braindecode.preprocessing.FindBadChannelsLof#

class braindecode.preprocessing.FindBadChannelsLof(n_neighbors=20, *, picks=None, metric='euclidean', threshold=1.5, return_scores=False, verbose=None)[source]#

Braindecode preprocessor wrapper for find_bad_channels_lof().

Find bad channels using Local Outlier Factor (LOF) algorithm.

Parameters:

rawinstance of Raw: Raw data to process.
n_neighborsint: Number of neighbors defining the local neighborhood (default is 20). Smaller values will lead to higher LOF scores.
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.
metricstr: Metric to use for distance computation. Default is “euclidean”, see sklearn.metrics.pairwise.distance_metrics() for details.
thresholdfloat: Threshold to define outliers. Theoretical threshold ranges anywhere between 1.0 and any positive integer. Default: 1.5 It is recommended to consider this as an hyperparameter to optimize.
return_scoresbool: If True, return a dictionary with LOF scores for each evaluated channel. Default is False.
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:

noisy_chslist: List of bad M/EEG channels that were automatically detected.
scoresndarray, shape (n_picks,): Only returned when return_scores is True. It contains the LOF outlier score for each channel in picks.