test_regress: Regression tests

In andy1764/FCHarmony: Functional Connectivity Harmonization

Regression tests

Description

Examine if harmonization affects batch and covariate associations. Options to regress edge weights and mean edge weights within and between subnetworks.

Usage

test_regress(
  ...,
  bat = NULL,
  mod = NULL,
  roi.names = NULL,
  labs = c("Raw", "Out"),
  tests = c("Elem", "Group"),
  fisher = TRUE,
  to.corr = FALSE,
  debug = FALSE
)

Arguments

...	p x p x n covariance or correlation matrices where p is the number of ROIs and n is the number of subjects. These are generally unharmonized or harmonized datasets and should have the same dimensions.
bat	Factor (or object coercible by as.factor to a factor) of length n designating batch IDs.
mod	Optional design matrix of covariates to regress on, usually from the output of model.matrix.
roi.names	Vector of names for regions of interest. For "Group" tests, roi.names defines subnetworks of interest. If NULL, the dimension names of each dataset are used.
labs	Vector of labels for the harmonization methods, in order of the inputted datasets.
tests	Vector of tests to apply. "Elem" refers to regression on each edge weight. "Group" refers to regression on mean edge weights within and between each subnetwork, as defined by roi.names.
fisher	Whether to z-transform input FC matrices
to.corr	Logical, whether input should be forced to be a correlation matrix using cov2cor
debug	Whether to return intermediate objects for debugging

andy1764/FCHarmony documentation built on April 4, 2022, 10:41 a.m.