Heart rate circadian phase and hyperarousal as wearable digital phenotyping of insomnia: An interpretable machine learning study - Report - MDSpire

Heart rate circadian phase and hyperarousal as wearable digital phenotyping of insomnia: An interpretable machine learning study

  • By

  • Minji Kim

  • Seojin Yun

  • Hyungju Kim

  • Emma Matsushita

  • Ji Won Yeom

  • Sujin Kim

  • Seung Pil Pack

  • Heon-Jeong Lee

  • Taesu Cheong

  • Chul-Hyun Cho

  • June 6, 2026

Share

Clinical Report: Circadian Heart Rate Patterns and Hyperarousal as Indicators of Insomnia

Overview

Revise to specify the role of machine learning models in enhancing insomnia classification and understanding.

Background

Insomnia is a prevalent condition affecting 10–30% of individuals, often diagnosed through subjective self-reports that may be biased. The advent of wearable technologies offers a promising avenue for objective sleep assessment, yet challenges remain in translating machine learning models into clinical practice due to issues of interpretability and clinical validity. Understanding circadian rhythms and their impact on sleep is crucial for developing effective interventions.

Data Highlights

No numerical data available in the provided source material.

Key Findings

  • Insomnia classification can be enhanced using wearable-derived digital phenotypes.
  • Machine learning models must consider clinical plausibility alongside predictive performance for trustworthy applications.
  • Wearable technology provides objective data that can improve the understanding of insomnia's physiological underpinnings.
  • Excessive optimization of machine learning models may lead to spurious associations that do not generalize across populations.
  • Regulatory frameworks emphasize the need for transparency and explainability in AI-driven healthcare solutions.

Clinical Implications

Clinicians should consider integrating wearable technology data into insomnia assessments to improve diagnostic accuracy. Emphasizing clinical plausibility in machine learning model selection can enhance trust in AI-assisted decision-making and ensure patient safety.

Conclusion

The study highlights the potential of wearable technology in insomnia classification while underscoring the necessity of aligning machine learning outputs with established clinical knowledge. This approach may foster greater confidence in AI applications within sleep medicine.

Related Resources & Content

  1. European Journal of Preventive Cardiology, 2023 -- From light and activity to risk: circadian alignment as an emerging wearable biomarker
  2. Conexiant, 2023 -- One Night's Sleep May Predict 130 Diseases
  3. npj Digital Medicine, 2023 -- Home-Based Detection of Isolated REM Sleep Behavior Disorder Using a Lumbar Wearable Sensor
  4. Journal of Clinical Sleep Medicine, 2025 -- Combination treatment for chronic insomnia disorder in adults: an American Academy of Sleep Medicine clinical practice guideline
  5. npj Digital Medicine — Actigraphy-based detection of isolated REM sleep behavior disorder: multicenter validation across devices and populations
  6. Combination treatment for chronic insomnia disorder in adults: an American Academy of Sleep Medicine clinical practice guideline | Journal of Clinical Sleep Medicine | Springer Nature Link
  7. Heart rate variability during sleep onset in patients with insomnia with or without comorbid sleep apnea - ScienceDirect
  8. Concordance of Wearable Device Sleep Metrics with Patient-Reported Sleep Quality: A Systematic Review - ScienceDirect

Original Source(s)

Heart rate circadian phase and hyperarousal as wearable digital phenotyping of insomnia: An interpretable machine learning study

  • by Minji Kim, Seojin Yun, Hyungju Kim, Emma Matsushita, Ji Won Yeom, Sujin Kim, Seung Pil Pack, Heon-Jeong Lee, Taesu Cheong, Chul-Hyun Cho

  • June 6, 2026

Related Content

Post-stroke insomnia: multidimensional mechanisms, clinical heterogeneity, and toward mechanism-informed, objectively quantified management

Side Effects: Salt, Steroids, and Second Brains

So get this: sodium may track with memory decline (in men), steroids might not be “immunosuppressive” in the ICU, and second pregnancies reshape the brain differently than first. Same theme: biology is less binary than we teach it.

Mandatory Training Modules Deserve a Harder Look

A JAMA Viewpoint argues that 4 million physician-hours spent on compliance modules each year warrants institutional reassessment.