A continuous approach to explain insomnia and subjective-objective sleep discrepancy.

Area of Science:

• Sleep Medicine
• Computational Neuroscience
• Artificial Intelligence in Healthcare

Background:

• Insomnia diagnosis and treatment are hindered by discrepancies between subjective patient complaints and objective sleep quality measures.
• Subjective-Objective Sleep Discrepancy (SOSD) represents a significant challenge in accurately assessing insomnia.
• Objective sleep measures often fail to capture the full spectrum of sleep disturbances reported by individuals with insomnia.

Purpose of the Study:

• To investigate the underlying physiological differences in insomnia, particularly focusing on subjective-objective sleep discrepancy (SOSD).
• To develop and apply machine learning models for analyzing sleep dynamics and identifying sleep intrusions and instability.
• To establish a principled framework for measuring sleep quality by integrating subjective and objective sleep data.

Main Methods:

• Utilized polysomnographic recordings from a large clinical database to measure sleep intrusions and instability.
• Developed personalized machine learning models to infer hypnodensities, a probabilistic measure of sleep dynamics.
• Applied information theory to quantify sleep intrusions and instability from hypnodensity data.

Main Results:

• Insomnia with SOSD was characterized by sleep intrusions occurring during intra-sleep wakefulness.
• Insomnia without SOSD exhibited wake intrusions that occurred during sleep periods.
• These distinct patterns suggest different etiological pathways for insomnia with and without SOSD.

Conclusions:

• The study provides a novel, continuous, and interpretable framework for measuring sleep quality by mapping identified metrics to standard sleep features.
• This approach offers a more accurate assessment of sleep quality and disorders by integrating subjective complaints with objective physiological data.
• The findings highlight the importance of distinguishing between different types of sleep disturbances in insomnia for improved diagnosis and treatment.