very surprised if you did not have the need for a slicedt value given is it a 2D readout. however, I am not sure what the default slice timing value would be.
Michael
> On 5 Feb 2016, at 19:49, Sara De Simoni <[log in to unmask]> wrote:
>
> Dear FSL users
>
> I have collected Pulsed ASL data (specifically Siemens PICORE Q2TIPS) and was hoping to get some feedback on my analysis pipeline.
>
> I have used FSL BASIL tools to do this. Running the following oxford_asl command works: i.e. produces a calibration perfusion image in both native and structural T1 space (along with the other output files).
>
> oxford_asl -i diffdata.nii.gz -o processed --tis 0.7 –bolus 1.8 --fixbolus --artoff -c stripped_MO_merge.nii.gz –tr 2.5 –s strippedT1.nii.gz --regfrom stripped_MO.nii.gz
>
> My questions are the following:
>
> (1) The PASL sequence acquires 101 images: M0-tag-control-tag-control etc. To calculate the difference images appropriately (using asl_file) I am separating out the first image (M0) from the rest of the file using fslroi.
> Has anyone done this with PASL before and is this a sensible thing to do?
>
> (2) The PASL parameters include inversion times T1 = 0.7s, T2 = 1.8s and saturation stop time = 1.6s. Am I correct in assuming the bolus duration (--bolus) corresponds to the T2 value of 1.8?
> Similarly the --tis (inversion time) corresponds to the T1 value of 0.7s?
>
> (3) If I want to perform complete registration (i.e. into standard space) do I need to add in both the -t (struct2std.mat) and -S (MNI template) options or can I just supply the -S option? I.e. will oxford_asl perform T1 to standard registration automatically if both -s and -S are defined?
>
> In relation to this - does oxford_asl concatenate the registration parameter files and perform registration in one go?
>
> (4) The PASL sequence was acquired in slices (14). It appears from the scanner protocol that there is no slice time delay - is this possible? I.e. there is no need to define --slicedt?
>
> Thanks very much for your help!!
> Sara
>
|