Beyond AHI: An Interpretable Causal-Discovery-Guided Framework for Sleep Recovery in Connected Health

2026-06-16

AI Summary

This study addresses the limitations of conventional sleep assessment metrics, such as the Apnea–Hypopnea Index (AHI), which inadequately capture patients’ subjective sense of restorative sleep and lack explanatory power regarding multidimensional physiological mechanisms. To overcome this, the authors propose an interpretable, hierarchical Sleep Restoration Score (SRS) derived from multimodal polysomnography (PSG) data. The SRS integrates five key physiological domains—respiratory events, hypoxia, sleep fragmentation, sleep architecture, and autonomic regulation—into a causal model tightly aligned with self-reported restoration. Methodologically, a two-stage variable selection framework is introduced, combining physiological prior constraints with large language model–assisted auditing, followed by directed acyclic graph learning under causal discovery constraints to ensure mechanistic plausibility and interpretability. Validation in the MESA and MrOS cohorts demonstrates that SRS correlates with subjective restoration up to 2.5 times more strongly than AHI, highlighting its clinical utility and potential for wearable health applications.

Abstract

Objective sleep assessment relies on polysomnography (PSG), yet clinical impact is often better reflected in patient-reported outcomes (PROs) such as sleepiness and fatigue. Existing summary indices, including the Apnea-Hypopnea Index (AHI), provide limited insight into the multidomain physiology underlying functional recovery. We propose an interpretable, causal-discovery--guided framework for deriving a hierarchical Sleep Recovery Score (SRS) from multimodal PSG. Using two large population cohorts (MESA: n=1540; MrOS: n=825), we apply directed acyclic graph (DAG) learning to identify candidate physiological drivers spanning respiratory burden, hypoxic burden, sleep fragmentation, sleep architecture, and autonomic regulation. Although derived from clinical PSG, these domains map naturally to sensing streams increasingly available in connected health technologies, including wearable ECG, oximetry, and sleep-stage estimation devices. To preserve mechanistic plausibility, we introduce a two-stage screening process that combines physiology-based constraints with constrained LLM-assisted auditing to identify and remove structural confounders and construct-overlapping variables. Across cohorts, these five domains emerge as recurrent physiological domains associated with recovery, and the resulting SRS shows up to 2.5$\times$ stronger alignment with perceived recovery than AHI. By linking multimodal sleep physiology to patient-centered outcomes through an interpretable, bias-aware, and domain structured framework, this work provides a practical foundation for recovery modeling across both clinical sleep studies and emerging smart and connected health settings.

Problem

• sleep recovery
• patient-reported outcomes
• Apnea-Hypopnea Index
• multimodal physiology
• functional recovery

Innovation

• causal discovery
• sleep recovery score
• multimodal PSG
• interpretable framework
• connected health