mne_denoise.dss.denoisers.RobustTanhDenoiser

class mne_denoise.dss.denoisers.RobustTanhDenoiser(alpha: float = 1.0)

Robust tanh denoiser (FastICA / RobustICA formulation).

Implements:

$s_{new} = s - \tanh(\alpha \cdot s)$

This form is often used in deflationary FastICA schemas (like pow3) where strictly structure relates to optimizing specific cost functions (like negentropy).

Parameters:

alpha (float) – Scaling factor. Default 1.0.

Examples

>>> # Use for robust ICA
>>> from mne_denoise.dss.denoisers import RobustTanhDenoiser, beta_tanh
>>> denoiser = RobustTanhDenoiser()
>>> dss = IterativeDSS(denoiser=denoiser, beta=beta_tanh)

References

Särelä & Valpola (2005). Section 4.2.2 “BETTER ESTIMATE FOR THE SIGNAL VARIANCE”

__init__(alpha: float = 1.0) → None

Methods

__init__([alpha])
denoise(source)	Apply robust tanh denoising.