analiza aiqbench
This research addresses a critical gap in automated sleep analysis where deep learning models often generate physiologically impossible transitions that skew downstream metrics like total sleep time and REM latency. By integrating physiology-informed priors directly into the inference framework, StageGuard enforces realistic state durations without sacrificing classification accuracy or blocking rare pathological events. The method demonstrates significant improvements in derived statistics across multiple datasets, ensuring that large-scale studies yield more reliable clinical endpoints for conditions such as obstructive sleep apnea.
StageGuard: Physiologically Constrained Sleep Staging
arXiv:2607.23284v1 Announce Type: new Abstract: Automated sleep staging is increasingly used in large-scale studies to derive sleep-architecture endpoints: total sleep time, REM latency, sleep efficiency, and bout-duration statistics. Deep learning models achieve epoch-level accuracy approaching inter-rater agreement, yet often produce hypnograms that violate physiological invariants, such as rare transitions (e.g., direct Wake -> REM) or excessively fragmented sequences. Such violations can bias downstream sleep metrics, regardless of overall accuracy. We propose StageGuard, a plug-and-play, backbone-agnostic structured-inference framework that wraps any neural sleep-staging backbone with physiology-informed priors. StageGuard combines (1) a differentiable soft transition penalty that discourages physiologically rare transitions during training, and (2) a semi-Markov constrained decoder with a duration-augmented state space that jointly enforces transition penalties and minimum bout durations at inference. Unlike hard-prohibition methods, it admits rare transitions when emission evidence is overwhelming, leaving informative pathological events recoverable rather than blocked. StageGuard constrains staging outputs to satisfy known physiological priors rather than modeling sleep generatively. We quantify the validity gap using transition-violation rate (TVR) and fragmentation index (FI) and demonstrate that, across six backbones and four datasets, StageGuard reduces TVR to physiologically plausible levels and lowers FI by 56-62%, while maintaining or slightly improving classification accuracy. Crucially, improved constraint satisfaction translates into 59-79% lower error on derived sleep-architecture statistics not directly optimized by the method, and recovers the direction and effect size of expert-defined subgroup differences (OSA severity, age) more faithfully than the unconstrained baseline.
Abstract:Automated sleep staging is increasingly used in large-scale studies to derive sleep-architecture endpoints: total sleep time, REM latency, sleep efficiency, and bout-duration statistics. Deep learning models achieve epoch-level accuracy approaching inter-rater agreement, yet often produce hypnograms that violate physiological invariants, such as rare transitions (e.g., direct Wake -> REM) or excessively fragmented sequences. Such violations can bias downstream sleep metrics, regardless of overall accuracy. We propose StageGuard, a plug-and-play, backbone-agnostic structured-inference framework that wraps any neural sleep-staging backbone with physiology-informed priors. StageGuard combines (1) a differentiable soft transition penalty that discourages physiologically rare transitions during training, and (2) a semi-Markov constrained decoder with a duration-augmented state space that jointly enforces transition penalties and minimum bout durations at inference. Unlike hard-prohibition methods, it admits rare transitions when emission evidence is overwhelming, leaving informative pathological events recoverable rather than blocked. StageGuard constrains staging outputs to satisfy known physiological priors rather than modeling sleep generatively. We quantify the validity gap using transition-violation rate (TVR) and fragmentation index (FI) and demonstrate that, across six backbones and four datasets, StageGuard reduces TVR to physiologically plausible levels and lowers FI by 56-62%, while maintaining or slightly improving classification accuracy. Crucially, improved constraint satisfaction translates into 59-79% lower error on derived sleep-architecture statistics not directly optimized by the method, and recovers the direction and effect size of expert-defined subgroup differences (OSA severity, age) more faithfully than the unconstrained baseline.
12 pages. Accepted at KDD 2026 (32nd ACM SIGKDD Conference), AI for Sciences track
Machine Learning (cs.LG)
arXiv:2607.23284 [cs.LG]
(or arXiv:2607.23284v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2607.23284 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
https://doi.org/10.1145/3770855.3818916 Focus to learn more DOI(s) linking to related resources
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 350 cuvinte
Subiecte conexe
Beyond AHI: An Interpretable Causal-Discovery-Guided Framework for Sleep Recovery in Connected Health
arXiv · 2 sept. 2026
Subject-Relative Micro-Motion and Sleep Dynamics for Near-Infrared Video Sleep Staging
arXiv · 9 sept. 2026
NanoSleep: A Parameter-Efficient Hybrid Temporal Convolutional Network for Single-Channel Sleep Stage Classification
arXiv · 20 aug. 2026
SemPSG: A Semantic Channel-Aware Foundation Model for Polysomnography Analysis
arXiv · 30 sept. 2026