Electrical counts had no clear adjusted link to response change agreement in 87 C. elegans recordings

We tested whether electrical links mark equal changes across repeated neuron stimulation. We used a fixed C. elegans event export and all 302 neuron slots in five separate wiring maps. The primary Cook map had 3,671 chemical links and 1,091 electrical pairs. A bounded score compared changes in two target neurons between trials with the same source and animal. The panel had 323,928 animal-source-target-pair rows from 248 repeated conditions in 87 animals. Full controls included cell identity, distance, chemical wiring, shared partners and recording quality. The electrical coefficient was 0.00837 per log(1+section count). Its 95% animal-cluster interval was -0.00698 to 0.02372. Complete animal bootstrap refits also gave an interval across zero. Added descriptive fits showed a 78.9% decrease from the coefficient with animal effects alone. The White whole map gave a positive secondary estimate; both separate adult maps had intervals across zero. A stricter missing-data rule retained 51 animals and gave a wide interval. Thus the primary score gave no clear added electrical term. This does not show that electrical coupling has no function. The study extends an earlier SCI repeat-correlation test with a different score and larger panel. Processed missing values, sparse repeats, absent red-channel controls and anatomy from other animals limit inference. The package preserves code, full neuronal graphs, event arrangements, pair summaries, fitted states, all controls, source records and figures. Raw-source recovery matched all 104,079 input event values. A row solver agreed within 1.3e-12. A clean repeat matched 34 model files and two derived tables.

connectome · electrical synapses · C. elegans · repeated stimulation · trial variability · calcium imaging · reproducible benchmark