calibrain.BMN_joint#

calibrain.BMN_joint(L, y, noise_var=None, n_orient=1, max_iter=1000, tol=1e-06, init_gamma=None, init_lambda=None, learn_noise=True, verbose=False, normalization=False, track_history=True, logger=None, **kwargs)[source]#

BMN estimate with optional sLORETA normalization and optional adaptive common-noise learning.

Return type:

Dict[str, Any]

Supports#

n_orient = 1 -> fixed (EEG or MEG) n_orient = 2 -> reduced free MEG n_orient = 3 -> free EEG

Notes

  • posterior_mean and posterior_cov are returned in the original coefficient space.

  • gamma is the learned common scalar hyperparameter in the internal optimization parameterization, so it should be treated mainly as a diagnostic quantity, especially when normalization=True.