Hello everyone,
I am relatively new to nilearn and have a few questions.
-
I did a canICA via nilearn with my resting state data and now i would like to create a mask out of only one component (e.g. the default mode network) but i do not now how to do that properly. Can somebody help?
-
Later on, i would like to combine my mask of the ICA component with regions from an atlas to get the functional connectivity matrices of these regions altogether. Is it possible to create a combined mask out of atlas regions and ica components?
It would be awesome if somebody could help me!
Best,
Maria
Hi,
I did a canICA via nilearn with my resting state data and now i would like to create a mask out of only one component (e.g. the default mode network) but i do not now how to do that properly. Can somebody help?
Since this would be a continuous valued map and unthresholded, I advice to use some form of threshold that component image using nilearn.image.threshold_img
and use nilearn.masking.compute_background_mask
on the thresholded image to get a mask of network that survived this threshold.
Later on, i would like to combine my mask of the ICA component with regions from an atlas to get the functional connectivity matrices of these regions altogether. Is it possible to create a combined mask out of atlas regions and ica components?
What sort of atlas is it? Is it ICA-like or regions that are defined as labels (integers)?
Hope this helps atleast partially
Best wishes
Thanks for your help !
I will try that. But i guess, it would be probably easier to just use an ICA component from an available atlas?
I was thinking of regions which are defined as labels because i wanted to use limbic structures.
I think you may be find this interesting: Dictionaries of multiple dimensions | DiFuMo
This gives you ICA-like maps + the regions (Neuroanatomy nomenclature) that could be found in the limbic systems after searching from a high-dimensional atlas.
Best,
Kamalakar
1 Like
Thanks for the suggestion! I used the atlas now and it is exactly what i have been looking for.
Best,
Maria