braindecode.preprocessing.CombineChannels#

class braindecode.preprocessing.CombineChannels(groups, method='mean', keep_stim=False, drop_bad=False, *, on_missing='raise', verbose=None)[source]#

Braindecode preprocessor wrapper for combine_channels().

Combine channels based on specified channel grouping.

Parameters:
instinstance of Raw, Epochs, or Evoked

An MNE-Python object to combine the channels for. The object can be of type Raw, Epochs, or Evoked.

groupsdict

Specifies which channels are aggregated into a single channel, with aggregation method determined by the method parameter. One new pseudo-channel is made per dict entry; the dict values must be lists of picks (integer indices of ch_names). For example:

groups=dict(Left=[1, 2, 3, 4], Right=[5, 6, 7, 8])

Note that within a dict entry all channels must have the same type.

methodstr | callable

Which method to use to combine channels. If a str, must be one of ‘mean’, ‘median’, or ‘std’ (standard deviation). If callable, the callable must accept one positional input (data of shape (n_channels, n_times), or (n_epochs, n_channels, n_times)) and return an array of shape (n_times,), or (n_epochs, n_times). For example with an instance of Raw or Evoked:

method = lambda data: np.mean(data, axis=0)

Another example with an instance of Epochs:

method = lambda data: np.median(data, axis=1)

Defaults to 'mean'.

keep_stimbool

If True, include stimulus channels in the resulting object. Defaults to False.

drop_badbool

If True, drop channels marked as bad before combining. Defaults to False.

on_missing‘raise’ | ‘warn’ | ‘ignore’

What to do if one or several event ids are not found in the recording. Valid keys are ‘raise’ | ‘warn’ | ‘ignore’ Default is 'raise'. If 'warn', it will proceed but warn; if 'ignore', it will proceed silently.

Note

If none of the event ids are found in the data, an error will be automatically generated irrespective of this parameter.

Added in version 1.11.0.

verbosebool | str | int | None

Control verbosity of the logging output. If None, use the default verbosity level. See the logging documentation and mne.verbose() for details. Should only be passed as a keyword argument.

Returns:
combined_instinstance of Raw, Epochs, or Evoked

An MNE-Python object of the same type as the input inst, containing one virtual channel for each group in groups (and, if keep_stim is True, also containing stimulus channels).

See also

mne.channels.equalize_channels
mne.channels.rename_channels
mne.channels.unify_bad_channels

Methods

fn(groups, method='mean', keep_stim=False, drop_bad=False, *, on_missing='raise', verbose=None)[source]#

Combine channels based on specified channel grouping.

Parameters:

  • inst (instance of Raw, Epochs, or Evoked) – An MNE-Python object to combine the channels for. The object can be of type Raw, Epochs, or Evoked.

  • groups (dict) –

    Specifies which channels are aggregated into a single channel, with aggregation method determined by the method parameter. One new pseudo-channel is made per dict entry; the dict values must be lists of picks (integer indices of ch_names). For example:

    groups=dict(Left=[1, 2, 3, 4], Right=[5, 6, 7, 8])

    Note that within a dict entry all channels must have the same type.

  • method (str | callable) –

    Which method to use to combine channels. If a str, must be one of ‘mean’, ‘median’, or ‘std’ (standard deviation). If callable, the callable must accept one positional input (data of shape (n_channels, n_times), or (n_epochs, n_channels, n_times)) and return an array of shape (n_times,), or (n_epochs, n_times). For example with an instance of Raw or Evoked:

    method = lambda data: np.mean(data, axis=0)

    Another example with an instance of Epochs:

    method = lambda data: np.median(data, axis=1)

    Defaults to 'mean'.

  • keep_stim (bool) – If True, include stimulus channels in the resulting object. Defaults to False.

  • drop_bad (bool) – If True, drop channels marked as bad before combining. Defaults to False.

  • on_missing ('raise' | 'warn' | 'ignore') –

    What to do if one or several event ids are not found in the recording. Valid keys are ‘raise’ | ‘warn’ | ‘ignore’ Default is 'raise'. If 'warn', it will proceed but warn; if 'ignore', it will proceed silently.

    Note

    If none of the event ids are found in the data, an error will be automatically generated irrespective of this parameter.

    Added in version 1.11.0.

  • verbose (bool | str | int | None) – Control verbosity of the logging output. If None, use the default verbosity level. See the logging documentation and mne.verbose() for details. Should only be passed as a keyword argument.

Returns:

combined_inst – An MNE-Python object of the same type as the input inst, containing one virtual channel for each group in groups (and, if keep_stim is True, also containing stimulus channels).

Return type:

instance of Raw, Epochs, or Evoked

See also

mne.channels.equalize_channels, mne.channels.rename_channels, mne.channels.unify_bad_channels