Hi,
I don't understand why this is a problem. That cluster call is just used to fill in holes within the fieldmap. What you are describing about the inverse mask (zero for the brain, ones outside) is correct and what I'd expect. If there are no holes then it should quickly move on. Have you done brain extraction on the fieldmap magnitude scan?
Also, what exactly do you see in the log? Are there any error messages or does it all just seem to stop processing?
And have you tried running this with the latest version of FSL?
All the best,
Mark
On 29 Mar 2013, at 18:09, Ben Jarrett <[log in to unmask]> wrote:
> Hi,
>
> Many thanks for your help. Well, I left a little detail out that I shouldn't have. I'm working with mice, so I have to play some "tricks". Very first thing I do is the brain extraction, then:
> 1. I do motion control first, but don't do it within FEAT. I first need to multiply my voxel size by ~20 to make it big enough for MCFLIRT to work properly, then I shrink again before loading everything (fieldmap, motion corrected EPI, T2 image, atlas) into FEAT. We have a script to make a .fsf file to load into FEAT with all the parameters we'd like to run (so we don't have to click on a bunch of different options in the GUI every single time). This file looks fine for my new data.
>
> 2. Since my post I've gone through the report logs (in the FEAT generated folder) of data that has worked and then tried each of the FEAT processing steps in a terminal on an old data set and my new one.
>
> 3. I have a new question, as my data seems to fail at the cluster function; what is the purpose of this function at this step in processing of FEAT?
>
> So:
>
> My new data fails at:
> /usr/share/fsl/4.1/bin/cluster -i FM_UD_fmap_mag_brain_mask_inv -t 0.5 --no_table -o FM_UD_fmap_mag_brain_mask_idx
>
> the output is a matrix full of ones (no brain!).
>
> If I turn on the table output I get:
> Cluster Index Voxels Z-MAX Z-MAX X (vox) Z-MAX Y (vox) Z-MAX Z (vox) Z-COG X (vox) Z-COG Y (vox) Z-COG Z (vox)
> 1 648000 1 89 0 0 44.7 45.1 39.6
>
>
> Now, on my old data set that does work, the FM_UD_fmap_mag_brain_mask_idx has a matrix with values of zero for the brain and ones outside the brain. If I turn on the table output in this case I get:
> Cluster Index Voxels Z-MAX Z-MAX X (vox) Z-MAX Y (vox) Z-MAX Z (vox) Z-COG X (vox) Z-COG Y (vox) Z-COG Z (vox)
> 3 624274 1 0 0 0 44.4 45.2 39.3
> 2 1 1 36 34 37 36 34 37
> 1 1 1 52 26 37 52 26 37
>
>
> not sure what this implies, but I see where it is failing now.
>
> Also, just for reference, here are the steps that seem to work on my new data set (leading up to the failure at the cluster function):
> Prestats
>
> /usr/share/fsl/4.1/bin/fslmaths /group_shares/FAIR_LAB/Analyses/Ben/mouse/Analyses/jarrett18jan2013/processed_data/18jan2013M2Mcoil/18jan2013M2Mcoil_func_1_fslreorient_brain prefiltered_func_data -odt float
> Total original volumes = 450
> Deleting 4 volume(s) - BE WARNED for future analysis!
>
> /usr/share/fsl/4.1/bin/fslroi prefiltered_func_data prefiltered_func_data 4 446
>
> /usr/share/fsl/4.1/bin/fslroi prefiltered_func_data example_func 223 1
>
> /bin/mkdir -p unwarp
>
> /usr/share/fsl/4.1/bin/fslmaths ../example_func EF_D_example_func
>
> /usr/share/fsl/4.1/bin/fslmaths /group_shares/FAIR_LAB/Analyses/Ben/mouse/Analyses/jarrett18jan2013/processed_data/18jan2013M2Mcoil/uphase_rad FM_UD_fmap
>
> /usr/share/fsl/4.1/bin/fslmaths /group_shares/FAIR_LAB/Analyses/Ben/mouse/Analyses/jarrett18jan2013/processed_data/18jan2013M2Mcoil/18jan2013M2Mcoil_MAG_fslreorient_strip FM_UD_fmap_mag
>
> /usr/share/fsl/4.1/bin/fslmaths FM_UD_fmap_mag FM_UD_fmap_mag_brain
>
> /usr/share/fsl/4.1/bin/fslmaths FM_UD_fmap_mag -bin FM_UD_fmap_mag_brain_mask -odt short
>
> /usr/share/fsl/4.1/bin/fslmaths FM_UD_fmap -abs -bin -mas FM_UD_fmap_mag_brain_mask -mul -1 -add 1 -bin FM_UD_fmap_mag_brain_mask_inv
>
>
> Thanks,
> Ben
|