Darren,
Your message reminded me of a rendering question tha I have. When I extract my brains
I get poor resolution in the occipital area as the dura is just not stripped well. I
assumed this was due to minimal CSF space when subjects are lying down. However, at
HBM 2000, i saw many posters that used SPM to extract brains and they did not have
this problem. Has there been an up-grade in the extraction algorithm that I may have
missed. Your comments would be helpful.
Paul
Darren Gitelman wrote:
> Dear Alejandro:
>
> The templates in spm99 do not include the medial portions of cortex. I have been
> working with the original MNI brains that John Ashburner used to make the current
> templates, but unfortunately I have not yet figured out how to make the natural
> looking surfaces in the current templates. Current settings within spm_xbrain tend
> to still make metallic looking surfaces and not the natural ones included with
> spm99 so I've not made my surfaces available publicly. **JOHN** (sorry to shout)
> By the way the changes you've suggested to use non-normalized surface normals did
> not work out.
>
> You can make your own canonical brains. Take a brain which you feel has normalized
> particularly well and just include those images in the spm/canonical folder for
> all users to access or your own folder for yourself. If you want the brain to
> reflect a group then normalize the group and make an average using the mean
> button. Finally to make templates for rendering you can use spm_xbrain to make the
> extracted surfaces.
>
> HTH,
> Darren
|