Yep. Here are my commands for multiplying (using fslmaths) and then pulling the data (fslstats). I called my T1 realigned dti_FA file dti_FA_str
Multiplying by realigned dti_FA:
fslmaths NormedAndThresholded_fdt_paths -bin -mul dti_FA_str outputname -odt float
Pulling data (-M for mean FA; -V for volume -> first value is # of voxels, second is volume):
fslstats outputname -M -V
I also used a freesurfer white matter label to only include those voxels within the white matter, so my command was actually this
fslstats outputname -k WMmask -M -V
Hope that helps!
Kristen
-----Original Message-----
From: FSL - FMRIB's Software Library on behalf of Jeremy Young
Sent: Tue 4/14/2009 4:28 PM
To: [log in to unmask]
Subject: Re: [FSL] tracking in T1, and pulling out expected FA values
OK. So what you're saying is to register/flirt dti_FA to T1 space to
do the multiplication. My T1 is higher resolution as well, 1
isotropic vs. 3x3.1x3.1.
Do you remember the exact commands you used Kristen?
Thanks!
Jeremy
On Apr 14, 2009, at 3:36 PM, Lindgren, Kristen, Ann wrote:
> I'm not positive of the argument for doing it this way, but in the
> past I've been advised by this list to use ApplyXFM_gui and
> diff2str.mat to align dti_FA to my T1 image. I think it was
> decided that this was the best way since it would preserve the
> resolution of fdt_paths (my T1 slices are 1mm thick whereas my DTI
> slices are 2mm thick).
>
> Kristen
>
>
> -----Original Message-----
> From: FSL - FMRIB's Software Library on behalf of Jeremy Young
> Sent: Tue 4/14/2009 3:21 PM
> To: [log in to unmask]
> Subject: [FSL] tracking in T1, and pulling out expected FA values
>
> Hi all,
>
> After running probtrackx tracking in T1 space, I want to pull out
> expected FA values by multiplying the dtifit_FA.nii.gz volume with
> the normalized and thresholded path volume using fslmaths. The
> problem is dtifit_FA is in FA space and the path is in T1 space.
>
> What's the appropriate thing to do, flirt the dtifit_FA into T1
> space or flirting the path into FA space?
>
> I've tried it both ways and the #'s come out differently but not
> an order of magnitude differently.
>
> Thanks for your help!
>
> Jeremy
>
>
|