ROI based GLM analysis via nistats


I’m learning how to perform fMRI analysis via nistats, and I found the tutorials on the nistats webpage very helpful. But I didn’t find there any tutorials on ROI-based GLM. Can you please suggest something?


1 Like

Ahoi hoi @alado,

thank you very much for the interesting question.

Could you maybe explain what ROI-based GLM refers to in your case?
Restricting the entire analysis to a certain ROI or apply small volume correction?

Cheers, Peer

P.S.: just in case you missed it, nistats was merged into nilearn and is now available through the latter’s glm functionality.


  • First note that Nistats has recently been integrated into Nilearn, so rather use Nilearn.
  • What do you mean with ROi-based analysis ?
    1. A voxel-based analysis restricted to a given ROI ? Then the only thing you have to change is the `mask_img argument that needs to be tuned to display only the ROI you’re interested in.
    2. An analysis on the signal average of a given ROI ? Indeed, this is not showcased in the library and atm you would need to use the low-level functions of the library… We have not identified this as a use case of wide interest so far, but we’re interested in getting feedback on this.

Oops, just saw your response… Sorry for ignoring it !

@bthirion and @PeerHerholz, hanks a lot for your responses!

Indeed, I saw that Nistats will be merged with Nilearn, but I missed the moment when it happened. I’ll switch to Nilearn.

I meant N2. Ok, I’ll try and post my experience here.