Projection of deep sources on anatomical template

Discussions
, ,
1

Hi brainstorm community,

After computing source activation with deep structures, I projected my subjects on the reference template. For some reason I don't catch I know have the mention "| norm" at the end of the name of my projected files. I did not proceed any normalization, so I would like to ask you if this is brainstorm have to do automatically (and in that case what normalization exactly is happening) or if this is just a bug in the renaming of my files.

Best,

Aurelien

2

Norm does not stand for "normalized" but for really "norm".
You have some structures with unconstrained dipoles, the projection computes the norm of the three orientations at each grid point before computing the interpolation to the template brain.

I don't really recommend using these mixed head models for group analysis. The interpolation of most structures to the template is very poor. It is more reliable to work in surface (https://neuroimage.usc.edu/brainstorm/Tutorials/CoregisterSubjects#Using_FreeSurfer_registered_spheres) or volume (https://neuroimage.usc.edu/brainstorm/Tutorials/CoregisterSubjects#Volume_source_models) if the goal is to project everything to the same space.

3

Hi François, thanks for your answer.

Yeah it all make sense now. Hum the thing is I actually only have data in the surface cortex, I only used a mixed model to see if adding deep structures will change a lot my source reconstruction model. So I guess going back to the original individuals subjects surface and project on the templace surface will be better for what I want to do.

Best,

Aurelien

4

In other words: in my mixed model results I only have cortex activation. So

  1. Regarding what you just mentioned, is that an issue. I am just projecting cortex activations on cortex anatomy, the fact that interpolations of most structure is very poor shouldn't play to much for the cortex ones nop ?

  2. If that still an issue; is there another way to just project cortical activation over cortex anatomy

Best,

Aurelien

5

Cortex to cortex interpolation should use the FreeSurfer output and should be OK, even with mixed head models.
Double-check this by placing a breakpoint at line 162 in file tess_interp_tess2tess.m. Run the projection, then execute step by step and make sure it really goes into the block "get freesurfer spheres".

If it doesn't stop at all, delete the interpolation matrices before running your projection (right-click on the cortex surface > Remove interpolations.