Baseline correction of misc channel

💬 Support & Discussions
,
1

:question: If you have a question or issue with MNE-Python, please include the following info:

  • MNE version: 1.5.0
  • operating systemWindows 10

Dear colliges,
i have faced a problem with baseline correction of electric dermal activity which was written in misc channel. Neither baseline correction via .apply_baseline(), nor via epochs = mne.Epochs(for_scr, events_clean, picks=[‘MISC001’], event_id = event_id, tmin=-5.94, tmax=0, baseline=(-5.94, -3.94), preload = True, reject_by_annotation = True, on_missing=‘ignore’, event_repeated= ‘drop’) seems to work properly. Plots of baselined epochs and initial are identical. could you please provide some advices how to overcome this issue.
Thanks!

3

It’s likely that the EDA is not considered as a “data channel” by these methods. I guess we
can consider this a bug. Can you open a issue on github for this with some code snippet to replicate the issue?

thanks
Alex

4

cc @scott-huberty who has recently thought about this issue for eyetracking channels. He might already have a work-in-progress fix going?

5

this issue was solved by changing channel type to data channel:
for_scr.set_channel_types({‘MISC001’: ‘eeg’})

1 Like
6

Hmm, actually, baseline correction works fine with eyetrack channels (for example see the tutorial).

Which is nice because I do occasionally want to baseline correct pupil size data… But maybe technically baseline correction shouldn’t apply to eyetrack channels?

FYI I think apply_baseline excludes misc / stim / gsr etc at this line, which eventually calls _PICK_TYPES_DATA_DICT:

Maybe instead, apply_baseline could have a picks parameter which by default has the current behaviour, but let’s a user specifically pick other channels, for example if they are analyzing galvanic skin response data?

7

Ah right I forgot that we added a new data channel type for eyetrack (they’re not misc).

Reclassifying them as EEG is a good workaround, but maybe we should have a larger discussion among devs about how to handle this going forward.

8

seconding this!
I’m working with misc (Accelerometer data) and EMG channels, I’m trying to apply_baseline only to the EMG channel without much luck
I tried a workaround: create two objects by picking the EMG channel, then applying baseline and merging it with the original data again without much luck

epochs_emg = epochs.copy().pick(picks=['EMG'])
epochs_emg.apply_baseline(baseline=(None,None))
epochs_emg_acc = epochs_emg.add_channels(epochs.copy().pick(picks=["ACC_X","ACC_Y","ACC_Z"]))
epochs_emg_acc.plot(picks=["ACC_X","ACC_Y","ACC_Z","EMG"])

pb_mne