> - Why do we need to 'determine the parameters' first for coregistering the
> anatomy data, and then use the outcoming .mat file for the writing of the
> EPI images? Is there any difference with 'determine and write parameters'?
"Determine the parameters" is the actual coregistration step. In order to
do a rigid-registration between a pair of images, three translations and
three rotations (6 parameters) need to be estimated. Writing the coregistered
images is simply a matter of creating new image files, where each image is
resampled to match some reference.
>
> - I have seen that when using the T1 image template sometime the
> normalization is not fully correct, but with using the filT1 template is
> does a correct normalization? What is the difference between the two, and
> when should we use one or the other?
There is actually very little difference between the templates. The filT1 is
simply a template that I made based on the sequence we use in the FIL. It
actually looks very similar to the official T1 template.
It is very difficult to predict what would cause a spatial normalisation to
go "wrong". Often, better starting estimates can help (search the archives
for the keywords Reorient and Display). Sometimes using skull-stripped
images (and template) can help, or segmenting your T1 first, and matching the
GM to a GM template (as in "optimised VBM", but not really specific to VBM).
Other tunable parameters are possibly disabling the "brain masking", and maybe
increasing the amount of regularisation that is used.
>
> - When I use a different T1, e.g. 3D T1 (SPGR), the EPI images do not
> overlay on the anatomical images in the right way. The only way to get this
> done is normalize the EPI to the EPI template. Would it be the most correct
> method to normalize the T1 to the T1 template and the EPI to the EPI
> template? Any problems occurring with this method? I would think that the
> EPI data needs a slightly other normalization than the T1 data?
EPI images are distorted in the phase-encode direction (usually anterior-
posterior). Therefore, it is not always possible to get a good rigid-
registration between the EPI and anatomical data. If this is the case,
then spatially normalising the EPI according to warps estimated from
an anatomical image is not likely to work too well. There are ways of
trying to fix this though. see e.g.
http://www.fmrib.ox.ac.uk/fsl/fugue/index.html
Best regards,
-John
--
Dr John Ashburner.
Functional Imaging Lab., 12 Queen Square, London WC1N 3BG, UK.
tel: +44 (0)20 78337491 or +44 (0)20 78373611 x4381
fax: +44 (0)20 78131420 http://www.fil.ion.ucl.ac.uk/~john
|