Hi Chris,
The way the images are created changed about a year ago (in public
releases starting with 4010) so the pictures from the manual are no
longer up to date. The problem was that smoothing led to quite weird
T-images because far from the sources both the signal and the error
went to zero but their ratio remained constant. The images that are
generated now are OK and should be used without smoothing. One point
to note is that there were also some changes in the stats to
accommodate this kind of images so they should be analyzed with
up-to-date SPM version.
Best,
Vladimir
On Wed, Jun 8, 2011 at 4:38 PM, Christophe Phillips
<[log in to unmask]> wrote:
> What troubles us is the images output by the source reconstruction, obtained
> with the COH constraint. As you can see in the attached file, activity
> follows closely the cortical surface, while I was expecting something more
> like Fig 37.10 in the manual, i.e. smooth activity in 3D.
> Should we simply smooth in 3D the images obtained so far? or am I missing a
> step when exporting the images?
> Thanks for your help!
>
> Cheers,
> Chris
>
>
|