English static mirror for SEO/GEO · AI-assisted translation · Read Chinese original

When Does the Real Wiring Win? Testing Connectome Structure-Function Claims for Tuning Confounds

Forum topic · 小凯 · 2026-11-12

Summary

Connectome-constrained neural models are frequently used to argue that measured wiring outperforms shuffled or rewired connectivity, implying that structure shapes function. This arXiv paper (2611.14508) by Yeyuan Liu shows such comparisons can confound wiring with model tuning: a connectome fixes who connects to whom, but not the excitability, gains, or plasticity rates that modellers often tune until the real-wiring model works. The author separates three estimands: physiology tuned on real wiring and transferred to nulls (P1), per-wiring tuning (P2), and P2 constrained by independent data (P3), mapping the biases to known ML phenomena (under-tuned baselines, inherited weights, winner's curse). Using synthetic ground truths, placebo nulls, and three Drosophila circuit models, some conclusions hold across all protocols—segregated reward/punishment dopamine aids learning, and head-direction topology is needed for persistent activity—while others do not. Notably, real Kenyon cell output connectivity in the mushroom body wins under P1/P2 but loses under P3, with the verdict controlled by the fraction of active alpha'-beta' Kenyon cells. Credible structural claims require matched optimization and data-constrained physiology.

Paper Overview

Field: Machine Learning Author: Yeyuan Liu Date: 2026-11-10 arXiv: 2611.14508

Summary

Connectome-constrained models are often used to argue that a measured wiring diagram outperforms shuffled or rewired wiring, and hence that structure shapes function. Such comparisons can confound wiring with how the models are tuned: a connectome fixes who connects to whom, not the excitability, gains or plasticity rates the model also needs, and modellers often tune these until the real-wiring model works.

The paper separates three estimands:

  • P1: physiology tuned on the real wiring and transferred to nulls
  • P2: each wiring tuned on its own
  • P3: P2 restricted to physiology consistent with independent data
  • The biases involved are known in machine learning (under-tuned baselines, inherited weights, winner's curse). The author measures their effect on biological conclusions using synthetic ground truths, placebo null models, and three Drosophila loop models.

    Findings

  • Robust conclusions (hold under all protocols): segregated reward/punishment dopamine helps learning; head-direction topology is necessary to maintain persistent activity wave peaks.
  • Protocol-dependent conclusions: real Kenyon cell output connectivity in the mushroom body beats cross-subtype-identity shuffles under P1/P2 (margin 0.09), but loses by 0.09 under P3 (95% CI excludes zero; held-out runs confirm). The verdict is determined by the fraction of alpha'-beta' cells among active Kenyon cells—the fraction selected by P1 is too low. At the proportion implied by imaging data (~44%), the real-wiring model fails to learn, exposing model incompleteness rather than a defect in the fly connectome.
  • In flyvis, P1 penalizes the mirrored connectome more than spatial shuffling (for both 50/50 member types); rankings persist after 10x retraining, and depend on checkpoint choice at 50x.

Conclusion

Credible structural claims require matched optimization, physiology constrained by independent data, and models that still work under that constraint.

Tags

#machine-learning#connectome#neuroscience#drosophila#structure-function#model-tuning#arxiv#connectome-constrained-models

This page is an English static mirror generated for search and AI citation. It may be a full translation or structured summary of the Chinese original. Canonical interactive discussion lives on the Chinese page: https://zhichai.net/topic/178637066