brainGraph (version 1.0.0)

permute.group: Permutation test for group difference of graph measures

Description

This function draws permutations from linear model residuals to determine the significance of between-group differences of a global or vertex-wise graph measure. This function is intended for cortical thickness networks (in which there is only one graph per group), but can be extended to other types of data. If you would like to calculate differences in the area-under-the-curve (AUC) across densities, then supply a vector to the densities argument.

Usage

permute.group(permSet, densities, resids, level = c("graph", "vertex", "lobe",
  "other"), atlas, measure = c("btwn.cent", "degree", "E.nodal", "knn",
  "transitivity", "vulnerability"), .function = NULL)

Arguments

permSet

Numeric matrix of the set of permutations to loop through; the number of rows equals the desired number of permutations and the number of columns equals the total number of subjects across groups

densities

Numeric vector of graph densities

resids

A data table of the residuals (from get.resid)

level

A character string for the attribute level to calculate differences; either 'graph', 'vertex', 'lobe', or 'other'

atlas

Character string of the atlas name

measure

A character string, either 'btwn.cent', 'degree', 'E.nodal', 'knn', or 'transitivity', 'vulnerability' (specific to the vertex level)

.function

A custom function you can pass (if level is 'other')

Value

A data table with values for group differences in modularity, global efficiency, clustering, average path length, and assortativity (etc.)

Details

The graph "level" will calculate modularity (Louvain algorithm), clustering coefficient, average path length, degree assortativity, global efficiency, lobe assortativity, and edge asymmetry. The vertex "level" will calculate a vertex-wise measure. Currently, you can choose betweenness centrality, degree, nodal efficiency, k-nearest neighbor degree, transitivity, or vulnerability. The lobe "level" is intended to test for group differences in number of inter-lobar connections, e.g. from the temporal lobe to the rest of the brain. The other "level" allows you to pass your own function to do permutations with. This is useful if you want to calculate something that I haven't hard-coded (e.g. number of hubs between groups). It must take as its own arguments: "g1", "g2", and "density".

See Also

centr_betw, vulnerability, count_interlobar, edge_asymmetry, efficiency

Examples

## Not run: ------------------------------------
# m <- get.resid(all.thick, covars)
# myPerms <- shuffleSet(n=nrow(m$resids), nset=1e3)
# out <- permute.group(myPerms, densities[N], m$resids, 'graph', atlas='dk')
# out <- permute.group(myPerms, densities[N], m$resids, 'vertex')
# out <- permute.group(myPerms, densities[N], m$resids, 'other',
#   .function=myFun)
## ---------------------------------------------