Response reference choice reversed the pooled recent-change slope in 87 C. elegans recordings

We tested whether a negative relation between recent and later fluorescence changes identifies negative temporal dependence after neural stimulation. We used 49,367 event-target observations from 87 C. elegans recordings. We removed each animal-target mean. Recent change was log(B/A), where A and B were two windows before light. Later change was log(C/B), where C was after light. The pooled later-change slope was -0.126. Using A as the response reference gave +0.874. Uniform permutation of C within each animal-target group had an exact expected slope of -0.353. The matched later-window term was +0.227, with a 95% animal resample range of +0.158 to +0.299. Thus a negative change slope coexisted with positive covariance with the later level. The reference identity is established mathematics; the new contribution is its measured size and a reusable raw-window benchmark in this atlas. In an added unpenalized wiring fit, supplying B as a predictor made wiring terms invariant to subtraction of B. Direct chemical-only and electrical terms were positive in green and green/red fits. Their recent-change interaction ranges crossed zero. These are descriptive fits, not prediction in new animals. Permutation changes source and time assignments. It is not a sham stimulus or a neural independence test. The results do not identify inhibition or separate neural dynamics from fluorescence effects. The package preserves raw-source hashes, window means, code, figures, complete results, and repeat steps.

connectome · calcium imaging · baseline reference · mathematical coupling · Caenorhabditis elegans · optogenetics · state dependence