SleepGuru: Personalized Sleep Planning System for Real-life Actionability and Negotiability

Title of the Paper

SleepGuru: Personalized Sleep Planning System for Real-life Actionability and Negotiability

Paper Information

• Research Area: Human-Computer Interaction and Health Informatics
• Keywords: Personalized sleep planning, sleep modeling, computational optimization, real-life constraints, sleep quality, sleep feedback, health informatics
• Conference: The 35th ACM Symposium on User Interface Software and Technology (UIST '22)
• DOI: 10.1145/3526113.3545709

Research Background and Problem Statement

• Identified Problems or Challenges:

  • Existing sleep health guidelines emphasize fixed sleep routines and durations (e.g., maintaining 7-8 hours of sleep daily). However, many people face highly irregular schedules due to work, academic, or family responsibilities, making these standard guidelines difficult to follow.
  • Work obligations and societal expectations for productivity often lead to sleep being neglected, posing long-term threats to sleep health.
  • Standard sleep guidelines typically attribute non-compliance to individual self-management issues, overlooking the diverse needs and constraints of individuals' actual lives.

• Significance of the Research:

  • Prolonged unhealthy sleep not only causes immediate fatigue but also has profound impacts on cognitive function, physical health, career, and quality of life.
  • Providing actionable and personalized sleep recommendations based on individual life constraints can improve real-world adherence to healthy sleep practices.

• Motivation and Related Work:

  • Existing mobile applications and HCI systems for healthy sleep primarily offer generic advice or data tracking but lack adaptation to users' real-life constraints.
  • The challenge lies in integrating health-focused recommendations with users' daily irregularities to create actionable and flexible sleep plans.

Proposed Solution

• Proposed Solution:

  • SleepGuru: A dynamic sleep planning system that supports personalized and actionable sleep recommendations based on users' irregular daily schedules.
  • The system is grounded in sleep physiology theories, leveraging users' online calendars and wearable device data to predict and optimize sleep pressure changes.
  • SleepGuru generates personalized multi-day sleep plans and allows users to select alternative options when unavoidable circumstances arise.

• Innovations:

  1. Actionability and Adjustability: Sleep plans provide actionable guidance tailored to users' real-life constraints and dynamic physical activities, while allowing users to adjust the plans.
  2. Real-time Adaptation: The system automatically optimizes future sleep plans based on updated sleep and activity data.
  3. Explanatory Interface: The user interface visualizes minute-level changes in sleepiness predictions, intuitively explaining the scientific basis for the optimized plans.
  4. Multi-day Optimization: Sleep plans consider long-term sleep pressure changes, avoiding the negative effects of one-time compensatory long sleep durations.

• Implementation Steps and Key Technologies:

  1. Data Integration and Acquisition:
     • Use online calendars to obtain users' work and non-sleep schedules.
     • Collect heart rate, activity steps, and actual sleep times via wearable devices (e.g., Fitbit).
  2. Mathematical Modeling and Optimization:
     • Model sleep pressure based on "Sleep Drive" and "Circadian Rhythm."
     • Define an optimization objective function to balance users' sleep pressure and available time.
  3. User Interface and Interaction:
     • Provide a visual dashboard displaying sleep plans on the calendar and support real-time adjustments to recommended times.

Research Outcomes

• Specific Results:

  1. User Testing Results:
     • In an 8-week real-world test, SleepGuru users' sleep quality scores (LSEQ) improved by 15% compared to non-users.
     • Users experienced improved sleep efficiency, reducing fatigue pressure more effectively per unit of sleep time.
  2. User Adherence:
     • SleepGuru's recommended sleep plans were more actionable than standardized guidelines, achieving an adherence rate (58%) approximately three times higher than standard guidelines.
  3. User Feedback:
     • Compared to traditional sleep guidelines, most users reported that SleepGuru provided recommendations better suited to their lifestyle, increasing the likelihood of long-term use.
     • Users highly rated the explanatory and adjustable interface of the sleep prediction tool.

• Advantages Over Existing Solutions:

  • Existing sleep systems often focus on idealized goals, whereas SleepGuru offers practical, actionable solutions that balance user dynamics with scientific insights.
  • SleepGuru incorporates generalized multi-day effects in predicting users' physiological sleep states, rather than limiting analysis to single-night sleep.

• Experimental or Evaluation Results:

  1. Self-reported data and physiological observations showed that SleepGuru significantly reduced daytime sleepiness (ESS scores improved by 20%).
  2. The system automatically adjusted users' sleep recommendations, re-optimizing future plans when deviations occurred.
  3. SleepGuru's educational interface enhanced users' understanding of the importance of their sleep needs.

• Limitations and Future Directions:

  • Limitations:
    1. Some users' extremely busy schedules limited the flexibility for sleep adjustments.
    2. The sleep pressure model did not account for non-sleep factors (e.g., caffeine intake).
    3. Some users failed to update their online calendars in a timely manner, leading to discrepancies between recommendations and actual needs.
  • Future Directions:
    1. Incorporate more comprehensive constraint factors (e.g., nighttime light exposure, dietary habits).
    2. Develop technologies for automatic detection of users' schedule events.
    3. Support long-term deployment to enrich personalized user data and infer genetic attributes.

Conclusion

SleepGuru provides a practical and user-friendly approach to personalized sleep planning, offering a novel solution for optimizing both health and daily productivity. Testing results demonstrate significant improvements in users' sleep quality, efficiency, and long-term adherence, laying a foundation for future applications and in-depth research in this domain.