MNE 1.12.1, pupil data no longer being baseline-corrected

[Cathaway]

Description of the problem

In MNE 1.11, when epoching raw eyelink data, both ypos_right and pupil_right channels are baseline-corrected with the specified baseline interval.
In 1.12.1, ypos_right is still interpolated but pupil_right channel is not correction - neither eyetracking channels were baseline-corrected, I made this mistake because mne.preprocessing.eyetracking.convert_units re-centres gaze data but not pupil

Probably has to do with eyelink channels not recognised as data channels? I couldn't locate the problem from source codes so can only guess

Steps to reproduce

baseline_period = (-0.2,0)
trial_duration=(-0.2, 0.8)


events, event_idx = mne.events_from_annotations(raw=el_data, event_id=event_id)
epochs = mne.Epochs(raw=el_data, events=events, event_id=event_idx, tmin=trial_duration[0], tmax=trial_duration[1], 
                        baseline=baseline_period, reject_by_annotation=auto_reject_trial, preload=True, on_missing='warn')

Link to data

No response

Expected results

This one was taken from a different analysis of the same set of data, using MNE 1.11

Actual results

Here's result from 1.12

Additional information

Platform Windows-11-10.0.26200-SP0
Python 3.14.4 | packaged by conda-forge | (main, Apr 8 2026, 02:08:03) [MSC v.1944 64 bit (AMD64)]
Executable C:\Users\tc3n24\AppData\Local\anaconda3\envs\mne\python.exe
CPU Intel(R) Core(TM) Ultra 5 135U (14 cores)
Memory 15.5 GiB
Core

• mne 1.12.1 (latest release)
• numpy 2.4.3 (MKL 2026.0-Product with 12 threads)
• scipy 1.17.1
• matplotlib 3.10.9 (backend=module://matplotlib_inline.backend_inline)

Numerical (optional)

• sklearn 1.8.0
• pandas 3.0.2

• unavailable numba, nibabel, nilearn, dipy, openmeeg, cupy, h5io, h5py

Visualization (optional)

• qtpy 2.4.3 (PyQt6=6.11.0)
• pyqtgraph 0.14.0
• mne-qt-browser 0.7.4
• ipywidgets 8.1.8

• unavailable pyvista, pyvistaqt, vtk, ipympl, trame_client, trame_server, trame_vtk, trame_vuetify

Ecosystem (optional)

• mne-icalabel 0.8.1
• defusedxml 0.7.1

• unavailable mne-bids, mne-nirs, mne-features, mne-connectivity, mne-bids-pipeline, neo, eeglabio, edfio, curryreader, mffpy, pybv, pymef, antio

[Cathaway]

did some digging around and it seems that a recent re-classification of eyetracking channels as non-data was related to this.

I see the reason behind this change but it does break standard eyetracking data preprocessing a bit, since MNE's default baseline correction method is also the one recommended in Mathôt, S., Fabius, J., Van Heusden, E., & Van der Stigchel, S. (2018). Safe and sensible preprocessing and baseline correction of pupil-size data. Behavior Research Methods, 50(1), 94–106. https://doi.org/10.3758/s13428-017-1007-2

it seems manual correction is the way to go with this MNE version

[larsoner]

Yeah that makes sense. We could modify epochs.apply_baseline to add a picks=None, exclude=() kwarg (meaning "data" by backward-compatible default), and you'd be able to pass picks="pupil" to get it to process pupil channels.