Wiring features reduced calcium response error by 1.17% beyond cell identity in a fixed test across 103 C. elegans worms
We tested whether wiring and response data from other animals improve calcium response predictions for a new worm. We used all 110 processed files in a fixed DANDI release. Usable data from 103 wild type C. elegans worms gave 41,156 animal and direction means for 15,668 directed neuron pairs. Five test folds kept each animal absent from all fitting and model selection. A ridge model used source and target identity. Added features from the corrected Cook chemical and electrical maps reduced equal-animal mean squared error by 1.17%. A descriptive 95% animal resample range was 0.50% to 1.94%, with predictions fixed. Pair response means from other animals gave a further 0.51% decrease, with a range of 0.13% to 0.89%. A second animal split and other signal and transform choices kept the wiring gain. A check added after the main result removed pharyngeal endpoints and bilateral homolog directions. The wiring gain then fell to 0.40%, with a range of 0.01% to 0.82%. On 38,518 matched rows, pre-stimulus changes gave no gain from either addition. Most response error remained. These results show a small prediction gain across archived animals under the tested measurement and model rules. They do not identify a transmission mechanism. The package contains animal and event labels, derived responses, full retained graphs, code, source hashes, fitted states, predictions, figures and repeat instructions.
connectome · Caenorhabditis elegans · signal propagation · animal validation · calcium imaging · structure and function
References
5- usesdoi:10.1038/s41586-023-06683-4
This study supplied the optogenetic activation and calcium recordings.
- usesdoi:10.48324/dandi.001075/0.240930.1859
We use all 110 processed segmentation files in this fixed data release.
- usesdoi:10.1038/s41586-019-1352-7
We use the hermaphrodite chemical and electrical maps corrected in July 2020.
- usesWiring profiles reduced error by 4.60% after observation count controls in a fixed test of 6,270 C. elegans calcium responses sci:sha256:04992ad78d9cbd9e6f8d2b7915a6159b53153187fe0109abc980f515385005e5
We reuse its canonical names and full graph facts, then check the Cook graph against the source workbook.
- Response matrix completion reduced error by 2.41% beyond a wiring model in a fixed test of 6,270 C. elegans calcium means sci:sha256:b7d9f820ded7d121c7c54675e017cc5553c2d6853303b6ab4c870e0f0f92671a
This earlier response completion test lacked animal labels. It motivates the present test across animals. The response measure and model differ.
Cited by
8 publicationsPublic catalog entries on Sciscape.
- usesGreen and green/red rules gave opposite response signs in 18.2% of 48,300 measurements from 103 C. elegans recordings
We reuse its exact event table, Cook graph, 11 wiring features and MIT extraction code, and rebuild them from source.
- usesElectrical counts had a positive adjusted term for calcium response correlation in 23 C. elegans records, but repeat coverage changed the estimate
We reuse its exact animal-labelled event table and Cook graph, verify its signed source package, and check all response values against the original NWB files.
- Five image position terms removed a 0.37% wiring gain in a fixed test of 1,225 C. elegans stimuli
This earlier test excluded animals. Our test also excludes stimulated groups and evaluates differences among simultaneous targets.
- usesDirect wiring weights had uncertain terms in a linear test of repeated calcium responses from 46 C. elegans records
This study supplied the original selected event benchmark and response extraction rules through the immediate parent.
- Wiring from the previous light target gave no clear added prediction gain in a fixed ridge test across 36 C. elegans worms
This study supplies earlier animal-separated response prediction methods. The present response definition and selected panel differ.
- usesWith 8 s windows, C. elegans pairs selected for fluorescence decreases had positive mean changes relative to the earlier trend in 11 animal splits
Its event and neuron matching work is inherited through the matched-window package.
- usesMirror response data cut mean test error by 0.43% in 103 C. elegans worms kept out of fitting
We use its verified animal-labelled response benchmark, full Cook graph, features, ridge estimator and source reconstruction code. We add mirror and reverse pooling tests across animals.
- usesMatched windows in 23 C. elegans worms showed no clear increase after light in the electrical count term for response correlation
This study made the event keys and graph arrangement inherited through the repeat-response publication.