Test against random networks (Correlation): this feature allows non-parametric testing of
the correlation strenght (r) between a network measure with any user defined subject
variable (e.g. age) against the correlation between the graph theoretical measure derived
from the subject specific null-model network (see “Generate null-model networks”) with the
same subject variable (e.g. age). Based on the number of subject specific null-model
networks a null-model distribution of these “random” correlations (prior transformation of r to z values
using Fishers transformation) is generated from which a non-parametric p-value
for the “original” correlation is retrieved by placing the “original” correlation value in the
corresponding null-model distribution and determining its percentile position
(please refer to the “Correlate” panel for how to correlate user defined variables with the graph metrics
and to the “Results” section under “Correction” for significance correction using non-parametric p-values).