eelbrain.pipeline.Pipeline.load_trf

Pipeline.load_trf(x, tstart=0.0, tstop=0.5, *, estimator='boosting', data=None, mask=None, samplingrate=None, filter_x=False, path_only=False, **state)[source]

Load (or compute) the TRF for a model and the current subject

Parameters:

  • x (str) – Model (e.g. 'gammatone + word').

  • tstart (float) – Start of the TRF in seconds.

  • tstop (float) – Stop of the TRF in seconds.

  • estimator (str) – Name of the estimator in estimators (default 'boosting'). Estimator-specific parameters (basis, delta, mu, …) are set on the Estimator object.

  • data (str) – Sensor-space data kind to fit: a sensor type ('meg'/'mag'/'grad'/'eeg'), optionally aggregated (e.g. 'eeg.rms'/'eeg.mean'). The default (None) uses default_data. The analysis space is set by the inv state (inv='' for sensor space; a non-empty inverse for source space); in source space leave data unset. NCRF requires inv='' and leaves data unset.

  • mask (str) – Parcellation to mask source-space data (not implemented yet).

  • samplingrate (int) – Samplingrate in Hz for the analysis.

  • filter_x (bool | Literal['continuous']) – Filter predictors like the M/EEG data (see load_predictor()).

  • path_only (bool) – Return the path to the cache file instead of loading the TRF.

  • ... – State parameters.