Skip to contents

Extract Per-ROI Out-of-Fold Predictions from Feature RSA Results

Source: R/feature_rsa_connectivity.R
feature_rsa_predictions.Rd

Pull the out-of-fold predicted patterns (`Yhat`) retained by feature_rsa_model(..., return_predictions = TRUE). Each ROI keeps the merged prediction matrix, the matching observed neural patterns, the observation order, and the outer-fold id so later scoring cannot accidentally treat a training target as a candidate.

Usage

feature_rsa_predictions(x)

Arguments

x

A regional_mvpa_result returned by run_regional() for a feature_rsa_model, or a tibble/data frame that already has columns roinum and predicted.

Value

A tibble with one row per ROI and columns:

roinum

ROI id.

n_obs

Number of observations in the retained matrices.

observation_index

List-column of the observation order used for the matrices.

fold_id

List-column identifying the outer test fold for each observation. Identification and geometry scoring must stay inside these groups.

voxel_index

List-column of the spatial indices for matrix columns.

predicted

List-column of out-of-fold `Yhat` matrices (`n_obs` by `n_voxels`).

observed

List-column of the matching observed pattern matrices.

See also

feature_rsa_model, feature_rsa_rdm_vectors

Examples

# \donttest{
  set.seed(79)
  sample <- gen_sample_dataset(c(4, 4, 4), nobs = 24, blocks = 3)
  Fmat <- matrix(rnorm(24 * 6), 24, 6)
  des <- feature_rsa_design(
    F = Fmat,
    labels = paste0("t", seq_len(24)),
    max_comps = 3,
    block_var = sample$design$block_var
  )
  mdl <- feature_rsa_model(
    sample$dataset, des, method = "pca",
    ncomp_selection = "max",
    return_predictions = TRUE
  )
  region_mask <- neuroim2::NeuroVol(
    sample(1:2, length(sample$dataset$mask), replace = TRUE),
    neuroim2::space(sample$dataset$mask)
  )
  res <- run_regional(mdl, region_mask)
#> INFO [2026-09-09 12:45:29] 
#> MVPA Iteration Complete
#> - Total ROIs: 2
#> - Processed: 2
#> - Skipped: 0
#> INFO [2026-09-09 12:45:30] run_regional: 2 ROIs processed (success=2, errors=0)
  preds <- feature_rsa_predictions(res)
  dim(preds$predicted[[1]])
#> [1] 24 34
# }