Inconsistent behavior between Epochs and Evoked in get_data() method when picking EEG channels and bads are present

Description of the problem

Epochs.get_data(picks="eeg") excludes bad channels, but Evoked.get_data(picks="eeg") doesn't

Steps to reproduce

# %%
import mne

sample_dir = mne.datasets.sample.data_path()
sample_fname = sample_dir / "MEG" / "sample" / "sample_audvis_raw.fif"

raw = mne.io.read_raw_fif(sample_fname)
raw.crop(tmax=60)

events = mne.find_events(raw, stim_channel="STI 014")
raw.pick("eeg")

epochs = mne.Epochs(raw, events=events, preload=True)
evoked = epochs.average()

# %%
for inst_type, inst in zip(["epochs", "evoked"], [epochs, evoked]):
    print(f"{inst_type}:")
    print(f"    bads: {inst.info['bads']}")
    print(f"    data shape: {inst.get_data(picks='eeg').shape}")

Link to data

No response

Expected results

Both methods should behave the same in regards to bad channel treatment.

Actual results

epochs:
    bads: ['EEG 053']
    data shape: (86, 59, 421)
evoked:
    bads: ['EEG 053']
    data shape: (60, 421)

As you can see, the data from Evoked contains the bad channel, while the data from Epochs does not.

Additional information

This is with main

The docstrings are identical, so I think both methods should behave identically. In addition, the docstring says:

None (default) will pick all channels.

I think this should be:

None (default) will pick all good channels.

At least that's how it currently works for Epochs, but not for Evoked apparently.

https://mne.tools/dev/generated/mne.Epochs.html#mne.Epochs.get_data
https://mne.tools/dev/generated/mne.Evoked.html#mne.Evoked.get_data

@cbrnr

@cbrnr Note that I did not pass None in my MWE! I did pass picks="eeg"

The doctoring says:

Note that channels in info['bads'] will be included if their names or indices are explicitly provided.

So something about this point is inconsistent between epochs and evoked at least when passing picks="eeg".

I think what you're mentioning here is being tracked at #12197

So something about this point is inconsistent between epochs and evoked at least when passing picks="eeg".

Neither get_data has an exclude param, sounds like one is implicitly using exclude="bads" and the other exclude=() under the hood. To change the behavior of one to match the other and be consistent I think we'd need a deprecation cycle, not sure whether or not it's worth the code churn. To update the docstring of both to reflect their actual behavior it wouldn't require a deprecation cycle. Adding exclude to both (to make the behavior more explicit and controllable) is probably worthwhile either way.

Even if documented I think it would be a bit surprising if Epochs and Evoked used different pick defaults. So I'd document the behavior with whatever version we think makes more sense, be consistent for both methods and deprecate the inconsistent behavior. And yes, an explicit exclude parameter would definitely make sense to me.

I tested the MWE with Raw and now it seems it's actually Epochs who's the odd one:

# %%
import mne

sample_dir = mne.datasets.sample.data_path()
sample_fname = sample_dir / "MEG" / "sample" / "sample_audvis_raw.fif"

raw = mne.io.read_raw_fif(sample_fname)
raw.crop(tmax=60)

events = mne.find_events(raw, stim_channel="STI 014")
raw.pick("eeg")

epochs = mne.Epochs(raw, events=events, preload=True)
evoked = epochs.average()

# %%
for inst_type, inst in zip(["raw", "epochs", "evoked"], [raw, epochs, evoked]):
    print(f"{inst_type}:")
    print(f"    bads: {inst.info['bads']}")
    print(f"    data shape: {inst.get_data(picks='eeg').shape}")

raw:
    bads: ['EEG 053']
    data shape: (60, 36038)
epochs:
    bads: ['EEG 053']
    data shape: (86, 59, 421)
evoked:
    bads: ['EEG 053']
    data shape: (60, 421)

More digging:

BaseSpectrum.get_data() features exclude="bads'
BaseTFR ditto

My feeling is that we should add exclude parameters to all these get_data() methods if they don't have one already, retaining the current behavior; and do a deprecation cycle for transitioning to either exclude=() or exclude="bads" for all of them

@luisademelo14

Hello,

Me and two friends (@luisademelo14 @Fooxfeir) are taking a course and we need to choose a tool and bug to fix. We chose MNE-Python, but since we are new to the community, we decided to pick a bug marked as easy and this one seems good.
If there's a problem with that, please let us know!

@RenanFlorencio

#12197 is also related

FYI not all of this issue is easy... the first part is probably uncontroversial: namely adding exclude params to functions that don't have them, retaining old behavior, and documenting things properly (#12197). I think documenting current behavior and making it changeable by adding exclude=... in places we don't have it is already a good start here, so @RenanFlorencio @luisademelo14 @Fooxfeir I think you can work on this part! 👍 So feel free to start on that

The deprecation cycle to transition them all to have the same behavior will not be trivial. I suspect a lot of tests will need to be adjusted, and it will be hard to do without a reasonably deep working knowledge of the code.

I am also not 100% sure we should do the deprecation to make everything have either exclude=() or exclude="bads everywhere. There are probably functions and classes where we should exclude bads -- like in multivariate connectivity fitting for example we've talked about it and there are probably other examples -- and other functions/classes (like constructing epochs maybe) it is convenient to include them by default. We'd have to look at all use cases (decoding, inverses, stats, time-freq/spectra, connectivity, etc.) to see what makes sense here. Moreover, if we change the behavior of something like Epochs the risk of breaking people's code is quite high even with a deprecation cycle... so we should think really hard about whether the potential disruption is worth the uniformity, and how much it buys us.

