Dear Marco,
FNIRT is always our recommendation unless the image quality is too poor to allow FNIRT to get a good registration. If, for any reason, you cannot get a good result with FNIRT, as assessed by manual inspection, then do not use a bad result, but try and improve it. If it is not possible to get a FNIRT result which is better than the FLIRT one then it would be OK to go with the FLIRT one, but I would be surprised if this were the case with good images. We always use FNIRT in our lab.
All the best,
Mark
On 9 May 2013, at 04:46, Marco <[log in to unmask]>
wrote:
> Dear FSL experts,
>
> Is there a clear consensus on whether FNIRT should be preferred to FLIRT when performing normalization to MNI152 for group analyses of fMRI experiments?
>
> In discussing with various people, I noticed that some have a strong preference for FNIRT (and indeed the FNIRTed images do look very beautifully registered to the MNI space). However, I have also heard others expressing a preference for FLIRT, because they feel less comfortable in having the data more aggressively deformed than with a linear registration.
>
> What are your recommendations? Are there papers on the FNIRT vs FLIRT comparison that you would recommend?
>
> Thanks,
>
> Marco
|