# Converting between montages

**URL:** <https://mne.discourse.group/t/converting-between-montages/6725>\
**Category:** Support & Discussions\
**Tags:** eeg\
**Created:** [April 14, 2023, 8:08pm UTC](https://mne.discourse.group/t/converting-between-montages/6725 "2023-04-14T20:08:27Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![JohnAtl](https://yyz2.discourse-cdn.com/free1/user_avatar/mne.discourse.group/johnatl/32/1223_2.png) [@JohnAtl](https://mne.discourse.group/u/JohnAtl)\
**Post date:** [April 14, 2023, 8:08pm UTC](https://mne.discourse.group/t/converting-between-montages/6725/1 "2023-04-14T20:08:28Z")

</div>

I have several datasets from different clinics that I am processing.  
I also have research datasets that I use.  
These datasets sometimes use different electrode montages, i.e. 10-20, 10-10, 10-05, or something unique, like a 10-20 that has T1 and T2 electrodes added, but no O1 and O2.  
Another example is a research dataset that uses 10-05 (-ish), and does not include Fp1 and Fp2.  
I would like for them all to be usable as 10-20 datasets.  
Is it possible (and advisable) to convert between the systems using interpolation?  
Or perhaps running ICA, then projecting back to a different montage?

---

<div class="post-metadata">

**Author:** ![sappelhoff](https://yyz2.discourse-cdn.com/free1/user_avatar/mne.discourse.group/sappelhoff/32/3276_2.png) [@sappelhoff](https://mne.discourse.group/u/sappelhoff)\
**Post date:** [January 25, 2024, 10:00am UTC](https://mne.discourse.group/t/converting-between-montages/6725/2 "2024-01-25T10:00:43Z")

</div>

> [@JohnAtl](#):
>
> Is it possible (and advisable) to convert between the systems using interpolation?

That depends on how many channels you would have to interpolate. Based on how you describe the situation it sounds like you’d have to interpolate a different amount of different channels for each dataset … it could be possible, but might also be “a bit much”.

Why don’t you try and then write a snippet of a potential “methods section” → writing this section and then reading it will give you a good idea of how reasonable that approach sounds.

Alternatively you could select a small subset of electrodes that is present for all (most) datasets.