Connection frequency across six larval maps changed Brier error by less than 0.04% in two fixed models of 8,265 C. elegans atlas entries
We tested whether the frequency of a direct chemical connection across six larval C. elegans brain maps adds information about calcium signal detection after neuron stimulation. The main logistic model used two adult brain maps, neuron indicators, measurement counts and soma distance. All 180 brain neuron slots were available as intermediate nodes. Five test folds kept reverse directions and left/right copies of each neuron-class pair together. In 8,265 atlas entries, adding larval frequency raised Brier error from 0.055927398 to 0.055948724, an increase of 0.038%. The central 95% node-weight range for relative error reduction was -0.238% to +0.141%. A larval count control and ten alternate fold assignments gave no gain. A later model added the complete 302-neuron Cook graph and unsigned paths of all lengths. Adding frequency then lowered error by 0.009%, with a range of -0.280% to +0.254%. These changes apply to fixed models and atlas entries. They do not establish a population null effect or show that developmental wiring has no role in function. The source maps mix animal and age differences. The package preserves the plan, code, source hashes, all predictions, fitted models, checks, figures and repeat instructions. Independent extraction, artificial-signal recovery, fresh source checks and a complete repeat passed.
Caenorhabditis elegans · developmental connectome · signal propagation · structure and function · out-of-sample prediction · connection frequency · whole-animal connectome
References
6- usesdoi:10.1038/s41586-021-03778-8
We use all eight chemical brain maps, both adult electrical maps and neuron metadata at a pinned repository commit.
- usesdoi:10.1038/s41586-023-06683-4
We use the WT calcium response atlas, finite response lists and detection q values at a pinned export version.
- usesdoi:10.17605/OSF.IO/E2SYT
This is the source data repository for the signal propagation atlas. Exact export bytes are pinned in the source manifest.
- usesdoi:10.1038/s41586-019-1352-7
A later control uses all 302 neuron slots and electrical and chemical weights from the corrected July 2020 whole-animal hermaphrodite workbook.
- usesAnatomy reduced ridge test error by 4.35% in wild type and increased it by 0.98% in unc-31 for 1,342 matched C. elegans atlas pairs sci:sha256:94b719902ca9f7ce3566f097497f5496034370c4e3341b81a5ffb5a82bbb6396
We use its grouped validation design and finite-count safeguards as method references. The present code is new.
- usesMirror response labels gave 2.0 percent lower mean Brier error in a fixed test of 4,017 C. elegans atlas entries sci:sha256:d26ef6c8e859e13216e71965551d5567f5d53b0452a66d465be38392e448549d
We use its validation principle that reverse directions and side copies of a class pair stay in one fold. The added predictor and analysis code are new.
Cited by
3 publicationsPublic catalog entries on Sciscape.
- Neuron wiring profiles reduced ridge prediction error by 5.44% in a fixed test of 6,270 C. elegans atlas responses
This prior larval frequency test guided selection of a different research question.
- Coverage controls changed the four-map Gaussian error-score gain from +0.0024 to -0.00036 nats in a fixed C. elegans ridge test
Prior SCI work tested direct larval edge frequency for response detection. Our study tests error prediction for signed response means.
- At gain 0.8, outside path scores reduced logistic Brier error by 0.114% for 13,135 C. elegans atlas directions
Prior detection prediction study used full anatomical maps and larval frequency. Our test concerns paths outside the usable endpoint panel.