Built-in plotting methods for Raw objects

This tutorial shows how to plot continuous data as a time series, how to plot the spectral density of continuous data, and how to plot the sensor locations and projectors stored in :class:~mne.io.Raw objects. :depth: 2

As usual we'll start by importing the modules we need, loading some example data <sample-dataset>, and cropping the :class:~mne.io.Raw object to just 60 seconds before loading it into RAM to save memory:

We've seen in a previous tutorial <tut-raw-class> how to plot data from a :class:~mne.io.Raw object using :doc:`matplotlib

matplotlib:index, but :class:~mne.io.Raw` objects also have several built-in plotting methods:

• :meth:~mne.io.Raw.plot
• :meth:~mne.io.Raw.plot_psd
• :meth:~mne.io.Raw.plot_psd_topo
• :meth:~mne.io.Raw.plot_sensors
• :meth:~mne.io.Raw.plot_projs_topomap

The first three are discussed here in detail; the last two are shown briefly and covered in-depth in other tutorials.

Interactive data browsing with Raw.plot() ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

The :meth:~mne.io.Raw.plot method of :class:~mne.io.Raw objects provides a versatile interface for exploring continuous data. For interactive viewing and data quality checking, it can be called with no additional parameters:

It may not be obvious when viewing this tutorial online, but by default, the :meth:~mne.io.Raw.plot method generates an interactive plot window with several useful features:

• It spaces the channels equally along the y-axis.

  • 20 channels are shown by default; you can scroll through the channels using the :kbd:↑ and :kbd:↓ arrow keys, or by clicking on the colored scroll bar on the right edge of the plot.

  • The number of visible channels can be adjusted by the n_channels parameter, or changed interactively using :kbd:page up and :kbd:page down keys.

  • You can toggle the display to "butterfly" mode (superimposing all channels of the same type on top of one another) by pressing :kbd:b, or start in butterfly mode by passing the butterfly=True parameter.

• It shows the first 10 seconds of the :class:~mne.io.Raw object.

  • You can shorten or lengthen the window length using :kbd:home and :kbd:end keys, or start with a specific window duration by passing the duration parameter.

  • You can scroll in the time domain using the :kbd:← and :kbd:→ arrow keys, or start at a specific point by passing the start parameter. Scrolling using :kbd:shift:kbd:→ or :kbd:shift:kbd:← scrolls a full window width at a time.

• It allows clicking on channels to mark/unmark as "bad".

  • When the plot window is closed, the :class:~mne.io.Raw object's info attribute will be updated, adding or removing the newly (un)marked channels to/from the :class:~mne.Info object's bads field (AKA raw.info['bads']).

.. TODO: discuss annotation snapping in the below bullets

• It allows interactive :term:annotation <annotations> of the raw data.

  • This allows you to mark time spans that should be excluded from future computations due to large movement artifacts, line noise, or other distortions of the signal. Annotation mode is entered by pressing :kbd:a. See annotations-tutorial for details.

• It automatically applies any :term:projectors <projector> before plotting the data.

  • These can be enabled/disabled interactively by clicking the Proj button at the lower right corner of the plot window, or disabled by default by passing the proj=False parameter. See tut-projectors-background for more info on projectors.

These and other keyboard shortcuts are listed in the Help window, accessed through the Help button at the lower left corner of the plot window. Other plot properties (such as color of the channel traces, channel order and grouping, simultaneous plotting of :term:events, scaling, clipping, filtering, etc.) can also be adjusted through parameters passed to the :meth:~mne.io.Raw.plot method; see the docstring for details.

Plotting spectral density of continuous data ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

To visualize the frequency content of continuous data, the :class:~mne.io.Raw object provides a :meth:~mne.io.Raw.plot_psd to plot the spectral density_ of the data.

If the data have been filtered, vertical dashed lines will automatically indicate filter boundaries. The spectrum for each channel type is drawn in its own subplot; here we've passed the average=True parameter to get a summary for each channel type, but it is also possible to plot each channel individually, with options for how the spectrum should be computed, color-coding the channels by location, and more. For example, here is a plot of just a few sensors (specified with the picks parameter), color-coded by spatial location (via the spatial_colors parameter, see the documentation of :meth:~mne.io.Raw.plot_psd for full details):

Alternatively, you can plot the PSD for every sensor on its own axes, with the axes arranged spatially to correspond to sensor locations in space, using :meth:~mne.io.Raw.plot_psd_topo:

This plot is also interactive; hovering over each "thumbnail" plot will display the channel name in the bottom left of the plot window, and clicking on a thumbnail plot will create a second figure showing a larger version of the selected channel's spectral density (as if you had called :meth:~mne.io.Raw.plot_psd on that channel).

By default, :meth:~mne.io.Raw.plot_psd_topo will show only the MEG channels if MEG channels are present; if only EEG channels are found, they will be plotted instead:

Plotting sensor locations from Raw objects ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

The channel locations in a :class:~mne.io.Raw object can be easily plotted with the :meth:~mne.io.Raw.plot_sensors method. A brief example is shown here; notice that channels in raw.info['bads'] are plotted in red. More details and additional examples are given in the tutorial tut-sensor-locations.

Plotting projectors from Raw objects ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

As seen in the output of :meth:mne.io.read_raw_fif above, there are :term:projectors <projector> included in the example :class:~mne.io.Raw file (representing environmental noise in the signal, so it can later be "projected out" during preprocessing). You can visualize these projectors using the :meth:~mne.io.Raw.plot_projs_topomap method. By default it will show one figure per channel type for which projectors are present, and each figure will have one subplot per projector. The three projectors in this file were only computed for magnetometers, so one figure with three subplots is generated. More details on working with and plotting projectors are given in tut-projectors-background and tut-artifact-ssp.

.. LINKS