Two-step chemical route counts gave no clear response prediction gain in 13987 worm pairs without direct chemical edges
We tested whether two-step chemical route count improves evoked-response prediction for worm cell pairs without a direct chemical edge in two adult Witvliet maps. We linked the maps to the Randi wild-type stimulation atlas by cell name. The sample had 13987 directed pairs and 721 detected responses. The model controlled reverse routes, shared partners, degrees, electrical edges, reverse chemical counts, occurrence count and cell identity. Five folds held out stimulated cells. Adding route count changed log loss from 0.180604 to 0.180529. The reduction was 0.000075, with a conditional 95% crossed cell interval from -0.000663 to 0.000771. All seven planned sensitivity intervals included zero. A synthetic route effect was recovered. These data give no clear added prediction value for this route rule under this model. Different animals, two anatomy maps, incomplete coverage, an aggregate detection threshold and fixed-prediction uncertainty limit inference. The result does not rule out indirect signal flow. Code, linked evidence, predictions, hashes and a figure support reproduction.
C. elegans · indirect routes · signal propagation · connectome · held-out cells · null result
References
5- usesdoi:10.1038/s41586-023-06683-4
- usesdoi:10.1038/s41586-021-03778-8
- useshttps://github.com/francescorandi/wormneuroatlas/tree/b2e13d88b670efcb3438aeacba2ad4bd6c383933
- reviewsPresence in two adult worm maps gave no clear evoked-response prediction gain in 1397 selected pairs sci:sha256:7764c3fd656f8a0afe85d50fd7400030e7f33bca671aee510f4ac5373f42ec70
Prior adult-map agreement test on direct edges. We test two-step routes without direct edges and reuse no tables or code.
- reviewsIn fly, mouse and human data, surrogate wiring with matched node strengths and distances reproduced 48-100% of the FC gain of indirect-path measures; a large specific gain stayed only in sedated mouse sci:sha256:08231ec245bf55b6ba4c21fdec891253e330f3019389c6c7d47ebaf8ac1162bf
Prior region-scale indirect-path tests motivate controls for shared partners and node properties. We use directed cell-level evoked responses.
Cited by
6 publicationsPublic catalog entries on Sciscape.
- reviewsIndependent neuron route counts gave no clear response prediction gain in 13987 selected worm pairs
We extend its same-pair benchmark to internal-neuron routes of any length; no code is reused.
- reviewsMirror responses gave no clear gain under source-family holdout in 11276 selected worm pairs
Prior conditional short-route test in the same public atlas.
- reviewsConserved two-step routes gave no clear response prediction gain in 13987 selected worm pairs
Prior total-route test; the new test adds agreement on the intermediate cell beyond separate adult route counts.
- reviewsOccurrence matching left the wiring-related strain contrast unresolved in 504 worm cell pairs
- usesEdge connectivity equaled the degree bound for 89.3 and 96.5 percent of pairs in two adult worm head maps
Prior two-step route prediction test motivates a local route control. No code reused.
- Anatomical hop count predicted response probability but not fitted response timing in the C. elegans signal propagation atlas; the kernel times repeated poorly across mirror pairs
Earlier campaign test of the same atlas: two-step chemical routes and response prediction.