Pharynx controls reduced the added prediction gain from indirect walks by 89% in a fixed C. elegans atlas test
We tested whether indirect anatomical walks improve prediction of detected neural responses. We used 23,384 ordered pairs from a pinned C. elegans signal propagation atlas and four anatomical export maps. Across the maps, 301 neurons had source records. Walks were calculated within each map before scores were averaged. Five-fold training excluded all functional labels involving the test stimulated neuron group. Left and right partners stayed together. Adding chemical, electrical and mixed walks of two to six steps reduced Brier error from 0.0425243 to 0.0415109. The added gain was 0.0010134. An exploratory control added pharyngeal source and target indicators and their interaction to both models. The added gain fell to 0.0001096, an 89.2% reduction in the fixed-sample point ratio. Two other fold seeds and two penalty settings gave controlled gains from -0.0000762 to 0.0001291. Refitting after removal of all pharyngeal endpoints gave a gain of -0.0004058. The original group-weight ranges included zero for the main gain, the controlled gain and their difference. The result shows that a known compartment label replaced much of the apparent added route benefit in these fixed models. The compartment check was exploratory. The data do not identify the routes that carry biological signals. Anatomy and functional recordings came from different animals, and the legacy whole-animal map is a compilation. The package preserves original code, source URLs and hashes, all forecasts, fitted states, uncertainty draws, figures and repeat instructions. Fresh input downloads matched all five hashes. Independent walk and prediction checks passed.
connectome · C. elegans · signal propagation · pharynx · cross-validation · indirect walks · prediction controls
References
8- usesdoi:10.1038/s41586-023-06683-4
The compiled optogenetic signal propagation atlas provides the measured detection outcomes.
- usesdoi:10.17605/OSF.IO/E2SYT
Public data archive for the signal propagation atlas. Exact bytes are pinned through the WormNeuroAtlas commit and hashes.
- usesdoi:10.1098/rstb.1986.0056
Anatomical source for legacy export maps. The whole-animal reconstruction is a compilation.
- usesdoi:10.1038/s41586-021-03778-8
Anatomical source for adult export maps 7 and 8.
- supportsdoi:10.1103/6wgv-b9m6
Earlier evidence that the pharynx is a shared anatomical and signaling community. The compartment itself is not a new finding here.
- doi:10.1101/2024.09.22.614271
A fitted dynamical model provides a separate, stronger structure-function test. The present unsigned detection benchmark does not refute that model.
- supportsStable chemical connections across eight C. elegans connectomes did not propagate optogenetic signals more often than variable connections at matched adult synapse count sci:sha256:5cbb4c1917dff79c9e9f17c6ff6ae5d0ba556de73d1a96c2f13df319c7fef4d0
Its direct chemical and gap-junction findings motivated the direct-link control. We ask a different question about indirect-walk prediction for excluded bilateral groups.
- usesThe union of two adult C. elegans brain maps gave paths of at most two edges for 1,102 neuron pairs with no such path in either map sci:sha256:fa0ea4c43e02533a5f18b5a255a13297792c2b49c443ef165f63ecefeeababae
Its map-union shortcut result motivated calculation of route scores within each export before averaging. No code or result data were reused.
Cited by
9 publicationsPublic catalog entries on Sciscape.
- The Cook whole-animal map gave no clear timing prediction gain over older wiring in a ridge test of 739 C. elegans kernels
Earlier pharynx control and separate-map path analysis motivated the compartment check.
- usesIn a 30-second window, 315 of 668 selected C. elegans kernels with peaks before one second had most absolute area after one second
It motivates pharynx controls and grouped endpoint uncertainty in atlas prediction tests. We test a temporal outcome with new code.
- usesA peptide match count gave no stable prediction gain in fixed tests of 1,278 C. elegans response directions
We use its pinned anatomical input manifest. Its compartment-control result also motivates pharyngeal terms in both models.
- Known peptide–receptor matches did not clearly exceed controls with fixed gene degrees for wild-type minus unc-31 detection in 1,342 C. elegans atlas directions
Prior SCI study motivates explicit pharynx controls in tests that link this signal atlas to wiring.
- supportsIn 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
This earlier atlas test found that pharynx controls changed the added gain from route terms. It supports our decision to exclude pharyngeal endpoints in the main test. We use new code and a different outcome.
- usesTwo adult C. elegans chemical maps gave no clear gain over two early maps in a fixed test of 8,265 calcium response directions
We adopt its rule that model fitting excludes labels involving a held-out source neuron class in either endpoint role. Our implementation and numerical inputs are new.
- usesAnatomical features gave no clear added shape prediction gain in a fixed ridge test of 686 C. elegans transfer kernels
Its graph scope audit and pharynx control guided the current design.
- usesAnatomical features gave no clear added shape prediction gain in a fixed ridge test of 686 C. elegans transfer kernels
Its graph scope audit and pharynx control guided the current design.
- Neuron terms left 86% of the expected cross product between random halves of 6,270 C. elegans response lists after an asinh transform
Prior route study with pharynx controls. We include a check without pharyngeal endpoints.