Thank you Jesper. I will process as per your recommendations and will inform you how it turns out. Thank you again for your insight and help.
Nader
Nader Metwalli
Bioengineering Ph.D. Candidate
The Wallace H. Coulter Department of Biomedical Engineering,
Georgia Institute of Technology / Emory University,
Atlanta, GA, U.S.A.
Tel: (404) 712 4439
----- Original Message -----
From: "Jesper Andersson" <[log in to unmask]>
To: [log in to unmask]
Sent: Monday, November 17, 2008 7:42:58 AM GMT -05:00 US/Canada Eastern
Subject: Re: [FSL] FNIRT quick fix
Dear Nader,
>
> Any progress yet on getting FNIRT to register my 5 mm slice
> thickness diffusion data? Thanks,
as we discussed before the problem has to do with the thick slices
you are using. They are actually 7.5mm, rather than 5mm as I think we
thought before.
I would argue strongly against using a "study derived template".
Given your slice thickness the template will inevitably be of quite
poor quality. I would instead recommend you using the supplied
"standard template" by specifying the -T switch.
If you absolutely want to use one of your own scans as a template I
there are two possible ways forward. In the
FA_2_FMRIB58_1mm.cnf
configuration file you need to make sure that --warpres *
final_subsampling >= voxel_dim_of_your_template
In your case the critical dimension is z, where you have a voxel-size
of 7.5mm. One way to fulfill the criterion is to change the line
--warpres=10,10,10
to
--warpres=10,10,15
which means the criterion would be fulfilled. This is the option I
would recommend.
Your other option is to change
--subsamp=8,4,2,2
to
--subsamp=8,4,2,1
which would also mean you fulfill the criterion. I would _NOT_
recommend this.
Also, let me again say that my #1 recommendation for you is to use
the supplied standard FA-template, and not change the configuration
file.
Good Luck Jesper
|