Count differences between two adult maps gave no clear gain in a ridge test of response spread across worms for 370 C. elegans neuron pairs

We tested whether differences in chemical site counts between two adult C. elegans brain maps add information about calcium response spread across measured worms. We joined pinned Witvliet adult maps to event-labelled responses from DANDI 001075 version 0.240930.1859. We averaged repeated events within each worm and measured the variance of worm means for 370 directed pairs with at least three worms. The panel used 2,460 target observations from 97 worms. A nested ridge test kept source cell groups out of fitting. Controls included mean adult count, cell terms, electrical counts and measured response and assay summaries. Adding squared relative count mismatch reduced test error from 0.914173 to 0.913076, or 0.120%. The descriptive source-group range was -0.024% to 0.282%; the target-group range was -0.089% to 0.316%. Both include zero. Removing pairs with one total site gave a 0.456% error increase. A stricter test excluded test groups from both endpoint roles and gave a 0.045% increase. Two hundred refits with shared positive worm weights gave a gain range of -0.114% to 0.527%. Repeated stimuli gave an excess variance estimate of 0.0682 above sampling variance in 103 wired pairs, against within-worm variance 0.5138. Its source-group range included zero. This excess can include state and assay effects. The evidence gives no clear added mismatch gain in this fixed test. It does not establish equal function across different wirings. Only two anatomical animals were available, and they differed from the response animals. The package preserves inputs, source hashes, code, models, predictions, figures and repeat instructions. Raw spot checks matched 1,867 observations. A clean repeat matched 138 numerical files and 920 arrays.

connectome · Caenorhabditis elegans · signal propagation · animal variation · synapse counts · response variance · optogenetics · reproducibility