mne_denoise.dss.iterative_dss#

mne_denoise.dss.iterative_dss(data: ndarray, denoiser: Callable[[ndarray], ndarray], n_components: int, *, method: str = 'deflation', rank: int | None = None, reg: float = 1e-09, max_iter: int = 100, tol: float = 1e-06, w_init: ndarray | None = None, verbose: bool | str | int | None = None, alpha: float | Callable | None = None, beta: float | Callable | None = None, gamma: float | Callable | None = None, random_state: int | Generator | None = None, callback=None) tuple[ndarray, ndarray, ndarray, ndarray][source]#

Extract multiple components with iterative DSS.

Parameters:
datandarray, shape (n_channels, n_times) or (n_channels, n_times, n_epochs)

Channel-first data; epochs are flattened for the solve.

denoisercallable

Nonlinear source transformation.

n_componentsint

Number of components to extract, limited by the whitening rank.

method{“deflation”, “symmetric”}, default=”deflation”

Sequential deflation or simultaneous symmetric extraction.

rankint or None, default=None

Whitening rank.

regfloat, default=1e-9

Relative whitening threshold.

max_iterint, default=100

Maximum iterations per component or symmetric solve.

tolfloat, default=1e-6

Convergence tolerance.

w_initndarray or None, default=None

Initial filter vector or matrix.

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

Logging level.

alpha, beta, gammafloat, callable, or None

Optional fixed-point parameters passed to iterative_dss_one().

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

Random state for initialization.

callbackcallable or None, default=None

Synchronous progress callback for fixed-point iterations.

Returns:
filtersndarray, shape (n_components, n_channels)

Sensor-space spatial filters.

sourcesndarray, shape (n_components, n_times)

Extracted sources.

patternsndarray, shape (n_channels, n_components)

Sensor-space mixing patterns.

convergence_infondarray, shape (n_components, 2)

Iteration count and convergence flag for each component.

Notes

This follows the iterative DSS formulation [1].

References