Hi Thomas,
the thing is that, intra-subject or not, the FA maps
present some distortions that are not in your T1-w
images. So you should definitely go first for 12 DOF
in flirt. If it's not enough (you can check this by
overlaying the thresholded FA map - so that you can
only see white matter - onto the corresponding T1 scan
in fslview), I would recommend going for a non-linear
registration (areg then nreg commands see previous
emails on that topic).
Hope this helps,
Gwenaelle
--- Thomas Doring <[log in to unmask]> a écrit :
> Hallo FA specialists,
> i escorted the mails about the FA registration to T1
> images.
> I am also interrested in doing this and would like
> to know what
> registration parameters are the most convenient. As
> it is an intra-
> subject registration, i guess it would be
> correlation ratio for the cost
> function with 6 DOFs. Unfortunatly it doesn't work
> very well when i have
> the brainmasked FA map (after running DTIfit) as
> input, and my T1 image
> (doesn't matter if brainmasked or not brainmasked)
> as the reference
> image. Is there something what i am missing?
> Thanks for help,
> Best regards,
> Thomas
>
_____________________________________________________________________________
Ne gardez plus qu'une seule adresse mail ! Copiez vos mails vers Yahoo! Mail
|