Hi,
If you run first_flirt with the -d flag, it won't clean up the
intermediate registration files.
If it's the first stage that is causing problem, you can use flirt
independently to test the best options. The first stage is just a 12 dof
alignment with skull. Does the subject have a lot of neck in the image? If
so, I suggest cropping the image first.
Cheers,
Brian
> Hi,
>
> Im running FIRST on a group of participants, and for most of them the
> segmentations are very good. However, for one of the participants the
> caudate is not properly estimated (the segmentation is smaller than the
> actual image). I checked the registration and the participant's caudate
> does not overlap well with the MNI brain. I took a look at the first_flirt
> options but Im not sure which option would be most appropriate to fix this
> issue. Could you recommend how the registration could be improved? Should
> I try to use regular flirt and use this transformation as transformation
> matrix when I run run_first_all?
>
> Thank you,
>
> Anastasia
>
>
|