Dear Marcelo,
The B0 correction is run on the FMRI, not the structural, as it is the EPI sequence (used in FMRI and DTI) that suffers from distortions. Without performing this distortion correction the alignment between the functional and structural images will be inaccurate, particularly in the inferior frontal and temporal areas.
Brain extraction is run on the structural data in order to help the registration process (by eliminating the differences in the non-brain structures between the individual subject and the standard template).
Brain extraction is run on the functional data largely as a way of forming a mask, to restrict the statistical calculations, and also to assist in registration (though it is less necessary for this given that the functional images usually have fat-saturation applied and hence show much less signal outside of the brain).
These processes are largely concerned with registration/alignment/localization of the functional and structural images, as you thought.
All the best,
Mark
On 3 Dec 2012, at 19:08, Marcelo Mattar <[log in to unmask]> wrote:
> Hi all,
>
> First of all, I would like to apologize for the incredibly naive question. However, I've been wondering about these questions for quite a while now, so I thought I could ask the experts and eliminate the doubts once for all.
>
> Basically, I would like to know what's the importance of running:
>
> 1) BET on the structural data;
> 2) B0 correction;
> 3) BET on the functional data;
>
> Obviously, I've read the tutorials and user guides, so I know what the procedures do to the data. However, what would be the benefit of removing non-brain parts, or of performing B0 unwarping? I suspect that performing BET and B0 unwarping on the structural data improves registration between functional and anatomical images, but is that all? In other words, is the functional data unaffected by these procedures, the only benefit being the localization of a given effect found in the functional data? And what about performing BET on the functional data (within Feat), does it also improve registration, or is it only used for visualization purposes?
>
> Thanks in advance,
>
> --Marcelo Mattar
>
|