Brown CS News

Ji Won Chung, a third-year PhD student advised by Jeff Huang, Brown CS faculty member and researcher in human-computer interaction, has been collaborating with the developers of Sleep as Android, a popular sleep tracking app that supports vibration on alarms, anti-snoring measures, and lucid dreaming cues. Ji Won’s research focused on writing code to implement a scientifically-evaluated sleep regularity index (SRI), which is now being incorporated into the app itself, and is expected to impact the sleep patterns of millions of people worldwide.

The SRI, discussed in this paper, was developed to quantify changes in sleep patterns between consecutive days on a 24-hour timescale and calculates the probability that an individual is in the same sleep/wake state at the same time on consecutive days averaged over the study period. The researchers used the index to classify 61 college students into regularity quintiles based on their index over two weeks, and the SRI was found to be independent of total sleep duration and positively correlated with academic performance.

The Sleep as Android app implemented SRI to provide users information about the consistency of their daily sleep patterns, giving users a clearer understanding of how they can modify their daily routines. The app also measures total sleep duration, complementing the SRI to give users a holistic view of their sleep patterns over time; tracking both aspects provides actionable insights to improve overall sleep health. 

“We’ve really enjoyed working on incorporating a measure from academic research into one of the most popular sleep apps out there,” Jeff says. “What started as a simple idea became challenging when we thought through all the use cases and accounted for uncommon scenarios that could happen when there are a million users across the world.”

“It’s exciting to see how research in collaboration with industry professionals can lead to real-world impact,” Ji Won says. “I learned a lot of practical lessons on how to design a system that minimizes misinterpretations of the model for the end-user.”

“I’m eager to see how users understand their sleep consistency as part of using the app, which is an important aspect of sleep that goes beyond the usual approach of counting hours of sleep,” Jeff adds. “The Sleep as Android team have been terrific partners over the last five years, and we’ve continued our collaboration from studying sleep changes during COVID-19 and hope to expand our computationally-guided sleep research projects in the future.”

Ji Won’s research expands upon prior research conducted by Jeff’s team:

• In a comprehensive analysis encompassing data from over 64,000 users of the Sleep as Android app between 2019 and 2020, the researchers delved into self-reported sleep patterns to unravel the impact of the COVID-19 pandemic on global sleep behaviors. Their investigation, conducted both on a global scale and for 20 representative countries, revealed a noteworthy shift in sleep times towards later hours, particularly in the early months of 2020, with this effect tapering off as the year progressed. Substantial delays were observed in the midpoint of sleep, bedtimes, and wake times. Notably, larger shifts were evident on weekdays compared to weekends, hinting at a reduction in social jetlag. Their study suggests that the surge in remote work and the implementation of lockdown policies likely contributed to the trend of later sleep schedules by eliminating daily commutes and disrupting established routines. Furthermore, the observed changes may indicate a more harmonious alignment with individual circadian preferences amid the disruptions caused by the pandemic. 

• Tailored to individual sleep needs, SleepCoacher serves as a personalized sleep coach, recognizing that people vary in their sleep requirements, chronotypes, and sensitivities to environmental factors like noise. To enhance a user’s sleep quality, the app provides a curated list of sleep export-endorsed recommendations for them to integrate into their routine. The app transmits sleep data to a system that analyzes sleep history, discerning patterns and connections between behavior and sleep quality, observing the outcomes of behavioral adjustments and informing the user about what specifically works best for them.

“Given that this implementation could impact people’s health, we had to carefully think about the biases we, as developers, may introduce, such as the way we handle missing data and visually display aggregate sleep data to the end-user,” Ji Won says. “We hope the implementation encourages users to think more critically about their day-to-day sleep behaviors and take actionable steps to have more consistent sleep routines, all of which have been deemed important by the sleep research community.”

Brown CS regularly publishes news articles about our pioneering and innovative students. We have no financial involvement in any of the companies mentioned above and have not been compensated in any way for this story.

For more information, click the following link to contact Brown CS Communications Manager Jesse C. Polhemus.