Volume atlases in subject space

Hi François,

I am creating volume scouts by loading the aseg.mgz (Volume mask or atlas [subject space]) as mentioned in the tutorial
http://neuroimage.usc.edu/brainstorm/Tutorials/TutVolSource#Volume_atlases

However, when performing coherence/TF/etc. analyses using volume scouts (Process options > Select scouts > aseg > Putamen R in this example), I got an error message:

** Scout "Putamen R" spans over multiple regions of the "Source model" atlas **
** Could not calculate the clusters time series **

Do you have any idea? Thank you.

Cheers,
Arnaud

Hi Arnaud, François is away.

I could not reproduce your issue with the tutorial data. It seems your atlas has identified the right Putamen in more than one consecutive region. What does the scout look like when you visualise it? Maybe you could slightly increase its size until it becomes one consecutive region.

Cheers,
Martin

Hi Martin,
Thanks for the quick reply. Yes if I slightly reduce its size (a few vertices) then it works, thanks!
I know how to fix it now, but just to let you know that this error is not systematic across subjects. Quite strange…
Cheers,
Arnaud

1 Like

Arnaud:

The tutorial you are referring to is use a volume source model, but what you have computed is a “mixed head model”.
I would recommend you do not mix these two approaches, as the chances of mistakes are very high.
If you want to use the ASEG atlas and a set of volume scouts, then you a volume head model. If you want to use a mixed head model, use directly the scouts you have in you “Structures” atlas.

Cheers,
Francois

Hi Francois,
Ok, thanks for the answer. But then if I use the “structures” atlas it does not work for the volumes (e.g. Scout “Putamen R” is not included in the source model. If you use this region as a volume, create a volume scout instead). Shoud I create volume scouts manually?
Cheers,
Arnaud

Indeed… I didn’t think of that limitation.
You should be able to create a new “Volume atlas” and add scouts to it. The region created will be better constrained and should be limited to one structure only.

Ok, thanks Francois!