Rethinking World Models for Biosignals
Abstract
World models promise to learn how systems evolve and respond to actions, enabling simulation, planning and decision-making. Bringing this promise to biosignals requires more than accurate prediction of physiological recordings. We examine existing biosignal world models through two requirements for intervention use: a genuine action and evidence of its causal effect. We identify three fundamental bottlenecks rooted in physiology: Temporal Latency, the variable delay between an action and its observable effect; System Openness, the influence of drivers beyond the modelled intervention; and Endogenous Dynamics, the system's ongoing evolution without that intervention. Drawing on published evidence, we outline corresponding research directions and propose an Intervention-Effect Validation Checklist to guide evaluation. The aim is to help the biosignal community realise the core promise of world models by learning the latent mechanisms that govern physiological evolution and responses to interventions.
Related
- The political turn of the extended mind: rethinking autonomy beyond individualism in a technological world
- Rethinking Platelet Identity from an Extracellular Vesicle-Centric View: A Proof-of-Concept Study
- Health prevention research in France: Increasing investment is a priority, rethinking targets is equally crucial
- Editorial: Rethinking addiction: technological advances, mechanistic insights, and the re-evaluation of substance use disorders
- Rethinking Syndecan-1 as a Marker of Trauma-Induced Endothelial Injury
BibTeX
@misc{bao2026rethinking,
title = {{Rethinking World Models for Biosignals}},
author = {Xianglong Bao and Bosong Huang and Xiaoda Wang and Keqi Han and Carl Yang and Ming Jin and Shirui Pan},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.17510},
eprint = {2610.17510},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.17510},
}
RIS
TY - UNPB
TI - Rethinking World Models for Biosignals
AU - Xianglong Bao
AU - Bosong Huang
AU - Xiaoda Wang
AU - Keqi Han
AU - Carl Yang
AU - Ming Jin
AU - Shirui Pan
PY - 2026
DA - 2026/10/13
PB - arXiv
DO - 10.48550/arxiv.2610.17510
UR - https://arxiv.org/abs/2610.17510
AB - World models promise to learn how systems evolve and respond to actions, enabling simulation, planning and decision-making. Bringing this promise to biosignals requires more than accurate prediction of physiological recordings. We examine existing biosignal world models through two requirements for intervention use: a genuine action and evidence of its causal effect. We identify three fundamental bottlenecks rooted in physiology: Temporal Latency, the variable delay between an action and its observable effect; System Openness, the influence of drivers beyond the modelled intervention; and Endogenous Dynamics, the system's ongoing evolution without that intervention. Drawing on published evidence, we outline corresponding research directions and propose an Intervention-Effect Validation Checklist to guide evaluation. The aim is to help the biosignal community realise the core promise of world models by learning the latent mechanisms that govern physiological evolution and responses to interventions.
ER -