Dear Armin,
The chances are that this is just showing an extended interpolation
and that the registration is fine. If it is only one slice that is missing
and the "stretched" bit is only about one slice extra then this would
be it. In that case there is little to worry about and I would say that
you can continue with your analysis. However, to be safe it would
help if you could post a picture (you cannot attach it to a list email
so post it on the web somewhere) showing the EPI to atlas registration,
the T1 to atlas registration and the EPI to T1 registration.
All the best,
Mark
On 3 Jan 2012, at 14:43, Armin wrote:
> Dear Mark,
>
>
> For some subjects , the EPI image is not contain the top slices of the brain and when I do registration, FNIRT stretches the brain (in EPI data) to fill the the top of the brain and even beyond that. But registration from T1 to Atlas look good and I donot know what should I do.
>
> I think it doesn't make sense. If the non-linear transform that get in transforming the T1 to Atlas is directly apply to transform the EPI (which is now registered in T1) to Atlas, we shouldn't have this much stretch in the EPI (when we transform the EPI_registered_T1 to atlas)
>
> Bests,
>
>
>
|