Two electrical pairs each caused growth above finite weights in a 302-cell worm model with unit leak, EM section weights and chemical gain 0.95

We asked whether a reciprocal electrical contact can cause growth in a stable neural model. A steady voltage test gives a sufficient rule. Apply equal and opposite currents to two cells. If the steady voltage contrast r is negative, a real mode must grow above gap weight -1/r. This rule assumes a stable finite linear voltage system with positive capacitances and proper chemical filters. A signed model can lose stability earlier. We tested 1,091 electrical pairs in one 302-cell Cook worm model. Chemical weights came from EM section totals and unique receptor sign predictions. We omitted 1,883 edges without a unique sign. Each chemical matrix was divided by its largest eigenvalue real part. With unit leak and chemical gain 0.95, ALA-AVER and PHBL-PVCL had thresholds of 1.1294 and 9.4896 times leak. Direct spectra confirmed growth above both thresholds. No pair had negative static contrast at gains 0.5 or 0.8. Making unknown signs positive gave 21 cases at gain 0.95. Adding all gaps together kept each main model stable at all 82 sampled gains. Changes of source gains and weight rules altered the pair results. The package contains the proof, code, all pair scores, source audits, figures and a clean replay. These are model results. They are not measured worm responses or conductances.

gap junction · electrical synapse · linear stability · C. elegans · steady voltage response · chemical synapse · connectome · feedback · sign uncertainty