NeuronSifter: An Agent-in-Twin for Prioritizing Central Nervous System Interventions with World Action Models
Abstract
Effective central nervous system (CNS) intervention requires more than predicting biological responses: a scientific system must determine which intervention, observation, or additional evidence can improve a decision under uncertainty. However, in partially observed neuronal microenvironments, identical predicted outcomes may correspond to different exposure states, biological mechanisms, and intervention risks. We propose NeuronSifter, a decision-centric Agent-in-Twin framework, where an AI agent operates with an evolving biological digital twin rather than a fixed predictive model. NeuronSifter compiles interventions into state-conditional target-occupancy representations, propagates them through a mechanism-grounded World Action Model (WAM), and selects informative measurements based on their expected reduction in decision uncertainty. A shared probabilistic belief state couples intervention simulation, observation selection, and outcome assimilation, enabling the digital twin to continuously refine its decision-relevant representations through externally verified evidence. We evaluate NeuronSifter in Alzheimer's disease intervention scenarios using mechanistic neuronal microenvironment simulations and retrospective clinical endpoint references. The evaluation framework examines adaptive evidence acquisition, uncertainty-aware intervention planning, and robustness under model discrepancy through controlled environments and independent comparisons. NeuronSifter demonstrates how an AI system can actively identify missing evidence, refine biological hypotheses, and optimize intervention decisions under limited experimental budgets. Code is available at https://anonymous.4open.science/r/NeuronSifter.
Related
- Association between H19 polymorphisms and susceptibility to pediatric central nervous system tumors in Chinese children
- LightSuite: scalable atlas registration and cell detection across the central nervous system
- CSF pharmacokinetics-pharmacodynamics of linezolid in critically brain injured patients, with or without central nervous system healthcare-associated infection. The PK-Pop-LCR Study: A Multicentre Pharmacokinetics and Pharmacodynamics Population Study
- Central Nervous System Reservoirs and HIV Cure: Therapeutic Opportunities and Challenges.
- Wnt/beta-catenin signaling regulates Vascular Endothelial Growth Factor (VEGF) receptors in central nervous system endothelial cells
BibTeX
@misc{xu2026neuronsifter,
title = {{NeuronSifter: An Agent-in-Twin for Prioritizing Central Nervous System Interventions with World Action Models}},
author = {Haowei Xu and Wanyi Fu and Hongbin Han and Zhaoheng Xie},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2609.35445},
eprint = {2609.35445},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.35445},
}
RIS
TY - UNPB
TI - NeuronSifter: An Agent-in-Twin for Prioritizing Central Nervous System Interventions with World Action Models
AU - Haowei Xu
AU - Wanyi Fu
AU - Hongbin Han
AU - Zhaoheng Xie
PY - 2026
DA - 2026/10/14
PB - arXiv
DO - 10.48550/arxiv.2609.35445
UR - https://arxiv.org/abs/2609.35445
AB - Effective central nervous system (CNS) intervention requires more than predicting biological responses: a scientific system must determine which intervention, observation, or additional evidence can improve a decision under uncertainty. However, in partially observed neuronal microenvironments, identical predicted outcomes may correspond to different exposure states, biological mechanisms, and intervention risks. We propose NeuronSifter, a decision-centric Agent-in-Twin framework, where an AI agent operates with an evolving biological digital twin rather than a fixed predictive model. NeuronSifter compiles interventions into state-conditional target-occupancy representations, propagates them through a mechanism-grounded World Action Model (WAM), and selects informative measurements based on their expected reduction in decision uncertainty. A shared probabilistic belief state couples intervention simulation, observation selection, and outcome assimilation, enabling the digital twin to continuously refine its decision-relevant representations through externally verified evidence. We evaluate NeuronSifter in Alzheimer's disease intervention scenarios using mechanistic neuronal microenvironment simulations and retrospective clinical endpoint references. The evaluation framework examines adaptive evidence acquisition, uncertainty-aware intervention planning, and robustness under model discrepancy through controlled environments and independent comparisons. NeuronSifter demonstrates how an AI system can actively identify missing evidence, refine biological hypotheses, and optimize intervention decisions under limited experimental budgets. Code is available at https://anonymous.4open.science/r/NeuronSifter.
ER -