Dear Ariana,
> I've been having trouble getting the nonlinear registration to work
> when
> using MELODIC. This problem occurs when I try to use multi-session
> temporal
> concatenation, and only occurs when I turn on the non-linear
> registration
> option. The analysis works fine using the regular registration, but
> I get
> error messages when using non-linear for the same analysis and same
> files.
>
> I understand this is possibly due to me not having the full
> structural files,
> but these are not available. I only have the 4D scans available.
when you say "4D scans", do you by that mean your EPI scans? I am
afraid that you cannot use non-linear registration base only on these.
Primarily because we don't have an EPI template in standard space
(unlike e.g. flirt, fnirt needs a template with similar-ish contrast
to the image you want to register). Even if we were to supply an EPI
template it is often the case that these contain so little/poor
structural information that the non-linear registration would have
little to work with.
So, I am afraid that my best advice to you is to be happy with the
linear registration.
Good Luck Jesper
|