NeuronDiscover: Agent-in-Twin for Mechanistic Discovery in Neuronal Microenvironments with World Action Models

2026-09-28

AI Summary

This study addresses the challenge of distinguishing twin errors from genuine mechanistic changes under sparse observations in neuronal microenvironment mechanism discovery. To this end, it proposes the Agent-in-Twin framework, which couples prediction, intervention, and observation design via a world action model to iteratively refine MIOY graphs for confounder disentanglement. The approach innovatively introduces joint mechanism-difference belief reasoning alongside difference-adjusted acceptance boundaries, enabling automatic compilation and verification from experiments to executable programs, arbitrated by an independent reference solver. Evaluated on both simulated and real-world data, the framework achieves 4.0 resolved relationships per unit budget with a false positive rate of only 5%, significantly outperforming existing baseline methods.

Abstract

Mechanistic discovery in neuronal microenvironments requires interventions and measurements that separate competing explanations of solute transport and neuronal response. Predictive accuracy cannot settle the question: a real mechanistic change and an error in the computational twin leave the same signature in sparse observations. We formalize this twin confounding and reason over a joint mechanism--discrepancy belief, designing experiments that separate the two. NeuronDiscover is an Agent-in-Twin framework whose shared, mechanism-grounded World Action Model (WAM) couples prediction, intervention proposals, and observation design; independently adjudicated outcomes revise a scoped Mechanism--Intervention--Observation--Outcome (MIOY) graph, whose supported relations compile into executable programs carrying discrepancy-adjusted acceptance bounds. We evaluate on simulated brain-fluid tracer-transport worlds adjudicated by an independently frozen finer-mesh reference solver, and on donor-disjoint public current-clamp recordings of cortical neurons. Counting only relations that reach a certified terminal status, and scoring abstentions as unresolved for every method, at a matched budget of 16 experiments over 32 source units NeuronDiscover resolves 4.0 relations per assigned world against 3.4 for the strongest baseline and 3.2 without graph revision, at 5% false support and 82% scope accuracy. Joint mechanism--discrepancy acquisition resolves 3.8 relations versus 2.9 for plug-in expected information gain; discrepancy-adjusted verification lowers accepted-program failure from 15% to 9% at 60% acceptance coverage; and transfer to the recordings yields 1.94 versus 1.53 relations per assigned world. Correctness is adjudicated within declared model worlds and archival recordings.

Problem

• mechanistic discovery
• neuronal microenvironments
• twin confounding
• solute transport
• computational twin

Innovation

• Agent-in-Twin
• World Action Model
• Twin Confounding
• MIOY Graph
• Mechanistic Discovery