Hello Gwenaelle,
I have had some of the same issues that Dr. Jou was talking about and I attempted to use the your solution to the first problem (fslmaths -i tfce_corrp_tstat1 -thr 0.95 -bin mymask), however when I use the -i I get the following error:
[son@penfield test_cluster]$ fslmaths -i tbss_vox_p_tstat1 -thr 0.95 -bin mymask
** ERROR (nifti_image_read): failed to find header file for '-i'
** ERROR: nifti_image_open(-i): bad header info
Error: failed to open file -i
Cannot open volume -i for reading!
[son@penfield test_cluster]$
When I do not include the -i (fslmaths tfce_corrp_tstat1 -thr 0.95 -bin mymask) I do not get the error. That said my question is whether or not the -i (-inm <mean> : (-i i ip.c) intensity normalisation (per 3D volume mean)) is necessary?
Thank you!
-----Original Message-----
From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On Behalf Of Gwenaëlle DOUAUD
Sent: Monday, March 08, 2010 11:29 AM
To: [log in to unmask]
Subject: [FSL] Re : [FSL] Coordinates and FA of all signifcant voxels in skeleton
Hi Roger,
> 1. Is there any way to use the cluster program (or
> any other program in
> FSL) to capture both MNI coordinates and corresponding mean
> FA values (for
> patient and control groups) of ALL voxels that are
> significantly different
> between the compared groups?
You can use fslmeants on your mean FA map, using a mask that you can create by thresholding/binarising your significant results, e.g.:
fslmaths -i tfce_corrp_tstat1 -thr 0.95 -bin mymask
fslmeants -i mean_FA -m mymask -o all_voxels_significant.txt --showall
> 2. Can a report be generated listing the
> total count of significantly
> different voxels for each involved white matter structure,
> as identified by
> the voxels’ MNI coordinates and corresponding labels in
> the JHU White-Matter
> Tractography Atlas?
I would then intersect each ROI of the atlas (see previous posts on this) with your mask of significant results, e.g.:
fslmaths mymask -mas uncinate mymask_unc
To get the number of voxels, just do a fslstats -V on each of your new masks.
> 3. Can mean FA also be calculated (for patient and
> control groups) for
> these significantly different voxels grouped by white
> matter structure?
Just repeat the fslmeants command above with the appropriate new masks.
Hope this helps,
Gwenaelle
|