GSR and temperature channels do not show up in plot

[hoechenberger]

MWE:

# %%
from pathlib import Path
import openneuro
import mne
import mne_bids

dataset_name = 'ds003061'
bids_root = Path('/tmp') / dataset_name

# %%
openneuro.download(
    dataset=dataset_name,
    include='sub-001',
    target_dir=bids_root,
)

# %%
bp = mne_bids.BIDSPath(
    subject='001',
    task='P300',
    run='1',
    datatype='eeg',
    suffix='eeg',
    extension='.set',
    root=bids_root,
)
raw = mne_bids.read_raw_bids(bp)

with mne.viz.use_browser_backend('matplotlib'):
    raw.copy().pick(['GSR1', 'GSR2' 'Temp']).plot()

produces:

[larsoner]

This is probably a scaling issue. There were no data to test with at the time of addition, so the mne/defaults.py scale factors are probably way off. In particular, the raw scale for temperature is almost certainly a problem here since room temperature (~20 deg C) will be off the plot given a unity plot_raw scale factor (the default right now IIRC)

[hoechenberger]

This is probably a scaling issue.

You sure? I don't even see channel names on the ordinate … I think we're just not doing … anything. No?

Same problem with the qt-browser too btw

cc @marsipu

[larsoner]

Ahh I didn't even notice that. Yeah it seems like a picking problem

[marsipu]

@hoechenberger Yes, when it is in matplotlib and qt-backend and no traces/y-axis-marks show I think too that it is a picking problem inside mne-python-code

[larsoner]

I can try to look tomorrow

[larsoner]

@hoechenberger running your example I get a bunch of junk from MNE-BIDS, followed by a real error:

/home/larsoner/python/mne-bids/mne_bids/read.py:60: RuntimeWarning: Unknown types found, setting as type EEG:
temp: ['Temp']
  raw = reader[ext](raw_path, **kwargs)
/home/larsoner/python/mne-bids/mne_bids/read.py:60: RuntimeWarning: Data will be preloaded. preload=False or a string preload is not supported when the data is stored in the .set file
  raw = reader[ext](raw_path, **kwargs)
/home/larsoner/python/mne-bids/mne_bids/read.py:60: RuntimeWarning: Not setting positions of 14 gsr/misc/resp channels found in montage:
['EXG1', 'EXG2', 'EXG3', 'EXG4', 'EXG5', 'EXG6', 'EXG7', 'EXG8', 'GSR1', 'GSR2', 'Erg1', 'Erg2', 'Resp', 'Plet']
Consider setting the channel types to be of EEG/sEEG/ECoG/DBS/fNIRS using inst.set_channel_types before calling inst.set_montage, or omit these channels when creating your montage.
  raw = reader[ext](raw_path, **kwargs)
/home/larsoner/python/mne-bids/mne_bids/read.py:586: RuntimeWarning: The unit for channel(s) EXG1, EXG2, EXG3, EXG4, EXG5, EXG6, EXG7, EXG8, Erg1, Erg2, Plet has changed from NA to V.
  raw.set_channel_types(channel_type_bids_mne_map_available_channels)
/home/larsoner/python/mne-bids/mne_bids/read.py:586: RuntimeWarning: The unit for channel(s) Temp has changed from V to C.
  raw.set_channel_types(channel_type_bids_mne_map_available_channels)
/home/larsoner/python/mne-bids/mne_bids/dig.py:540: RuntimeWarning: There are channels without locations (n/a) that are not marked as bad: ['EXG1', 'EXG2', 'EXG3', 'EXG4', 'EXG5', 'EXG6', 'EXG7', 'EXG8', 'GSR1', 'GSR2', 'Erg1', 'Erg2', 'Resp', 'Plet', 'Temp']
  warn(f"There are channels without locations "
/home/larsoner/python/mne-bids/mne_bids/dig.py:550: RuntimeWarning: Not setting positions of 15 gsr/misc/resp/temperature channels found in montage:
['EXG1', 'EXG2', 'EXG3', 'EXG4', 'EXG5', 'EXG6', 'EXG7', 'EXG8', 'GSR1', 'GSR2', 'Erg1', 'Erg2', 'Resp', 'Plet', 'Temp']
Consider setting the channel types to be of EEG/sEEG/ECoG/DBS/fNIRS using inst.set_channel_types before calling inst.set_montage, or omit these channels when creating your montage.
  raw.set_montage(montage, on_missing='warn')
/home/larsoner/Desktop/temp.py:29: RuntimeWarning: Channel(s) ['GSR2Temp'] could not be picked, because they are not present in the info instance.
  raw.copy().pick(['GSR1', 'GSR2' 'Temp']).plot()
/home/larsoner/python/mne-python/mne/viz/_mpl_figure.py:441: UserWarning: Attempting to set identical low and high ylims makes transformation singular; automatically expanding.
  ax_vscroll.set_ylim(len(ch_order), 0)
/home/larsoner/python/mne-python/mne/viz/_mpl_figure.py:1733: UserWarning: Attempting to set identical low and high ylims makes transformation singular; automatically expanding.
  self.mne.ax_main.set_ylim(ylim)
Traceback (most recent call last):
  File "/home/larsoner/Desktop/temp.py", line 29, in <module>
    raw.copy().pick(['GSR1', 'GSR2' 'Temp']).plot()
  File "/home/larsoner/python/mne-python/mne/io/base.py", line 1529, in plot
    return plot_raw(self, events, duration, start, n_channels, bgcolor,
  File "<decorator-gen-155>", line 12, in plot_raw
  File "/home/larsoner/python/mne-python/mne/viz/raw.py", line 355, in plot_raw
    fig = _get_browser(show=show, block=block, **params)
  File "/home/larsoner/python/mne-python/mne/viz/_figure.py", line 659, in _get_browser
    fig = backend._init_browser(**kwargs)
  File "/home/larsoner/python/mne-python/mne/viz/_mpl_figure.py", line 2278, in _init_browser
    fig._update_trace_offsets()
  File "/home/larsoner/python/mne-python/mne/viz/_mpl_figure.py", line 1738, in _update_trace_offsets
    self._update_yaxis_labels()
  File "/home/larsoner/python/mne-python/mne/viz/_mpl_figure.py", line 1642, in _update_yaxis_labels
    ticklabels = self.mne.ch_names[self.mne.picks]
IndexError: arrays used as indices must be of integer (or boolean) type

I see in the output and a quick sanity check a few things:

1. MNE-BIDS not setting the Temp channel as a temperature channel
2. Missing comma between 'GSR2' and 'Temp' causing them not to be picked
3. Us not handling empty picks correctly, which I'm fixing
4. The data appearing broken on read: raw.get_data('temperature').max() is -5.005000233650207e-07

I'll open a WIP PR to fix the picking bug (3), but two of the other things (1, 4) probably need to be fixed at the MNE-BIDS/OpenNeuro end I think.

[larsoner]

Please also check the GSR, I see:

>>> raw.get_data('gsr').max()
-4.843750298023223e-07
>>> raw.get_data('gsr').min()
-4.843750298023223e-07

If these are reasonable values we can figure out how to plot them nicely. But it appears totally constant, which seems suspicious.