Hi Luis,
> I was wondering if someone could help me figure out where the masking
> takes place in a coregistration / normalization process. I coregister a
> full brain image to my functionals (estimate only). The result looks
> great.
>
> I then normalize that coregistered structural image and the resulting
> image is properly normalized, but the pixels that were outside the field
> of view in the functional image (my template for coregistration) are
> zeroed.
I'm guessing that actually the regions zeroes are those outside the
default bounding box for the normalise-write routine. You could set
this to nan(2,3) to get the whole template FOV covered, or if this is
insufficient, then you should specify a larger bb (in mm).
The following script might help you to get an idea of an appropriate
bounding box size, by running it on existing images:
http://www.cs.ucl.ac.uk/staff/gridgway/vbm/world_bb.m
Hope that helps,
Ged.
|