Figure 7

Reciprocity and correlation of structurally constrained networks. (a) Reciprocity as a function of FR for networks without self-connections and Dale’s principle with 1:1 Ex:In ratio. Reciprocity shows a parabolic relationship with FR, decreasing as FR approaches 0.5 spikes/time step. F networks show strong negative reciprocity, while T networks reciprocity is close to zero. (b) Pairwise correlation for T + F networks as a function of firing rate. The correlation was computed over pairs of excitatory neurons (Ex-Ex), pairs of inhibitory neurons (In-In), and pairs of one excitatory and one inhibitory neuron (Ex-In). Correlation has a maximum close to 0.5 spikes/time step. The In-In pairs show the highest correlations, followed by the Ex-In pairs. The Ex-Ex pairs show the lowest correlation. (c) pairwise correlation for F networks as a function of firing rate. The CC/FR relationship is similar to the one observed for T + F networks, although F networks correlation is displaced towards higher values. Mean ± SD are shown; n = 20 networks were constructed for each target FR and network type. Firing rates of excitatory or inhibitory neurons are displayed for Ex-Ex and In-In curves, respectively. For Ex-In curves, the average FR over all neurons is shown.