One sign exception is needed for 91 selected worm responses, but stricter samples permit cell signs

We tested whether separate source and target cell signs explain mean response signs on anatomical links in C. elegans. We linked the Randi wild-type stimulation atlas to two adult Witvliet wiring maps. At q below 0.05 and at least three occurrences, 91 selected chemical-linked responses required exactly one pair-specific sign exception. An eight-edge negative-product cycle and a feasible fit proved this minimum without relying on an optimizer bound. A uniform additive mean-response change of 0.041506 delta F/F was necessary and sufficient to permit the model with zero responses allowed. Requiring chemical links in both adults, at least five occurrences, or q below 0.01 removed the conflict. The both-adult role graph had no cycles and could not test this model. No measured response-error bound was available. Different animals, pooled responses, sparse constraints and indirect propagation prevent a biological sign-rule claim. Code, exact certificates, source hashes, evidence tables and a figure support reproduction.

C. elegans · connectome · effective response signs · signed graph · cycle certificate · cell factors