Maybe you can use the TOM8 toolbox (https://irc.cchmc.org/software/tom.php) to create your own Tissue Probability Map (TPM). Take a look in Christian Gaser's vbm8 manual on page 24.
Von: SPM (Statistical Parametric Mapping) [mailto:[log in to unmask]] Im Auftrag von John Ashburner
Gesendet: Dienstag, 10. Mai 2011 16:44
An: [log in to unmask]
Betreff: Re: [SPM] inaccurate segmentation of gray matter
I don't have any good suggestions at the moment, other than maybe try Christian Gaser's toolbox to see if this handles the data better. I suspect that it might, as its final segmentation relies less heavily on the tissue probability maps.
On 9 May 2011 10:37, Moeller, C. <[log in to unmask]> wrote:
> Dear SPM experts,
> I am doing VBM analysis on scans of AD patients and healthy controls.
> I just finished the segmentation and found a lot of inaccurate
> segmentations, especially in scans with a lot of atrophy. I attached a
> screenshot of the segmentation for illustration of the problem. It
> seems as if the space between the cortex and the dura is also
> segmented as gray matter. I used 'New Segment' of SPM8.
> Does anyone have an idea why this happens and how I can solve this
> Another issue is, that often the dura around the brain is also
> segmented as gray matter. In the older version of the 'Segment' step
> there was a 'Clean up' option. Does anyone know how to solve this in
> the 'New Segment' step of SPM?
> Thanks in advance.
> Best regards,