mne_denoise.bss_cca.BSSCCA#

Reference-free BSS-CCA estimator.

The estimator learns fixed channel-space operators from the fitted data and reuses them during transform.

Parameters:

lag_samplesint or None, default=None: Positive lag in samples.
lag_secondsfloat or None, default=None: Positive lag in seconds; NumPy input then requires sfreq.
sfreqfloat or None, default=None: Sampling frequency for NumPy data.
n_removeint or None, default=None: Number of components to remove.
rho_thresholdfloat or None, default=None: Correlation threshold used instead of n_remove. Exactly one selection rule is required.
reject{“low”, “high”}, default=”low”: End of the correlation spectrum treated as artifactual.
threshold_on{“rho”, “rsq”}, default=”rho”: Scale for rho_threshold.
segment_lenfloat or None, default=None: Continuous-data block length in seconds.
overlapfloat, default=0.0: Fraction shared by neighboring blocks.
preserve_meanbool, default=True: Add the fitted channel mean after cleaning.
verbosebool, str, int, or None, default=None: Logging level.

Attributes:

cleaning_matrix_ndarray or tuple of ndarray: Fitted channel-space operator(s).
filters_ndarray: Canonical filters ordered by decreasing correlation.
patterns_ndarray: Least-squares channel patterns.
correlations_ndarray: Canonical correlations.
autocorrelations_ndarray: Signed lagged autocorrelations.
filter_asymmetry_ndarray: Canonical-filter asymmetry diagnostic.
kept_mask_ndarray of bool: Retained component mask.
training_mean_ndarray: Fitted channel mean.
spans_tuple: Block spans when segmented.
lag_samples_int: Resolved lag in samples.

See also

mne_denoise.icanclean.ICanClean: Reference-based CCA cleaning using physical or derived reference signals.
compute_bss_cca: One-shot functional interface for array data.

Notes

With segment_len set, transform requires the same number of samples used during fit. MNE Raw, Epochs, and Evoked inputs preserve their container type and channel metadata [1][2].

References

Examples

>>> import numpy as np
>>> from mne_denoise.bss_cca import BSSCCA
>>> rng = np.random.default_rng(0)
>>> data = rng.standard_normal((8, 2000))
>>> model = BSSCCA(sfreq=250.0, lag_samples=1, n_remove=1)
>>> clean = model.fit_transform(data)

fit(X: Any, y=None, *, callback=None, verbose: bool | str | int | None = None) → BSSCCA[source]#

Fit the BSS-CCA operators.

Parameters:

Xarray-like or MNE Raw, Epochs, or Evoked: Data used to learn the operators.
yNone, default=None: Ignored for scikit-learn compatibility.
callbackcallable or None, default=None: Synchronous progress callback in segmented mode.
verbosebool, str, int, or None, default=None: Logging level.

Returns:

BSSCCA: The fitted estimator.

fit_transform(X: Any, y=None, *, callback=None, verbose: bool | str | int | None = None, **fit_params) → Any[source]#

Fit BSS-CCA and apply the fitted operators to X.

Parameters:

Xarray-like or MNE Raw, Epochs, or Evoked: Data to fit and transform.
yNone, default=None: Ignored for scikit-learn compatibility.
callbackcallable or None, default=None: Synchronous progress callback in segmented mode.
verbosebool, str, int, or None, default=None: Logging level.
**fit_paramsdict: Reserved for scikit-learn compatibility.

Returns:

same type as X: A cleaned copy.

transform(X: Any, y=None, *, verbose: bool | str | int | None = None) → Any[source]#

Apply the fitted BSS-CCA operators.

Parameters:

Xarray-like or MNE Raw, Epochs, or Evoked: Data with the fitted channel layout.
yNone, default=None: Ignored for scikit-learn compatibility.
verbosebool, str, int, or None, default=None: Logging level.

Returns:

same type as X: A cleaned copy.