Peptide channel terms raised ridge test error by 0.27% for an early minus late calcium contrast in 104 C. elegans worms

We tested whether possible peptide channels help predict an early minus late calcium response contrast in other C. elegans animals. We joined 92 peptide-receptor layers to 19,811 mean responses for directed neuron pairs from 104 worms. Five outer folds excluded whole animals. A nested ridge model controlled for direct chemical and electrical counts, neuron identity, imaging measures, and each possible channel count. Adding channel terms raised mean animal squared error from 0.01500821 to 0.01504845, an increase of 0.27%. A central 95% range from 199 complete refits with changed animal weights gave an error reduction of -1.17% to +0.32%. A separate channel penalty reduced the error increase to 0.025%. For the primary contrast, sample and split checks also had negative point gains. The known assignment of receptor profiles gave no superior gain among 199 matched-profile controls, with descriptive rank 0.970. A later empirical pair-template diagnostic gave a small 0.106% error reduction. This test gives no clear added timing prediction from fixed channel support. It does not refute peptide effects. The response is a coarse fluorescence contrast, many directions have one trial, and wiring and expression came from other animals. The package preserves code, data facts, source versions and hashes, predictions, all simulation draws, figures, and repeat instructions.

connectome · neuropeptide · calcium imaging · Caenorhabditis elegans · response timing · cross validation · ridge regression