Plot the cluster-test grid for a pairwise discriminability report

Renders one cluster t-map (or signed TFCE map) per pair, laid out in a square-ish grid. Each panel shows the per-pixel statistic with black contours bounding FWE-significant clusters (or FWE-corrected pixels under TFCE). Color convention matches the rest of the package: blue = first condition larger, red = second condition larger.

Usage

# S3 method for class 'rcisignal_rel_pairwise_report'
plot(x, ..., ncol = NULL, max_pairs = 12L, base_image = NULL, alpha_max = 0.7)

Arguments

x

A run_discriminability_pairwise() result.

...

Reserved for future use.

ncol

Optional integer. Columns in the panel grid. When NULL (default), ceiling(sqrt(n_pairs)) is used.

max_pairs

Integer. Above this many pairs a warning is emitted (panels become illegible). Default 12L. The grid is still drawn; pass max_pairs = Inf to silence the warning.

base_image

Optional. Either a numeric matrix (nrow x ncol, grayscale, values in 0-1) or a path to a PNG/JPEG file. When supplied, each per-pair t-map is composited on top of the grayscale base; otherwise panels render on a flat background.

alpha_max

Numeric in [0, 1]. Maximum opacity of the t-map overlay at the color-scale top when base_image is supplied. Default 0.7.

Value

Invisibly the input x.

Details

To compare overall magnitudes across pairs on a shared axis, pass the per-pair $dissimilarity children to plot_dissimilarity_grid() instead.

See also

plot_dissimilarity_grid() for a shared-axis comparison of bootstrap dissimilarity distances across pairs.