mne.viz.plot_epochs#

mne.viz.plot_epochs(epochs, picks=None, scalings=None, n_epochs=20, n_channels=20, title=None, events=False, event_color=None, order=None, show=True, block=False, decim='auto', noise_cov=None, butterfly=False, show_scrollbars=True, show_scalebars=True, epoch_colors=None, event_id=None, group_by='type', precompute=None, use_opengl=None, *, theme=None, overview_mode=None, splash=True)[source]#

Visualize epochs.

Bad epochs can be marked with a left click on top of the epoch. Bad channels can be selected by clicking the channel name on the left side of the main axes. Calling this function drops all the selected bad epochs as well as bad epochs marked beforehand with rejection parameters.

Parameters:

epochsinstance of Epochs

The epochs object.

picksstr | array_like | slice | None

Channels to include. Slices and lists of integers will be interpreted as channel indices. In lists, channel type strings (e.g., ['meg', 'eeg']) will pick channels of those types, channel name strings (e.g., ['MEG0111', 'MEG2623'] will pick the given channels. Can also be the string values 'all' to pick all channels, or 'data' to pick data channels. None (default) will pick good data channels. Note that channels in info['bads'] will be included if their names or indices are explicitly provided.

scalings‘auto’ | dict | None

Scaling factors for the traces. If a dictionary where any value is 'auto', the scaling factor is set to match the 99.5th percentile of the respective data. If 'auto', all scalings (for all channel types) are set to 'auto'. If any values are 'auto' and the data is not preloaded, a subset up to 100 MB will be loaded. If None, defaults to:

dict(mag=1e-12, grad=4e-11, eeg=20e-6, eog=150e-6, ecg=5e-4,
     emg=1e-3, ref_meg=1e-12, misc=1e-3, stim=1,
     resp=1, chpi=1e-4, whitened=1e2)

Note

A particular scaling value s corresponds to half of the visualized signal range around zero (i.e. from 0 to +s or from 0 to -s). For example, the default scaling of 20e-6 (20µV) for EEG signals means that the visualized range will be 40 µV (20 µV in the positive direction and 20 µV in the negative direction).

n_epochsint

The number of epochs per view. Defaults to 20.

n_channelsint

The number of channels per view. Defaults to 20.

titlestr | None

The title of the window. If None, the event names (from epochs.event_id) will be displayed. Defaults to None.

eventsbool | array, shape (n_events, 3)

Events to show with vertical bars. You can use plot_events as a legend for the colors. By default, the coloring scheme is the same. True plots epochs.events. Defaults to False (do not plot events).

Warning

If the epochs have been resampled, the events no longer align with the data.

New in v0.14.0.

Changed in version 1.6: Passing events=None was disallowed. The new equivalent is events=False.

event_colorcolor object | dict | None

Color(s) to use for events. To show all events in the same color, pass any matplotlib-compatible color. To color events differently, pass a dict that maps event names or integer event numbers to colors (must include entries for all events, or include a “fallback” entry with key -1). If None, colors are chosen from the current Matplotlib color cycle. Defaults to None.

orderarray of str | None

Order in which to plot channel types.

New in v0.18.0.

showbool

Show figure if True. Defaults to True.

blockbool

Whether to halt program execution until the figure is closed. Useful for rejecting bad trials on the fly by clicking on an epoch. Defaults to False.

decimint | ‘auto’

Amount to decimate the data during display for speed purposes. You should only decimate if the data are sufficiently low-passed, otherwise aliasing can occur. The ‘auto’ mode (default) uses the decimation that results in a sampling rate at least three times larger than info['lowpass'] (e.g., a 40 Hz lowpass will result in at least a 120 Hz displayed sample rate).

New in v0.15.0.

noise_covinstance of Covariance | str | None

Noise covariance used to whiten the data while plotting. Whitened data channels are scaled by scalings['whitened'], and their channel names are shown in italic. Can be a string to load a covariance from disk. See also mne.Evoked.plot_white() for additional inspection of noise covariance properties when whitening evoked data. For data processed with SSS, the effective dependence between magnetometers and gradiometers may introduce differences in scaling, consider using mne.Evoked.plot_white().

New in v0.16.0.

butterflybool

Whether to directly call the butterfly view.

