Different number of vertices for subjects with warped anatomy

Discussions
1

Hi,

My goal is to compare the difference in power between two groups for all the vertices in the head model (using warped ICBM152 anatomies).

I have made a mixed head model consisting of the default cortex (constrained) and the thalamus (unconstrained). So in anatomy view, each of my subjects have a head model consisting of 16557 vertices (15002 for the cortex and 1555 for the thalamus). Subsequently I have warped each subject's anatomy using the polhemus head points. Then I created a source model (Minimum Norm) and computed the average power in a frequency band. This worked well for each individual, resulting in the figure below (this is one subject with 1 value for every vertex):

image
image822×500 75.2 KB

However, the number of values (when I click view file contents) is different from what I expected; 14475 in stead of 16557. Also, for every subject the number of values varies (14474, 14450, 14480, etc). Am I doing something wrong or do you know what causes the difference in number of values?

From the tutorial I understood that I should not have to project the results back on the default anatomy for the group analysis, because I'm using warped templates. However, because I had a different number of values for each subject, I did try this. This worked well, however the number of values again didn't make sense to me (15386) and I'm not sure how reliable this group average would be?

I hope my question is clear, if not please let me know!

Thanks,
Bart

2

Hi Bart,

These mixed head models are complicated to deal with, and impossible to use the context of a group analysis.

However, the number of values (when I click view file contents) is different from what I expected; 14475 in stead of 16557.

I guess 16557 is the number of the vertices on the surfaces (eg. the surface of the amygdala + the surface of the cortex), and 14475 is the number of source points (the vertices of the amygdala surface are not used as source points, but instead a regular grid of source points is defined inside the volume of the amygdala).

From the tutorial I understood that I should not have to project the results back on the default anatomy for the group analysis, because I'm using warped templates.

This hold only if you are using a surface head model, not a mixed head model.

I recommend you use simpler source models: either a surface head model or volume head model.
Instructions for group analysis with volume head models:
https://neuroimage.usc.edu/brainstorm/Tutorials/CoregisterSubjects#Volume_source_models

3

Hi Francois,

Thank you for your response! I will try the volume head model.

Thanks,
Bart