Dear Torsten,
We register the fieldmap with the EPI (example_func) as part of the processing
in FEAT, so the difference is already taken into account.
As for differential effects of distortions changing with head movement - this is
a very small effect (since head movement is rarely more than a degree or two)
and so we neglect this in the corrections and simply deal with one fieldmap
and one example_func in order to get the best registration with the structural.
In this context it doesn't really matter how much the fieldmap is rotated or
translated as long as the registration is good.
If you want to see how much movement was required between the fieldmap
and the EPI then you can run (from your feat directory):
rmsdiff unwarp/FM_2_EF.mat $FSLDIR/etc/flirtsch/ident.mat example_func
which will give you the average (RMS) mm displacement.
All the best,
Mark
On 14 Jul 2011, at 08:59, Torsten Ruest wrote:
> Hi there,
>
> in our dataset there is some degree of head movement between tasks that have been acquired within one session. That's obviously also evident in the fieldmap relative to the EPI data to be unwarped. So I was wondering about how much head movement is acceptable so that the fieldmap is still validly applied to unwarp the EPI and how I can assess this quantitatively.
>
> Ideally I'd like to have a similar displacement assessment as for MCFLIRT motion correction of the EPI data (prestats in feat), with a displacement (in mm) between the EPI and the fieldmap data. Is this possible at all, and if so how?
>
> Thanks very much in advance for any suggestions.
>
> Best wishes,
>
> Torsten
>
|