Fit the core GLM on a transformed time-series matrix — fit_glm_on_transformed

This helper performs IID, non-robust OLS inference. It rejects controls that request AR or robust fitting; use fit_glm_with_config() for those models.

Usage

fit_glm_on_transformed_series(
  model,
  Y,
  cfg = NULL,
  dataset = NULL,
  strategy = "external",
  engine = "external"
)

Arguments

model: An fmri_model describing the design.
Y: Numeric matrix with nrow(Y) time points and columns matching voxels.
cfg: Optional fmri_lm_control; defaults to fmri_lm_control() and must request IID noise and no robust fitting.
dataset: Optional dataset whose feature axis describes the columns of Y. It is never inferred from model$dataset, because a transformed response can have a different feature space.
strategy: Character label recorded on the returned object. Defaults to "external".
engine: Character label indicating the engine that produced the fit. Defaults to "external".

Value

An object of class fmri_lm.