MRI: issue with ico

💬 Support & Discussions
1
  • MNE version: 1.1.1
  • operating system: Windows 10-10.0.19045-SP0
  • python: 3.10.7 (venv)

Dear all,

I computed individual source estimates and morhped them into a fsaverage brain, but I noticed inconsistency in ico parameter:
When I created individual BEM-models, I set ico=4

model = mne.make_bem_model(subject=subject, ico=4,
                           conductivity=[0.3], # for single layer - suitable for MEG data
                           subjects_dir=mris_dir)

After that I created fsaverage using mne.datasets.fetch_fsaverage()
fs_ave = mne.datasets.fetch_fsaverage(subjects_dir=mris_dir)

And get a file ‘fsaverage-ico-5-src.fif’ to use for morphing.
However, I have noticed that in this new file ico=5, which increases a number of sources from around 8000 to 20000.

The morphing went without errors, but I am wondering how this inconsistency can influence the data?
I am also curious if there is a way to set ico=5 in the mne.datasets.fetch_fsaverage() function?

All the best,
Katya

2

@bonokat Hi,

I think you have mixed up with different ico parameters.

First,

model = mne.make_bem_model(subject=subject, ico=4,
                           conductivity=[0.3], # for single layer - suitable for MEG data
                           subjects_dir=mris_dir)

BEM modeling has nothing to do with the source morphing. BEM modelling is typically done to compute the conductivity for different brain layers ( e.g., one-layer model for MEG, or a three-layer model for EEG). ico parameter for BEM model represents the sub-sampling technique (e.g. 5=20484, 4=5120, 3=1280, mesh grid no.) to model each layer.

Second,
Your main question is: why do you see the change of the source vertices from ~8000 to 20000 once you morph the subject stc data to fsaverage atlas?

I believe, you have created your individual subject source space with default spacing: oct6. This will give you ~8k (in total) source locations. Afterwards, you have morphed your subject data to fsaverage atlas (source space with spacing: ico5 ~20k source locations). Therefore, the morphed data is now consist of ~20k source locations/vertices. I don’t see an issue here.

If you want to see the morphing takes place between equal no. of source locations (subject -> fsaverage), you have to create your individual subject source space with spacing: ico5 which will give you ~20k source locations.

Check the following API: (you have an option to change the spacing from oct6 to ico5 here)
https://mne.tools/stable/generated/mne.setup_source_space.html

best,
Dip

2 Likes