Mapping Sleep Architecture: How a Markov Model Reveals Sleep State Transitions
Together with a good diet and regular exercise, sleep plays a vital role in our overall health, but it’s more complex than closing your eyes and drifting off.
Our brains go through different stages of sleep, controlled by brain neurons and influenced by internal factors like circadian rhythm and environmental factors like light and noise. As we sleep, the brain switches between different sleep states, such as light sleep, deep sleep, and rapid eye movement (REM) sleep. The way these states switch back and forth can give us important clues about sleep quality. When people experience sleep disorders like insomnia, they might have trouble switching smoothly between these states. To effectively diagnose and treat sleep issues, we need to better understand sleep patterns.
A common way to study sleep is by using something called a hypnogram, which is a visual representation of sleep stages throughout the night. However, the traditional ways of analyzing these hypnograms often focus too much on how long someone spends in each sleep stage (which we often see next to the hypnograms produced by commercial wearables), rather than the patterns of switching between them. New research suggests we’re not getting the full picture of how people move from one sleep state to another.
A new solution for understanding sleep
To solve this problem, researchers at VCU proposed using a more advanced model called a seven-state continuous-time Markov model. Led by Jonathon Jacobs, Ph.D. student in biostatistics with Shanshan Chen, Ph.D., associate professor of biostatistics at VCU’s School of Public Health, Caitlin E. Martin, M.D., associate professor of obstetrics and gynecology, and Bernard Fuemmeler, Ph.D., M.P.H., associate director of population sciences at VCU’s School of Medicine, the study explored how demographic factors, like age, sex, race, and job schedule, impact sleep and sleep transitions, in the dataset from the Multi-ethnic Study of Atherosclerosis (MESA) Sleep Study, which followed 2,056 adults to examine their sleep patterns.
This model helps us see and quantify how transitions between sleep states occur over the course of sleep. One key improvement here is that it breaks down wakefulness into three types:
- Wake before falling asleep
- Wake after falling asleep but before waking up for good (also called WASO – Wake After Sleep Onset)
- Final awakening after sleeping
By focusing on these three types of wakefulness, the model can provide greater insights into three types of insomnia: trouble falling asleep, trouble staying asleep, and waking up too early.
This seven-state sleep architecture can be represented by the figure below:
Insights into how age, gender, and race affect sleep
The researchers found that age, sex, and race all play a role in how people move between sleep stages. For example:
- Aging: Older people tended to wake up more during the night and had trouble reaching deeper sleep stages like deep sleep and REM.
- Sex: Men fell asleep faster than women but had more trouble moving from light sleep to deep sleep and staying in deep sleep.
- Race: Black participants tended to fall asleep faster than White participants, while Hispanic participants took longer to fall asleep than White participants. However, all minority groups had more difficulty moving into deep sleep compared to White participants.
- Job Schedule: Participants with a day job had better sleep patterns than those who are retired or have shift jobs. They tended to fall asleep faster, and transition faster to REM sleep.
These findings show that sleep quality can vary widely based on a person’s demographics, and that these differences are linked to specific patterns of sleep stage transitions.
Why this matters
This work provides a new way of looking at sleep hypnograms and a better understanding of how sleep stage transition patterns vary among different demographic groups. By looking more closely at how people transition between sleep stages, we can identify irregular sleep patterns that occur in patients with health issues like cardiovascular disease. For future research, this model could be a valuable tool in studying the connection between sleep quality and specific health outcomes.
The Future of Sleep Studies with Enhanced Data Models
The seven-state continuous-time Markov model offers a new, detailed way to study sleep and insights into the complex links between sleep, health, and quality of life. Researchers are also working to make these tools available for broader use in sleep studies.
Find the full study published in the Journal of Sleep Research and a guest blog overview written by Dr. Chen on the National Sleep Research Resource (NSRR).
Categories public health research