Print

Print
Dear Mingxia,

I would advise you to use the pve output images and threshold these with fslmaths to get the tissue masks.
For example:
   fslmaths basename_pve_0 -thr 0.8 -bin csf_mask

There is no option in FAST for directly creating this output (it just creates the pve outputs).  But it is very easy to run fslmaths as above.

As for separating columns, it does not matter.  Any combination of spaces or tabs is fine (including multiple spaces and/or tabs).

All the best,
Mark




On 9 Apr 2013, at 11:33, zhang mingxia <[log in to unmask]> wrote:

Dear FSL experts,

I am analyzing resting state data. As previous studies did, I planned to use nuisance signals from white matter, cerebrospinal fluid and six motion parameters as the confounding EVs.

I planned to use the script of fast to segment the T1 image of every subject and get the mask, extracting timecourse of white matter and cerebrospinal fluid.

Then I planned to combine the files of timecourse of white matter, cerebrospinal fluid with prefiltered_func_data_mcf.par to be one file as the confouding EVs.

Does anyone know some simpler ways to do it?

When I used the method I mentioned above,I had two basic question: first, "the segmented WM and CSF images were always thresholded to ensure 80% tissue type probability" in previous studies. I didn't find an option of fast to do it. I planned to use -p to get the probability of WM and CSF images and then use fslmaths to threshold them and convert them into binary masks for further analysis. Is that right? I saw binary segmentation of classes automatically created. What's the threshold of them to define as one class?  Second, when I combine different files of nuisance signals,what should be used to delimit columns of different files? TABs? It seems two SPACEs delimiting six columns in prefiltered_func_data_mcf.par.Should i also use two SPACEs to delimit columns of different files?

Any suggestion is appreciated!

Mingxia Zhang