Comments on the mailing list are a good place to start. If you use
the "new segment" toolbox of SPM8, then the procedure is usually a lot
more robust - particularly for the initial affine registration, which
often goes a bit wrong with the older segmentation. On the negative
side, some of the subcortical GM (striatum etc) may be identified a
bit less accurately, although this could be resolved with better
tissue probability maps.
I've heard occasional reports that VBM analyses appear to show fewer
differences with the newer segmentation in studies of AD, but not seen
anything published. However, it's possible that the extra CSF volume
in subjects with AD led to biases in how the GM was modelled using
SPM5 for these patients. Given that cortical thickness differences
are at the sub-voxel scale, it is difficult to figure out which
segmentation approach is "better".
Best regards,
-John
On 12 January 2011 10:12, Miorelli, Ana <[log in to unmask]> wrote:
> Hi there SPMers,
> I think it is a quite simple question but I have struggled to find a straight forward answer.....
> I am running an analysis on structural MRIs only. I have already segmented all the images with SPM5 but I am now interested in using SPM8. Can anyone advice where to look for the advantages of segmentation on SPM8?
>
> Many thanks
>
> Ana
|