New in v0.18.0.

show_scrollbarsbool

Whether to show scrollbars when the plot is initialized. Can be toggled after initialization by pressing z (“zen mode”) while the plot window is focused. Default is True.

New in v0.19.0.

show_scalebarsbool

Whether to show scale bars when the plot is initialized. Can be toggled after initialization by pressing s while the plot window is focused. Default is True.

New in v0.24.0.

epoch_colorslist of (n_epochs) list (of n_channels) | None

Colors to use for individual epochs. If None, use default colors.

event_idbool | dict

Determines to label the event markers on the plot. If True, uses epochs.event_id. If False, uses integer event codes instead of IDs. If a dict is passed, uses its keys as event labels on the plot for entries whose values are integer codes for events being drawn. Ignored if events=False.

New in v0.20.

group_bystr

How to group channels. 'type' groups by channel type, 'original' plots in the order of ch_names, 'selection' uses Elekta’s channel groupings (only works for Neuromag data), 'position' groups the channels by the positions of the sensors. 'selection' and 'position' modes allow custom selections by using a lasso selector on the topomap. In butterfly mode, 'type' and 'original' group the channels by type, whereas 'selection' and 'position' use regional grouping. 'type' and 'original' modes are ignored when order is not None. Defaults to 'type'.

precomputebool | str

Whether to load all data (not just the visible portion) into RAM and apply preprocessing (e.g., projectors) to the full data array in a separate processor thread, instead of window-by-window during scrolling. The default None uses the MNE_BROWSER_PRECOMPUTE variable, which defaults to 'auto'. 'auto' compares available RAM space to the expected size of the precomputed data, and precomputes only if enough RAM is available. This is only used with the Qt backend.

New in v0.24.

Changed in version 1.0: Support for the MNE_BROWSER_PRECOMPUTE config variable.

use_openglbool | None

Whether to use OpenGL when rendering the plot (requires pyopengl). May increase performance, but effect is dependent on system CPU and graphics hardware. Only works if using the Qt backend. Default is None, which will use False unless the user configuration variable MNE_BROWSER_USE_OPENGL is set to 'true', see mne.set_config().

New in v0.24.

themestr | path-like

Can be “auto”, “light”, or “dark” or a path-like to a custom stylesheet. For Dark-Mode and automatic Dark-Mode-Detection, qdarkstyle and darkdetect, respectively, are required. If None (default), the config option MNE_BROWSER_THEME will be used, defaulting to “auto” if it’s not found. Only supported by the 'qt' backend.

New in v1.0.

overview_modestr | None

Can be “channels”, “empty”, or “hidden” to set the overview bar mode for the 'qt' backend. If None (default), the config option MNE_BROWSER_OVERVIEW_MODE will be used, defaulting to “channels” if it’s not found.

New in v1.1.

splashbool

If True (default), a splash screen is shown during the application startup. Only applicable to the qt backend.

New in v1.6.

Returns:

figmatplotlib.figure.Figure | mne_qt_browser.figure.MNEQtBrowser: Browser instance.

Notes

The arrow keys (up/down/left/right) can be used to navigate between channels and epochs and the scaling can be adjusted with - and + (or =) keys, but this depends on the backend matplotlib is configured to use (e.g., mpl.use(TkAgg) should work). Full screen mode can be toggled with f11 key. The amount of epochs and channels per view can be adjusted with home/end and page down/page up keys. h key plots a histogram of peak-to-peak values along with the used rejection thresholds. Butterfly plot can be toggled with b key. Left mouse click adds a vertical line to the plot. Click ‘help’ button at bottom left corner of the plotter to view all the options.

MNE-Python provides two different backends for browsing plots (i.e., raw.plot(), epochs.plot(), and ica.plot_sources()). One is based on matplotlib, and the other is based on PyQtGraph. You can set the backend temporarily with the context manager mne.viz.use_browser_backend(), you can set it for the duration of a Python session using mne.viz.set_browser_backend(), and you can set the default for your computer via mne.set_config('MNE_BROWSER_BACKEND', 'matplotlib') (or 'qt').

Note

For the PyQtGraph backend to run in IPython with block=False you must run the magic command %gui qt5 first.

Note

To report issues with the PyQtGraph backend, please use the issues of mne-qt-browser.

New in v0.10.0.