mvpa_iterate: Iterate MVPA Analysis Over Multiple ROIs

In bbuchsbaum/rMVPA: Multivoxel Pattern Analysis in R

Iterate MVPA Analysis Over Multiple ROIs

Description

Performs multivariate pattern analysis (MVPA) across multiple regions of interest (ROIs) using batch processing and parallel computation.

Usage

mvpa_iterate(
  mod_spec,
  vox_list,
  ids = 1:length(vox_list),
  batch_size = as.integer(0.1 * length(ids)),
  verbose = TRUE,
  processor = NULL,
  analysis_type = c("searchlight", "regional")
)

Arguments

mod_spec	An MVPA model specification object containing the dataset to analyze, compute_performance (logical indicating whether to compute performance metrics), and return_predictions (logical indicating whether to return predictions).
vox_list	A list of voxel indices or coordinates defining each ROI to analyze.
ids	Vector of identifiers for each ROI analysis. Defaults to 1:length(vox_list).
batch_size	Integer specifying number of ROIs to process per batch. Defaults to 10% of total ROIs.
verbose	Logical indicating whether to print progress messages. Defaults to TRUE.
processor	Optional custom processing function. If NULL, uses default processor. Must accept parameters (obj, roi, rnum) and return a tibble.
analysis_type	Character indicating the type of analysis. Defaults to "searchlight".

Details

The function processes ROIs in batches to manage memory usage. For each batch:

1. Extracts ROI data from the dataset.

2. Filters out ROIs with fewer than 2 voxels.

3. Processes each ROI using either the default or custom processor.

4. Combines results across all batches.

Value

A tibble containing results for each ROI with columns:

result

List column of analysis results (NULL if return_predictions=FALSE).

indices

List column of ROI indices used.

performance

List column of performance metrics (if computed).

id

ROI identifier.

error

Logical indicating if an error occurred.

error_message

Error message if applicable.

warning

Logical indicating if a warning occurred.

warning_message

Warning message if applicable.

bbuchsbaum/rMVPA documentation built on June 10, 2025, 8:23 p.m.