[MINC-users] motion correction and lid closure

[Andrew Janke]

Hi Marc,

I suspect in order to do this you will have to edit fmr_preprocess
itself to add the masking to the minctracc calls.

I do know of some eye masks that are already in talairach space, but
you are likely best to make your own in mni (approximate talairach)
space that suits your data.  Make the eye mask in Display, then invert
it (mincmath -invert).

Then add your new mask to the minctracc registration calls using -source_mask.


Andrew



On 26/09/05, Marc Schoenwiesner <marc.schoenwiesner at mcgill.ca> wrote:
> Hi,
>
> this is a question about intra-subject intra-modality image registration to
> correct motion within a run of functional images. In their 2002 neuroimage
> paper Stephan et al. showed that closing and opening the eyes may lead to small
> misalignments in the registration, because the intensity of the eye voxels
> changes. Some people recommend to mask out the eyes in functional runs before
> motion correction. How could this be done with tools at the BIC? Zeroing the
> eye voxels in all images may not be the optimal solution, because this creates
> an artificial structural feature which might mislead the registration even
> more... Roger Woods AIR package for instance has a -mask option which allows to
> disregard the eye signal without creating structural features. How could that be
> done in the frm_preprocess step?
>
> Thanks for any pointers!
> Cheers
> Marc
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>


--
Andrew Janke      (a.janke at gmail.com || www.cmr.uq.edu.au/~rotor)
Canada->Montreal                                   Cell: +1 (514) 924 2012