For MRI, I would not really recommend using the Normalise button to
spatially normalise. If you use the segment button, then this will
write out some parameters that can be used for spatially normalising
your data. In principle, the "New Segment" toolbox should do a better
job than the algorithm you get from the segment button, but I have to
do a bit more work in terms of integrating the deformation fields that
this produces into SPM.
There's a paper by Arno Klein et al in NeuroImage, which shows how
poorly the Normalise button does compared to other approaches
(including other approaches available in SPM). A paper by Crinion et
al also shows that the Normalise button does not work very well when
compared with the Segment button.
Best regards,
-John
On 4 October 2010 18:46, Asaf Kaftory <[log in to unmask]> wrote:
> Hello dear SPM'ers,
> I have two questions regarding my pre-processing batch file:
>
> 1. Under Normalization, should the "source image" be "Bias corr" (a dependent product of the segmentation phase) ? - I thought i should supply a *.mat parameter file that is the product of the segmentation process, however we don't have an insertion of parameter file option and besides, that *.mat file hasn't been created yet (since this is the batch configuration which we have yet to run + it doesn't appear in the dependencies) ?
>
> 2. Under the anatomic Normalization, what should the "images to write" be ? the same as the source image ("Bias corr") ?
>
> thanks,
> asaf.
>
|