mne.epochs.
average_movements
(epochs, head_pos=None, orig_sfreq=None, picks=None, origin=’auto’, weight_all=True, int_order=8, ext_order=3, destination=None, ignore_ref=False, return_mapping=False, mag_scale=100.0, verbose=None)[source]¶Average data using Maxwell filtering, transforming using head positions.
Parameters:  epochs : instance of Epochs
head_pos : array  tuple  None
orig_sfreq : float  None
picks : arraylike of int  None
origin : arraylike, shape (3,)  str
weight_all : bool
int_order : int
ext_order : int
regularize : str  None
destination : str  arraylike, shape (3,)  None
ignore_ref : bool
return_mapping : bool
mag_scale : float  str
verbose : bool, str, int, or None


Returns:  evoked : instance of Evoked

Notes
The Maxwell filtering version of this algorithm is described in [R41], in section V.B “Virtual signals and movement correction”, equations 4044. For additional validation, see [R42].
Regularization has not been added because in testing it appears to decrease dipole localization accuracy relative to using all components. Fine calibration and crosstalk cancellation, however, could be added to this algorithm based on user demand.
New in version 0.11.
References
[R41]  (1, 2) Taulu S. and Kajola M. “Presentation of electromagnetic multichannel data: The signal space separation method,” Journal of Applied Physics, vol. 97, pp. 124905 110, 2005. 
[R42]  (1, 2) Wehner DT, Hämäläinen MS, Mody M, Ahlfors SP. “Head movements of children in MEG: Quantification, effects on source estimation, and compensation. NeuroImage 40:541–550, 2008. 