Evaluate model performance for feature RSA

Computes correlation-based metrics (diag correlation, mean correlation, voxel correlation), MSE, R^2, and optionally performs permutation tests (via a helper function).

Usage

evaluate_model.feature_rsa_model(
  object,
  predicted,
  observed,
  nperm = 0,
  save_distributions = FALSE,
  ...
)

Arguments

object

The feature RSA model

predicted

Matrix of predicted values (from feature space F to voxel space X)

observed

Matrix of observed values (actual voxel space X)

nperm

Number of permutations for statistical testing (default: 0, no permutation)

save_distributions

Logical indicating whether to save full permutation distributions

...

Additional arguments

Value

A list containing performance metrics and optional permutation results