Wiring from the previous light target gave no clear added prediction gain in a fixed ridge test across 36 C. elegans worms

We tested whether wiring from the previous light target helps predict the calcium response after the next pulse in C. elegans. We added exact pulse history to a signed matched-window dataset and checked the original processed files in a fixed DANDI release. The selected panel had 124 source switches in 36 worms and 2,786 downstream measurements. A fixed ridge test kept each test worm out of fitting and tuning. The model supplied cell labels, cell state, source activity, position and current-source wiring. Adding 18 terms for previous-source wiring raised mean test error by 0.40%. A central 95% range from 300 animal refits allowed a 2.13% error increase to a 0.77% decrease. The before-period test and five planned checks also gave no clear added gain. Most pulse gaps were near 30 seconds. Label and trace rules retained only 36 of 90 parent worms. The result limits this static prediction rule; it does not show that previous pulses have no biological effect. The package preserves the new event-linked table, code, provenance, model states, predictions, figures and repeat instructions.

connectome · Caenorhabditis elegans · signal propagation · stimulation history · calcium imaging · animal-separated prediction · reproducible dataset