A nested ridge test selected no wiring terms for response direction differences in 96 C. elegans worms

We tested whether anatomical wiring improves prediction of calcium response differences between opposite stimulation directions. A fixed DANDI release gave 5,069 paired observations from 96 C. elegans worms. Both directions were measured in the same animal. Five outer folds excluded complete animals from fitting and tuning. A baseline used cell names and measured assay values. A separate ridge penalty controlled added unsigned counts, two-step paths, walk sums and source-response products. The main difference test selected no wiring terms in all five folds. Requiring finite wiring penalties raised test error by 0.764%; its descriptive animal-resample range crossed zero. Excluding test animal and class-pair groups also gave no clear gain. Known added direction signals were recovered. A later bounded-distance check reduced common-level extrapolation and left the direction result unchanged. Most pairs had one event per direction, and the median directional time gap was 431 seconds. These results limit the tested conditional prediction rules. They do not show equal biological responses or absent anatomical direction information. The package supplies the animal-labelled paired benchmark, complete graph facts, source provenance, code, all results, figures and repeat steps.

connectome · Caenorhabditis elegans · signal propagation · direction differences · animal-separated prediction · paired benchmark · ridge regression