# How to define spacing when computing source space?

**URL:** <https://mne.discourse.group/t/how-to-define-spacing-when-computing-source-space/10738>\
**Category:** Support & Discussions\
**Created:** [January 24, 2025, 12:27pm UTC](https://mne.discourse.group/t/how-to-define-spacing-when-computing-source-space/10738 "2025-01-24T12:27:05Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![Jess](https://avatars.discourse-cdn.com/v4/letter/j/2bfe46/32.png) [@Jess](https://mne.discourse.group/u/Jess)\
**Post date:** [January 24, 2025, 12:27pm UTC](https://mne.discourse.group/t/how-to-define-spacing-when-computing-source-space/10738/1 "2025-01-24T12:27:05Z")

</div>

MNE version: 1.4.2  
Operating system: Ubuntu 22.04.3

I have been assigned a task to compute LCMV beamforming on mixed sources. However, I cannot understand which type of spacing to choose when computing source space.

src = mne.setup\_source\_space(subject, spacing=“ico5”, add\_dist=False, subjects\_dir=subjects\_dir)

My supervisor and me worked on a project where we always used spacing=oct6 with LCMV beamformer. But the script for mixed sources that I have been given uses oct5. And my colleagues who worked with the same data but differently used ico5. While the MNE website recommends oct6 for mixed sources ([Compute MNE inverse solution on evoked data with a mixed source space — MNE 1.9.0 documentation](https://mne.tools/stable/auto_examples/inverse/mixed_source_space_inverse.html)).

I am a bit confused, I know there is information on the MNE website but it does not make it clearer for me which spacing I should use.

Thanks!

---

<div class="post-metadata">

**Author:** ![wmvanvliet](https://yyz2.discourse-cdn.com/free1/user_avatar/mne.discourse.group/wmvanvliet/32/100_2.png) [@wmvanvliet](https://mne.discourse.group/u/wmvanvliet)\
**Post date:** [January 28, 2025, 3:57pm UTC](https://mne.discourse.group/t/how-to-define-spacing-when-computing-source-space/10738/2 "2025-01-28T15:57:35Z")

</div>

As a general rule, I think you want at least 4000 source points (counted across both hemispheres), but preferably a bit more. Sometimes there is a good reason to use very few points, like when you are writing an example script that just needs to run fast but is not meant to do a serious analysis.