Determine whether a smaller odor-specific associative-learning interaction exists

Determine whether a small odor-specific interaction between odor reversal and plasticity blocking exists in the evolved agents, despite the absence of a statistically significant interaction in the experiments reported.

Background

The study tested whether the randomized networks’ stronger dependence on olfaction was mediated by odor-specific associative learning. No statistically significant odor-reversal-by-plasticity interaction was detected in any condition, although plasticity itself produced a measurable fitness benefit.

Because the analysis used only ten seeds and the confidence intervals were compatible with a small interaction, the authors explicitly leave unresolved whether a weaker odor-specific associative-learning effect exists. Establishing this would require greater statistical power or a more sensitive experimental design.

References

We detected no odor-specific associative learning in any condition: in an odor-reversal assay crossed with plasticity blocking, the interaction never reached significance, although plasticity itself was worth 0.09 to 0.30 fitness units. With 10 seeds a small interaction cannot be excluded.

— Null-model treatment of the sensory-motor boundary changes an evolutionary connectome comparison  (2609.39248 - Park, 30 Sep 2026) in Appendix, Section “Further checks,” subsection “What survives from the first experiment”