With independent site states, equal rates and unit weights, eight 162-cell C. elegans panels gave maximum readout variance ratios of 2.12–2.60 against independent contacts

We tested how shared chemical synapse sites change uncertainty in linear input count readouts. We used the same 162 embryonic neuron names in eight C. elegans brain maps. The primary model gave each site an independent binary event, with a uniform probability and unit contact weights. The comparison gave each contact an independent event. Both models gave every target the same full count distribution. Generalized covariance eigenvalues gave exact bounds over all real readout weights. Per-map maximum variance ratios were 2.118–2.601. Minimum ratios were 0.122–0.386. Count sum ratios were 2.069–2.368. The maximum was only 2.37–12.13% above the count sum ratio. A source mixture preserved source correlation in both models. At correlation 0.1, adult count sum ratios fell to 1.127 and 1.120, while maximum ratios were 1.455 and 1.505. Maximum weights selected in one adult gave ratios 2.201 and 2.061 in the other at zero source correlation. At correlation 0.1, these transfer ratios were 0.982 and 1.252. Thus separate model bounds do not establish a common readout. Partial site sharing scaled each variance excess exactly. Selection checks and site size weights retained substantial differences. Exact rational enumeration, independent pair counts, direct simulation and a fresh source rebuild checked the work. The package preserves new code, data facts, source hashes, sufficient matrices, attaining weights, figures and an inspection tool. These are conditional count-model results. They do not measure release correlations, neural activity, stimulus information or animal population effects.

Caenorhabditis elegans · connectome · polyadic synapse · count covariance · linear readout · shared event · source correlation · variance bound · reproducible research