Skip to main content
. 2014 Oct 21;25(10):3743–3757. doi: 10.1093/cercor/bhu252

Figure 4.

Figure 4.

High numbers of connections occur more than chance in clusters of up to 6 neurons. Each plot shows how often a given number of connections occurred for clusters of up to 6 neurons. Curves of dotted lines show connections expected by chance. In chance shuffling, the numbers of directed inputs, directed outputs, and bidirectional connections was preserved. Curves of sold lines show connections observed in data. In general, solid lines tended to depart from dotted lines at high numbers of connections, suggesting nonrandomness of connections. Top row (a–d) shows results for networks constructed from effective connectivity (TE); bottom row (e–h) shows results for networks constructed from patch-clamp experiments (synaptic connections; Perin et al. 2011). Small networks inset in (ad) show schematic example clusters. Number of neurons in each cluster goes from 3 (left) to 6 (right). Note qualitative similarity between effective connections measured by TE and synaptic connections assessed by patch-clamp. Larger groups of neurons tended to have more connections than chance (*P < 0.01, **P < 0.001; z-test).