Hi,
I would think that using fnirt (the non-linear registration) would give you even better results than first_flirt. A good way to check visually your data is to concatenate them and load them in fslview where you can activate the movie loop and see how the edges are aligned.
If you're dealing with patients that might be atrophic or elderly subjects, you might consider using a VBM approach (grey matter onto grey matter) and apply the warpfields generated this way to your coregistered PET images.
Cheers,
Gwenaëlle
--- En date de : Mer 18.3.09, Buyean Lee <[log in to unmask]> a écrit :
> De: Buyean Lee <[log in to unmask]>
> Objet: [FSL] Better spatial normalization of subcortial regions with FSL FIRST?
> À: [log in to unmask]
> Date: Mercredi 18 Mars 2009, 19h35
> Hi FSL users,
>
> I am a PET researcher and use a dopamine D2/3 receptor
> ligand whose binding sites are mainly in the subcortical
> areas.
>
> I wonder anyone knows if one can better spatially normalize
> MRIs (in terms of the subcortical areas) if one uses FSL
> FIRST (specifically, first_flirt), which normalize a MRI
> twice.
>
> In relation to this, I would like to know how one can
> quantitatively evaluate if the normalized images
> (specifically, the subcorital areas) are closely aligned to
> each other (I am only able to do this visually by diplaying
> the images like a movie).
>
> Thank you,
>
> Buyean Lee
|