topomap plots of dual- or triaxial OPM data

[laurabpaulsen]

Describe the new feature or enhancement

Plotting of topomaps fail for dual- or triaxial OPM systems because multiple channels share locations but differ in orientation. This holds for many plotting functions, such as Evoked.plot_joint() and ICA.plot_components().

Describe your proposed implementation

A potential solution would be to handle OPM axes analogously to how magnetometers and gradiometers are handled, where a plot is generated for each channel type: automatically group channels by orientation (e.g., radial vs tangential) and generate separate topomaps.

Describe possible alternatives

The grouping (tangential/radial) could either be specified by the user or automatically inferred from sensor orientation vectors relative to a fitted head sphere

Additional context

No response

[PragnyaKhandelwal]

This is a great point, @laurabpaulsen. It feels like a pretty deep architectural shift since the topomap engine would need to learn how to split these axes automatically @larsoner, given the 1.12 push is happening right now, should this be there for 1.13 roadmap instead?

[larsoner]

Yeah it would be great to have something like this! It'll require a bit of work, see #13144 and #11579 for some related discussions

[PragnyaKhandelwal]

Thanks for the pointers, @larsoner. I’ll dig into those issues and the coordinate math once the current release rush is over. Definitely feels like a bigger project to plan out once things settle down.