# How to use altases unavailable in MNE to segment ROI?

**URL:** <https://mne.discourse.group/t/how-to-use-altases-unavailable-in-mne-to-segment-roi/8862>\
**Category:** Support & Discussions\
**Created:** [May 28, 2024, 2:58am UTC](https://mne.discourse.group/t/how-to-use-altases-unavailable-in-mne-to-segment-roi/8862 "2024-05-28T02:58:24Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![ScarletVV](https://yyz2.discourse-cdn.com/free1/user_avatar/mne.discourse.group/scarletvv/32/3694_2.png) [@ScarletVV](https://mne.discourse.group/u/ScarletVV)\
**Post date:** [May 28, 2024, 2:58am UTC](https://mne.discourse.group/t/how-to-use-altases-unavailable-in-mne-to-segment-roi/8862/1 "2024-05-28T02:58:24Z")

</div>

I am a student new to EEG, and I have completed source localization with a template MRI([EEG forward operator with a template MRI](https://mne.tools/stable/auto_tutorials/forward/35_eeg_no_mri.html)) and got the source spaces(src) and source estimate(stc).

- My question is how to segment ROI, such as using an atlas(AAL2, Schaefer2018, etc) that is not available in `fsaverage/label` ?

After using `mne.read_labels_from_annot()`, `mne.extract_label_time_course()` I will get a single stc in every ROI if the atlas is available in `fsaverage/label`. As far as I know Nilearn have abundant atlases. Can I use the Nilearn atlases to label vertexes(dipole?) in src so that I can get a single stc in ROI based on AAL2.