PeterKohler Wrote:
-------------------------------------------------------
> Hi Ziad,
>
> thanks for getting back to me. The reasoning
> behind my ROI query was that we load
> functionally-defined, bonified SUMA ROIs into
> Matlab as part of our EEG source localization
> pipeline. We recently started using a separate set
> of anatomically defined ROIs, that are defined not
> as bonafide SUMA ROIs, but as niml.dset file where
> the vertices belonging to each ROI are assigned
> different integers. It would be nice to be able to
> use the same procedure for all of our ROIs, so I
> was looking to convert the dset ROIs to bonafide
> SUMA ROIs. I think I have come up with a
> reasonable solution on the Matlab side, but I
> would be interested in hearing more about how to
> compute surface ROIs (bonafide or not) based on
> thresholded surface maps automatically, rather
> than just drawing them manually.
>
Hi Peter,

Sorry, I almost forgot about this thread... I am not sure I understand correctly here but it seems to me that the answer to the last question would be SurfClust with option -out_roidset. The ROI format in SUMA can be quite complicated if you don't use the 1D format. NIML ROIs contain brush-stroke by brush-stroke information that would be unpleasant to parse outside of SUMA.
Perhaps you're looking for a way to get at node indices that correspond to values in a dataset. If that is the case, to get at all the values you can do this:
ConvertDset -i something.niml.dset -o_1D_stdout
To get all the node indices corresponding to the rows output above:
ConvertDset -i something.niml.dset'' -o_1D_stdout


> The color map thing works great - thanks!
>
> I have another SUMA-question: I have analyzed a
> functional dataset using afni_proc, and used
> SurfAnat+orig from each subject's SUMA folder as
> the anatomical dataset. I make afni_proc align the
> anatomical to the EPIs, rather than the other way
> around, because I want to manipulate my EPIs as
> little as possible. It works fine, everything is
> well-aligned and my data seems to make sense. Now
> I want to map the data onto the surface: Do I need
> to use @SUMA_AlignToExperiment to make a version
> of the SurfAnat dataset that is aligned to my
> experiment, or can I simply use the aligned
> SurfAnat+orig generated by afni_proc.py
> (anat_final)?

In general, if you're not using the -GIFTI option with @SUMA_Make_Spec_FS, then the answer in you need to run @SUMA_AlignToExperiment. But the safest thing to do always is to run suma with the same -spec and -sv options you would use for your 3dVol2Surf etc. Launch afni , have SUMA talk to AFNI and verify the alignment with your data. If all looks swell then you're good to go. Otherwise, you'll need SUMA_AlignToExperiment.



I think the latter is the case,
> because when I run afni and suma with anat_final
> as my surface volume, everything appears to line
> up. Similarly, when I run 3dVol2Surf with
> anat_final as the surface volume, my data appears
> to be mapped onto the surface in a reasonable way.
> Can you confirm that @SUMA_AlignToExperiment is
> not necessary in this specific case?
I guess I should have read to the end :) . If it looks good with the -spec and -sv options you used for SUMA, it will be good for the 3d progs.

cheers
z