Significance measure P
of selected quadruplet over-representation for different thresholds. Multiple hypothesis testing corrected p-values for the quadruplet over-representation in datasets 1 and 2 (n = 1000). The patterns shown satisfy the following selection criteria: There must be a significant value P
< 0.1 for at least one Θ and the count of this pattern in C. elegans must be at least 5. All random matrices in the ensemble (n = 1000) preserve the number of uni- and bi-directional connections for each neuron as well as the count in all triplet classes for the whole network.