mne_denoise.dss.denoisers.LagAverageBias#

class mne_denoise.dss.denoisers.LagAverageBias(lags: int | ndarray = 10, weighting: str = 'uniform')[source]#

Lag-averaging bias for emphasizing temporally smooth signals.

Creates a bias by averaging time-shifted versions of the data, emphasizing signals that remain similar across the selected lags. This is a lightweight package bias for ordinary sensor-space DSS; it is not the lag-augmented TimeShiftDSS estimator.

Parameters:

  • lags (int or array-like) – If int, use lags from 1 through lags. If array, use specified lag values in samples. Default 10.

  • weighting ({'uniform', 'inverse_lag'}) – Weight every lag equally or weight it by inverse absolute lag. Neither option estimates an autocorrelation function or fits prediction coefficients.

Examples

>>> bias = LagAverageBias(lags=[1, 2, 5, 10], weighting="inverse_lag")
>>> biased_data = bias.apply(data)

See also

SmoothingBias

Bias for low-frequency signals.

__init__(lags: int | ndarray = 10, weighting: str = 'uniform') → None[source]#

Methods

__init__([lags, weighting])

apply(data)

Apply time-shift bias.