Subsample Searchlight Centers — subsample

Selects a representative subset of searchlight centers for permutation runs, optionally stratifying by the number of features per center.

Usage

subsample_centers(
  dataset,
  searchlight,
  n_centers = NULL,
  fraction = 0.1,
  stratify_by = "nfeatures",
  redundancy_map = NULL,
  seed = NULL
)

Arguments

dataset: An mvpa_dataset.
searchlight: A searchlight iterator (list of integer voxel-index vectors) as returned by get_searchlight().
n_centers: Optional integer count of centers to select. If NULL, derived from fraction.
fraction: Numeric in (0, 1]. Fraction of all centers to select when n_centers is NULL.
stratify_by: Character. Covariate used for stratification ("nfeatures" only for now).
redundancy_map: Optional named numeric vector of per-center redundancy values (not used unless stratify_by = "redundancy").
seed: Optional integer seed.

Value

A list with:

center_ids: Integer vector of selected center IDs.
vox_list: Subsetted searchlight list.
covariates: A data.frame with at least column nfeatures.