Trial, wiring and distance count rules forced 88 of 94 two-way detections in a C. elegans atlas of 111 neurons

We tested how graph control rules change the reference for two-way response detections in one published C. elegans signal atlas. The main set had 2,470 neuron pairs among 111 neurons. Each direction had a finite q value and at least five finite trial values. There were 500 detected directions at q < 0.05 and 94 two-way pairs. Controls that kept each neuron's input and output detection counts gave a mean of 34.43 two-way pairs. Keeping these counts within trial and direct-wiring groups raised the mean to 64.21. An added exploratory control used both directions' trial groups and soma-distance groups. Its rules forced 469 detected directions and 88 recorded two-way pairs to stay present. We enumerated all 204,288 feasible graphs. Their exact mean was 92.5708 pairs. Their central 95 percent range was 90 to 95. The graph match therefore had little freedom to test the recorded two-way pattern. These results concern a fixed pooled atlas and stated graph rules. They do not prove physiological feedback or assign a causal share to anatomy. The package contains compact numerical inputs, source versions and hashes, new code, exact graph results, sampling checks, figures, and repeat instructions.

Caenorhabditis elegans · signal propagation · connectome · reciprocity · fixed margins · null models · trial coverage · exact enumeration