Get Covariance/Correlation for SVD

Main function for preparing the cross-block matrix for SVD decomposition. Handles all 6 PLS methods with appropriate mean-centering and correlation modes. Ported from MATLAB rri_get_covcor.m

Usage

pls_get_covcor(
  method,
  stacked_datamat,
  stacked_behavdata = NULL,
  num_groups,
  num_subj_lst,
  num_cond,
  bscan = seq_len(num_cond),
  meancentering_type = 0L,
  cormode = 0L,
  datamat_reorder = NULL,
  behavdata_reorder = NULL,
  datamat_reorder_4beh = NULL,
  compute_smeanmat = FALSE
)

Arguments

method

PLS method (1-6)

stacked_datamat

Stacked brain data matrix

stacked_behavdata

Stacked behavior data matrix (methods 3-6)

num_groups

Number of groups

num_subj_lst

Vector of subjects per group

num_cond

Number of conditions

bscan

Conditions for behavior block (multiblock)

meancentering_type

Mean-centering type (0-3)

cormode

Correlation mode (0, 2, 4, 6)

datamat_reorder

Optional reordering indices for datamat

behavdata_reorder

Optional reordering indices for behavdata

datamat_reorder_4beh

Optional reordering for behavior block datamat

compute_smeanmat

Logical, compute stacked_smeanmat for brain-score CIs

Value

List with:

datamatsvd

Cross-block matrix for SVD

datamatsvd_unnorm

Unnormalized version (multiblock)

datamatcorrs_lst

Correlation matrices by group (behavior PLS)

stacked_smeanmat

Matrix for brain score CIs (task PLS with bootstrap)