mne_denoise.sound.compute_sound_ref_best

mne_denoise.sound.compute_sound_ref_best(data: ndarray, leadfield: ndarray, *, lambda_: float = 0.1, n_iter: int = 5, tol: float | None = None, random_state=None, callback=None, verbose: bool | str | int | None = None) → tuple[ndarray, ndarray, ndarray, int]

Compute SOUND with a selected single-channel reference.

The function excludes the least-noisy reference channel during estimation and returns an average-referenced full-channel operator.

Parameters:

datandarray, shape (n_channels, n_times)

Sensor data.

leadfieldndarray, shape (n_channels, n_sources)

Lead-field matrix in the same channel order and reference as data.

lambda_float, default=0.1

Non-negative regularization scale.

n_iterint, default=5

Maximum number of iterations.

tolfloat or None, default=None

Convergence tolerance.

random_stateint, numpy.random.Generator, or None, default=None

Random state for channel-update order.

callbackcallable or None, default=None

Synchronous callback after each iteration.

verbosebool, str, int, or None, default=None

Logging level.

Returns:

operatorndarray, shape (n_channels, n_channels)

Average-referenced cleaning operator.

sigmasndarray, shape (n_channels - 1,)

Noise amplitudes for channels other than best_channel.

convergencendarray, shape (n_iter_run,)

Maximum relative noise-level change per iteration.

best_channelint

Index of the selected reference channel.

Notes

The reference channel is selected from the data-driven Wiener noise estimates; its index is returned so the reference choice remains explicit.