@larsoner

Okay @larsoner , we worked on the first part and we added the exclude parameter to all three functions (epochs, evoked and base, since spectrum and tfr already have it), this way we have three different behaviors for each. We would like you to confirm this is what's expected to write the docstring. We also want to ask you a question that is at the end of this comment.

Maintained Previous Behavior:

for inst_type, inst in zip(["raw", "epochs", "evoked"], [raw, epochs, evoked]):
    print(f"{inst_type}:")
    print(f"    bads: {inst.info['bads']}")
    print(f"    data shape: {inst.get_data(picks='eeg').shape}")

Results in:

raw:
    bads: ['EEG 053']
    data shape: (60, 36038)
epochs:
    bads: ['EEG 053']
    data shape: (86, 59, 421)
evoked:
    bads: ['EEG 053']
    data shape: (60, 421)

Behavior when exclude=()

Now, both functions will not exclude bad channels:

for inst_type, inst in zip(["raw", "epochs", "evoked"], [raw, epochs, evoked]):
    print(f"{inst_type}:")
    print(f"    bads: {inst.info['bads']}")
    print(f"    data shape: {inst.get_data(picks='eeg', exclude=()).shape}")

Gives us:

raw:
    bads: ['EEG 053']
    data shape: (60, 36038)
epochs:
    bads: ['EEG 053']
    data shape: (86, 60, 421)
evoked:
    bads: ['EEG 053']
    data shape: (60, 421)

Behavior when exclude="bads"

Now, both functions will exclude bad channels:

for inst_type, inst in zip(["raw", "epochs", "evoked"], [raw, epochs, evoked]):
    print(f"{inst_type}:")
    print(f"    bads: {inst.info['bads']}")
    print(f"    data shape: {inst.get_data(picks='eeg', exclude='bads').shape}")

Resulting in:

raw:
    bads: ['EEG 053']
    data shape: (59, 36038)
epochs:
    bads: ['EEG 053']
    data shape: (86, 59, 421)
evoked:
    bads: ['EEG 053']
    data shape: (59, 421)

Question

Lastly, we would like to ask you about the following lines of code in epochs.py at the _get_data function:

        if orig_picks is None:
            picks = _picks_to_idx(self.info, picks, "all", exclude=())
        else:
            picks = _picks_to_idx(self.info, picks, exclude=exclude)

Here you can see that we changed the line inside the else block to add the exclude parameter. However, the if block is still "hard coded" to exclude=(). Should it also use the new exclude parameter? If so, we would have to figure out a way to keep the old behavior, since the else block would need a default exclude=bads, which we are using, and the if block would need exclude=().

Conclusion

There are two functions, in raw and evoked, use default exclude=(). The other three, in epochs, spectrum and tfr use default `exclude='bads'. If we were to do a deprecation cycle, we would have to think more about which one makes more sense. Moreover, do you have any suggestions of issues we could work on?

Maintained Previous Behavior:

For this, rather than using print how about you write some tiny test on main that should pass that asserts something about shapes of returned data?

Then you add the exclude param defaults, and your test ensures that you haven't changed behavior.

Then you append to that same test some calls with different exclude values and assert something new/different about the obtained shapes.

This is a variant of test-driven development (strictly speaking there you'd write the whole test first even though it would fail, then modify code, but in practice it's nice to see stuff green / passing to start)

Lastly, we would like to ask you about the following lines of code in epochs.py at the _get_data function:

I would try not to mess with this one. Internal functions get split to handle specific use cases... the epochs code gets used all over the place so at the end of the day having CIs/tests green should tell us if things are okay or not.

If we were to do a deprecation cycle, we would have to think more about which one makes more sense.

Let's not do that for now... I think it would be more end-user pain that it's worth.

Moreover, do you have any suggestions of issues we could work on?

Anything with the EASY label should in principle be somewhat easy, at least that's the goal 😄

https://github.com/mne-tools/mne-python/issues?q=state%3Aopen%20label%3AEASY

Update evoked.py

Navigation Menu

Repository navigation

Inconsistent behavior between `Epochs` and `Evoked` in `get_data()` method when picking EEG channels and bads are present #12577

Description of the problem

Steps to reproduce

Link to data

Expected results

Actual results

Additional information

Maintained Previous Behavior:

Behavior when exclude=()

Behavior when exclude="bads"

Question

Conclusion

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Inconsistent behavior between Epochs and Evoked in get_data() method when picking EEG channels and bads are present #12577

Description

Description of the problem

Steps to reproduce

Link to data

Expected results

Actual results

Additional information

Activity

cbrnr commented on Apr 26, 2024

hoechenberger commented on Apr 26, 2024

larsoner commented on Apr 26, 2024

cbrnr commented on Apr 26, 2024

hoechenberger commented on Apr 26, 2024

hoechenberger commented on Apr 26, 2024

hoechenberger commented on Apr 26, 2024

RenanFlorencio commented on Nov 6, 2025

larsoner commented on Nov 7, 2025

RenanFlorencio commented on Nov 11, 2025

Maintained Previous Behavior:

Behavior when exclude=()

Behavior when exclude="bads"

Question

Conclusion

larsoner commented on Nov 25, 2025

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

Inconsistent behavior between `Epochs` and `Evoked` in `get_data()` method when picking EEG channels and bads are present #12577