Association Between Sleep Measures and Cognition Performance: Insights from a Digital Sleep Assessment Study

Huitong Ding, Chenglin Lyu, Edward Searls, Spencer Low, Zachary Popp, Zexu Li, Salman Rahman, Akwaugo Igwe, Kristi Ho, Phillip H Hwang, Ileana De Anda‐Duran, Robert J. Thomas, Vijaya B. Kolachalama, Rhoda Au, Honghuang LinAlzheimer 2024 · DOI

TL;DR

This study leverages wearable digital sleep trackers to investigate the link between objective sleep metrics and cognitive performance, as measured by the MoCA. The findings suggest that certain sleep disturbances, particularly reduced oxygen saturation during sleep, are significantly associated with poorer cognitive function, highlighting sleep as a potential modifiable factor for cognitive health.

What Problem Does This Paper Address?

Cognitive decline, especially in the context of aging and neurodegenerative diseases like Alzheimer's, is a major public health concern. While sleep disorders are increasingly recognized as potential contributors to cognitive impairment, objectively measuring sleep in a naturalistic setting has been challenging. Traditional sleep studies (polysomnography) are often conducted in clinical environments, which can be disruptive and not representative of a person's typical sleep. This paper addresses the gap by exploring how data from consumer-grade wearable devices can be used to assess sleep quality and its association with cognitive function in a real-world setting.

Methods

The researchers utilized a digital sleep assessment approach. Participants wore a "SleepImage Ring" for at least three nights per quarter over a two-week period, collecting various sleep metrics. These metrics included:

  • Duration of unstable non-rapid eye movement (NREM) sleep: This refers to periods of sleep that are not deep or restorative.
  • Percentage of time spent with SpO2 below 80%: This indicates periods of significant oxygen desaturation during sleep, which can be a marker for conditions like sleep apnea.
  • Percentage of stable sleep duration: The proportion of time spent in restorative sleep stages.
  • REM sleep duration: Time spent in rapid eye movement sleep, crucial for memory consolidation.
  • Sleep latency: The time it takes to fall asleep.
  • Number of transitions into wake stage: How often a person wakes up during the night.
  • Average and maximum heart rate: Physiological indicators during sleep.

These objective sleep measures were then statistically analyzed using linear regression models to determine their association with the Montreal Cognitive Assessment (MoCA) total score and individual MoCA sub-scores. The models were adjusted for potential confounding factors such as sex, age, and education level.

Datasets

The study utilized data from 75 participants enrolled in the Boston University Alzheimer's Disease Research Center (BU ADRC) Clinical Core. This cohort is part of a longitudinal study of aging and includes individuals with various Alzheimer's Disease-related clinical features. The data collected included:

  • Sleep data: From the SleepImage Ring wearable device.
  • Cognitive data: Montreal Cognitive Assessment (MoCA) scores.
  • Demographic and clinical data: Sex, age, and education level.

No new datasets were introduced; the study leveraged an existing clinical cohort and a commercially available sleep tracking device.

Key Insights and Ideas

  • Sleep as a Modifiable Factor: The study strongly suggests that sleep quality is not just a consequence of cognitive health but potentially a modifiable factor that can influence it. This opens avenues for interventions targeting sleep to improve or maintain cognitive function.
  • Objective Sleep Metrics Matter: Moving beyond subjective reports, objective measures from wearables can reveal significant associations with cognitive performance, providing a more nuanced understanding of sleep's impact.
  • Oxygen Desaturation is a Red Flag: The finding that reduced SpO2 during sleep is negatively associated with cognitive scores is particularly noteworthy. This points towards the potential role of conditions like sleep apnea in cognitive decline.

Key Results

The analysis revealed several significant associations between sleep measures and MoCA scores:

  • Percentage of time spent with SpO2 below 80% was negatively associated with the MoCA total score (P=0.016). This specific measure was also significantly linked to lower scores in:
  • Language – Naming (P<0.001)
    • Delayed recall – No cue (P=0.027)
    • Abstraction (P=0.034)
  • Other sleep measures also showed nominal significance (P<0.05) with various MoCA items. For instance:
  • Percentage of sleep duration that is stable sleep was positively associated with Visuospatial/executive – Cube (P=0.017).
  • Percentage of unstable sleep was negatively associated with Visuospatial/executive – Cube (P=0.035).
  • REM sleep duration was negatively associated with Memory – Registration (two trials) (P=0.039).

Limitations and Future Work

The authors acknowledge several limitations:

  • Sample Size: The study included 75 participants, which is a relatively small sample size.
  • Generalizability: The participants were from the BU ADRC, meaning they may have pre-existing cognitive concerns or specific health profiles, which could limit the generalizability of findings to the broader population.
  • Device Limitations: While consumer wearables offer convenience, they may not capture the same level of detail or accuracy as clinical polysomnography.

Future work suggested includes:

  • Larger and Independent Samples: Replicating these findings in larger, more diverse cohorts is crucial for validation.
  • Intervention Studies: Investigating whether interventions aimed at improving sleep quality (e.g., treating sleep apnea, optimizing sleep hygiene) can lead to measurable improvements in cognitive function.
  • Longitudinal Tracking: Examining how changes in sleep patterns over time correlate with cognitive trajectories.

Summary updated

Abstract

Abstract Background Sleep disorders as a contributing factor to cognitive impairment have spurred growing interest. The advent of digital technology facilitates the collection of comprehensive sleep measures in a home setting. The objective of this study is to examine the association between digital sleep measures and the Montreal Cognitive Assessment (MoCA). Method This study included participants from the Boston University Alzheimer’s Disease Research Center (BU ADRC) Clinical Core, a longitudinal study of aging which includes the Uniform Dataset (UDS) and other Alzheimer’s Disease related clinical features. Participants were asked to wear a SleepImage Ring when they went to bed at night at least three times in a two week span at quarterly intervals. A variety of sleep measures, such as duration of unstable non‐rapid eye movement (NREM) sleep and percentage of time spent with SpO2 below 80%, were collected from the device and analyzed. Linear regression models were used to assess the associations between these sleep measures and the MoCA total score as well as individual MoCA scores. All models were adjusted for sex, age, and education to account for potential confounding factors. Result Our study included 75 participants from the BU ADRC (mean age: 74.9± 7.9 years; 64.0% women). On average, the SleepImage Ring was worn for 17 nights over the duration of the study (interquartile range: 6‐23 nights). As shown in Table 1, 11 sleep measures were associated with at least one MoCA item with nominal significance (P&lt;0.05). Interestingly, the percentage of time spent with SpO2 below 80% was negatively associated with the MoCA total score (P=0.016) as well as three individual MoCA scores, including the Language – Naming (P&lt;0.001), Delayed recall – No cue (P=0.027) and Abstraction (P=0.034). Conclusion Our analysis revealed multiple suggestive associations between digital sleep measures and MoCA test scores, highlighting the potential of sleep as a modifiable lifestyle factor to assess and improve cognitive health. Further studies with larger and independent samples are required to further validate these findings.
BibTeX
@inproceedings{ding2024association,
  author = {Huitong Ding and Chenglin Lyu and Edward Searls and Spencer Low and Zachary Popp and Zexu Li and Salman Rahman and Akwaugo Igwe and Kristi Ho and Phillip H Hwang and Ileana De Anda‐Duran and Robert J. Thomas and Vijaya B. Kolachalama and Rhoda Au and Honghuang Lin},
  title = {{Association Between Sleep Measures and Cognition Performance: Insights from a Digital Sleep Assessment Study}},
  booktitle = {Alzheimer},
  year = {2024},
  doi = {10.1002/alz.088924},
}
Figure