This should work OK, providing you do the coregistration before
you estimate the spatial normalisation parameters. In this case,
if you do it the other way around, then you will not get what you
expected. It also means that the _sn3d.mat file may no longer work
to normalise your T1 mri.
This is because the orientation of the "image to determine paramaters
from" is stored in the _sn3d.mat. When the sn3d.mat file is applied
to another image, it works out the images orientation relative to the
one stored.
If you change the mat file of the "image to determine paramaters
from" after the sn3d.mat file has been created, then it will no
longer work properly.
This is all just an educated guess about what may have gone wrong,
but its the best explanation I can think of.
Best regards,
-John
| I am running into problems normalizing dynamic EPI scans using
| normalization
| parameters obtained from morfological T1 scan - result dynamic image
| seems to be
| rotated in opposite direction than coregistered morfology is rotated
| compared to
| un-coregistered morfology.
| I have coregistered morfological scan to the mean dynamic scan first.
| All articals I have found in this mail list seem that it should work.
| What is wrong?
|
| I have T1 morfological scan 256x256x56 voxels, 1.11x1.11x3 mm voxel size
| and 60 EPI dynamic scans 128x128x17 voxels, 2.21x2.21x6 mm voxel size.
|
| Steps I do are:
| coregister:
| 1 subject
| coregister only
| modality of first target - EPI
| modality of first object - T1 MRI
| target image - meanNEBUZ001
| object image - morf-3
| other images - none
|
| coregister step produces morf-3.mat file
| When I use <ChechReg>, it seems coregistration is done well
|
| normalize:
| 1 subject
| determine parameters only
| image to determine paramaters from - morf-3
| template - T1
|
| normalize:
| write normalized only
| 1 subject
| parameters - morf-3_sn3d.mat
| object image - meanNBUZ001
| bilinear interpolation
|
| nmeanNEBUZ001 is wrong (funny rotated)!
|
| If you are interested in data I have used, look at anonymous ftp:
| ftp://www.neuro.lf1.cuni.cz/pub/download/spm/
| There are raw data - mean dynamic image and morfology and
| results in "result" directory.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|