NeuronDiscover: Agent-in-Twin for Mechanistic Discovery in Neuronal Microenvironments with World Action Models
作者: Haowei Xu, Wanyi Fu, Hongbin Han, Zhaoheng Xie
分类: cs.LG, q-bio.NC
发布日期: 2026-09-28
备注: 54 pages
💡 一句话要点
提出NeuronDiscover以解决神经微环境中的机制发现问题
🎯 匹配领域: 支柱二:RL算法与架构 (RL & Architecture)
关键词: 神经科学 机制发现 Agent-in-Twin 世界行动模型 实验设计 干预提案 计算误差 生物医学工程
📋 核心要点
- 现有方法在神经微环境中难以有效区分真实机制变化与计算误差,导致预测准确性不足。
- NeuronDiscover通过Agent-in-Twin框架和机制-干预-观察-结果(MIOY)图,设计实验以分离机制与误差,提升干预效果。
- 在模拟实验中,NeuronDiscover在每个分配的世界中解决了4.0个关系,显著优于基线方法的3.4个关系,且误差验证降低了程序失败率。
📝 摘要(中文)
在神经微环境中进行机制发现需要通过干预和测量来区分溶质运输和神经反应的竞争解释。预测准确性无法解决这一问题,因为真实的机制变化和计算双胞胎中的误差在稀疏观察中留下相同的特征。本文提出了NeuronDiscover,一个基于Agent-in-Twin框架的世界行动模型(WAM),通过预测、干预提案和观察设计的耦合,设计实验以分离机制和误差。实验结果表明,NeuronDiscover在模拟脑液示踪运输世界中表现优异,解决了更多的关系,并在接受覆盖率和准确性方面优于现有方法。
🔬 方法详解
问题定义:本文旨在解决神经微环境中机制发现的挑战,现有方法无法有效区分真实机制变化与计算误差,导致结果不可靠。
核心思路:NeuronDiscover通过Agent-in-Twin框架,结合机制-干预-观察-结果(MIOY)图,设计实验以分离机制与误差,从而提高干预的有效性和准确性。
技术框架:该框架包括三个主要模块:世界行动模型(WAM),用于预测和干预提案;MIOY图,用于记录和修正实验结果;以及独立评估机制,用于验证实验的有效性。
关键创新:最重要的创新在于引入了机制-干扰信念的联合获取,能够有效地分离真实机制与计算误差,显著提升了实验的可靠性和准确性。
关键设计:在实验设计中,采用了经过验证的参数设置和损失函数,确保了模型的稳定性和准确性,同时通过调整接受范围来降低程序失败率。实验中,使用了独立的细网格参考求解器进行结果验证。
🖼️ 关键图片
📊 实验亮点
实验结果显示,NeuronDiscover在每个分配的世界中解决了4.0个关系,相较于最强基线的3.4个关系有显著提升。同时,误差验证将接受程序的失败率从15%降低至9%,在接受覆盖率达到60%的情况下,表现出更高的准确性和可靠性。
🎯 应用场景
该研究的潜在应用领域包括神经科学、药物开发和生物医学工程等。通过提高对神经微环境中机制的理解,NeuronDiscover可为新药物的靶点发现和神经疾病的治疗提供重要支持,具有广泛的实际价值和深远的未来影响。
📄 摘要(原文)
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.