metricT: Directional Fragility Diagnostics for EEG Connectivity

Implements Metric-T, a permutation-based diagnostic that flags when the reported direction of a group difference in electroencephalography (EEG) functional connectivity depends on the choice of metric. Direction consistency (DC) converts the weighted phase lag index (wPLI) and magnitude-squared coherence (MSC), both derived from the same cross-spectrum, into the percentage of channel-pair features for which one group exceeds the other; Metric-T is T = DC(wPLI) - DC(MSC). Significance is assessed by aligned label permutation with marginal and Westfall-Young max-T multiplicity control. For small samples the reference set is enumerated completely: exact two-group tests, the complete two-group labelling space (in which cut-points, exclusions and batches are located on one distribution), and a stratified exact test for a continuous covariate. Core functions use base R only; a 'shiny' graphical interface is included. Method: Kimura (2026) <doi:10.1016/j.neuri.2026.100286>.

Version:	0.2.0
Depends:	R (≥ 4.0.0)
Imports:	stats, graphics, grDevices, utils
Suggests:	shiny, testthat (≥ 3.0.0)
Published:	2026-10-16
DOI:	10.32614/CRAN.package.metricT
Author:	Akira Kimura [aut, cre]
Maintainer:	Akira Kimura <a-kimura at paz.ac.jp>
BugReports:	https://github.com/akimura753/metricT/issues
License:	MIT + file LICENSE
URL:	https://github.com/akimura753/metricT
NeedsCompilation:	no
Language:	en-GB
Materials:	README, NEWS
CRAN checks:	metricT results

Documentation:

Reference manual:

metricT.html , metricT.pdf

Downloads:

Package source:	metricT_0.2.0.tar.gz
Windows binaries:	r-devel: metricT_0.2.0.zip, r-release: metricT_0.2.0.zip, r-oldrel: metricT_0.2.0.zip
macOS binaries:	r-release (arm64): metricT_0.2.0.tgz, r-oldrel (arm64): metricT_0.2.0.tgz, r-release (x86_64): metricT_0.2.0.tgz, r-oldrel (x86_64): metricT_0.2.0.tgz

Linking:

Please use the canonical form https://CRAN.R-project.org/package=metricT to link to this page.