A different set of spatial normalisation parameters should be estimated
for each subject. Within subject, the images should be registered using
a six parameter rigid body model (works best when there is no image
distortion). Between subject, the different shapes and sizes of the
brains should be modelled - therefore needing more parameters.
Therefore:
Realign all the within subject functional images together.
Coregister the anatomical image to match the first, or a
mean of the functional images.
Estimate a set of spatial normalisation parameters for
each subject (either from the functional or anatomical image
- whatever works best).
Apply the parameters for a subject to all the subject's images.
Best regards,
-John
| I have a group with 6 datasets (datasets with 15
| slices), which I want to normalize en bloc. Shall I
| write all the images of all 6 datasets with the
| normalization parameters determined from the first
| image of the first dataset or shall I determine the
| normalization parameters for each dataset from the
| first image of each dataset? Then I would have separate
| normalisation parameters for each dataset. And what
| normalization parameters should I use for the
| anatomical image anat.img? Should I determine the
| normalisation parameters for anat.img from anat.img and
| then write anat.img or shall I determine the
| normalization parameters for anat.img from the first
| image of the first dataset of the 6?
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|