In four C. elegans maps with two-way electrical links, route length had rank correlations of 0.03–0.10 with fitted mean kernel peak time

We tested whether a short anatomical route marks an early peak in a fitted neural response. We used four C. elegans anatomy exports and a pinned WT signal propagation atlas. Chemical links kept their direction; electrical links could carry routes in both directions. All available neural relays were retained in each map. Edges were not pooled across maps. The main test had 594 detected, non-self entries with at least three finite trial amplitudes and a nonzero mean kernel. It excluded AWC and pharyngeal endpoints. Route length had a Spearman rank correlation of 0.0466 with fitted mean kernel peak time. Entries with one, two or three edges had median times of 0.44, 0.43 and 0.43 seconds. Adding chemical-only and combined shortest route terms to a fixed ridge model raised held-out log-time squared error from 0.18094975 to 0.18105042. The error reduction was -0.056%. Its central 95% group weight range was -1.00% to +1.18% of baseline error. The predictions stayed fixed; this range does not describe variation across animals. Across the four maps, rank correlations were 0.025-0.104 for 550-594 entries. None of 18 planned checks or one later panel check gave a positive lower weight-range limit. A clean run reproduced 115 result and figure files. Independent graph, decimal-kernel and saved-model checks passed. The package preserves code, input hashes, the derived benchmark, fitted models, predictions, weight draws and figures. The measure is the peak of the full exported mean kernel. It is not conduction delay or the source study's corrected rise time. Anatomy and imaging came from different animals, and the tests share one atlas.

Caenorhabditis elegans · connectome · signal propagation · transfer kernel · peak time · shortest path · electrical synapses · reproducible benchmark