These values were averaged per cluster, and, subsequently, over
all clusters in a network. This provided us with 4 Zsep values for each
network: for young and older adults, in the easy and difficult
learning conditions. Note that for each network, the Zsep value is
always significantly different from 0 in the condition and age group
that were used to define a network because all voxels that were
included to create the network showed a significant effect (e.g., in
the easy-negative network in young adults, the Zsep value of young
adults in the easy condition is necessarily significant below 0).
Next, to examine the effects of learning condition and age, we
calculated the normalized distance between the Z-transformed
correlation values: