Validation of an automated sleep detection algorithm using data from multiple accelerometer brands

Are nocturnal awakenings at age 1 predictive of sleep duration and efficiency at age 6: Results from two birth cohorts

Objective

To investigate the association of nighttime awakenings at 12 months with the duration and efficiency of nighttime sleep at 6 years of age.

Methods

Data from two population-based prospective studies (The Pelotas 2004 and The Pelotas 2015 Birth Cohorts) were used. Information on nighttime awakenings was provided by mothers during the 12-month follow-up interview. Infants who awakened >3 times after sleep onset at 12 months were considered frequent wakeners. Sleep duration and sleep efficiency were obtained by actigraphy at the 6-year follow-up. Children wore the device at the wrist of the non-dominant arm continuously for 3-7 days, including at least one weekend day. Unadjusted and adjusted beta coefficients were obtained by linear regression for each cohort separately.

Results

2500 children from the 2004 and 2793 from the 2015 cohort had full information on nighttime awakenings at 12 months and actigraphy at 6 years and were analyzed. Prevalence of frequent wakeners was 6.3 % and 5.9 % in the 2004 and 2015 cohort, respectively. Mean bedtime and wake-up time at 6 years were, respectively, 23:23 and 08:41 h in the 2004 cohort, and 00:10 and 09:00 h int the 2015 cohort. Nighttime sleep lasted on average 7.54 and 7.24 h respectively in the 2004 and the 2015 cohort, and the sleep efficiency was 81.1 and 82.5 % respectively. In adjusted analyses, no associations were found between awakening at 12 months and sleep duration or sleep efficiency at 6 years of age.

Conclusion

In both cohorts sleep duration and efficiency were below the recommendation for school-age children (respectively 9-11 h and 85 %). There was no relationship between the number of nighttime awakenings at 12 months and sleep duration or efficiency at 6 years.

A Machine Learning Model for Predicting Sleep and Wakefulness Based on Accelerometry, Skin Temperature and Contextual Information

Purpose

Body-worn accelerometers are commonly used to estimate sleep duration in population-based studies. However, since accelerometry-based sleep/wake-scoring relies on detecting body movements, the prediction of sleep duration remains a challenge. The aim was to develop and evaluate the performance of a machine learning (ML) model to predict accelerometry-based sleep duration and to explore if this prediction can be improved by adding skin temperature data, circadian rhythm based on the estimated midpoint of sleep, and cyclic time features to the model.

Patients and Methods

Twenty-nine adults (17 females), mean (SD) age 40.2 (15.0) years (range 17-70) participated in the study. Overnight polysomnography (PSG) was recorded in a sleep laboratory or at home along with body movement by two accelerometers with an embedded skin temperature sensor (AX3, Axivity, UK) positioned at the low back and thigh. The PSG scoring of sleep/wake was used as ground truth for training the ML model.

Results

Based on pure accelerometer data input to the ML model, the specificity and sensitivity for predicting sleep/wake was 0.52 (SD 0.24) and 0.95 (SD 0.03), respectively. Adding skin temperature data and contextual information to the ML model improved the specificity to 0.72 (SD 0.20), while sensitivity remained unchanged at 0.95 (SD 0.05). Correspondingly, sleep overestimation was reduced from 54 min (228 min, limits of agreement range [LoAR]) to 19 min (154 min LoAR).

Conclusion

An ML model can predict sleep/wake periods with excellent sensitivity and moderate specificity based on a dual-accelerometer set-up when adding skin temperature data and contextual information to the model.

Objectively assessed sleep quality parameters in Multiple Sclerosis at home: Association to disease, disease severity and physical activity

BACKGROUND

Multiple Sclerosis (MS) is a chronic inflammatory autoimmune, neurodegenerative disease that affects regular mobility and leads predominantly to physical disability. Poor sleep quality, commonly reported in MS patients, impacts their physical activity (PA). Accelerometers monitor 24-h activity patterns, offering insights into disease progression in daily life.

OBJECTIVE

To test if the sleep quality variables of MS patients, as assessed with wrist-worn accelerometers, differ from those of controls and are associated with PA and disease severity variables.

METHODS

Seven-day raw accelerometer data collected from 40 MS patients and 24 controls was processed using an open-source GGIR package, from which variables of sleep quality (sleep efficiency, wake after sleep onset (WASO), sleep regularity index (SRI), intradaily variability (IV)) and PA (of different intensities: inactivity, light (LPA), moderate (MPA), vigorous (VPA)) were analyzed. The variables were compared between the two study groups and in MS patients, correlation tested associations among the variables of sleep quality, PA, and disease severity (assessed with the Expanded Disability Status Scale, EDSS).

RESULTS

Sleep efficiency was the only variable that differed significantly between MS patients and controls (lower in MS, p = 0.01). Both SRI (positively) and IV (negatively) correlated with the time spent in LPA and MPA. WASO correlated negatively with inactivity.

CONCLUSION

This is one of the few studies with a wrist-worn accelerometer that shows a difference in sleep efficiency between MS patients and controls and, in MS, an association of sleep quality variables with PA variables.

Self-supervised learning of accelerometer data provides new insights for sleep and its association with mortality

Sleep is essential to life. Accurate measurement and classification of sleep/wake and sleep stages is important in clinical studies for sleep disorder diagnoses and in the interpretation of data from consumer devices for monitoring physical and mental well-being. Existing non-polysomnography sleep classification techniques mainly rely on heuristic methods developed in relatively small cohorts. Thus, we aimed to establish the accuracy of wrist-worn accelerometers for sleep stage classification and subsequently describe the association between sleep duration and efficiency (proportion of total time asleep when in bed) with mortality outcomes. We developed a self-supervised deep neural network for sleep stage classification using concurrent laboratory-based polysomnography and accelerometry. After exclusion, 1448 participant nights of data were used for training. The difference between polysomnography and the model classifications on the external validation was 34.7 min (95% limits of agreement (LoA): -37.8-107.2 min) for total sleep duration, 2.6 min for REM duration (95% LoA: -68.4-73.4 min) and 32.1 min (95% LoA: -54.4-118.5 min) for NREM duration. The sleep classifier was deployed in the UK Biobank with 100,000 participants to study the association of sleep duration and sleep efficiency with all-cause mortality. Among 66,214 UK Biobank participants, 1642 mortality events were observed. Short sleepers (<6 h) had a higher risk of mortality compared to participants with normal sleep duration of 6-7.9 h, regardless of whether they had low sleep efficiency (Hazard ratios (HRs): 1.58; 95% confidence intervals (CIs): 1.19-2.11) or high sleep efficiency (HRs: 1.45; 95% CIs: 1.16-1.81). Deep-learning-based sleep classification using accelerometers has a fair to moderate agreement with polysomnography. Our findings suggest that having short overnight sleep confers mortality risk irrespective of sleep continuity.

Associations between actigraphy-measured sleep duration, continuity, and timing with mortality in the UK Biobank

STUDY OBJECTIVES

To examine the associations between sleep duration, continuity, timing, and mortality using actigraphy among adults.

METHODS

Data were from a cohort of 88 282 adults (40-69 years) in UK Biobank that wore a wrist-worn triaxial accelerometer for 7 days. Actigraphy data were processed to generate estimates of sleep duration and other sleep characteristics including wake after sleep onset (WASO), number of 5-minute awakenings, and midpoint for sleep onset/wake-up and the least active 5 hours (L5). Data were linked to mortality outcomes with follow-up to October 31, 2021. We implemented Cox models (hazard ratio, confidence intervals [HR, 95% CI]) to quantify sleep associations with mortality. Models were adjusted for demographics, lifestyle factors, and medical conditions.

RESULTS

Over an average of 6.8 years 2973 deaths occurred (1700 cancer, 586 CVD deaths). Overall sleep duration was significantly associated with risk for all-cause (p < 0.01), cancer (p < 0.01), and CVD (p = 0.03) mortality. For example, when compared to sleep durations of 7.0 hrs/d, durations of 5 hrs/d were associated with a 29% higher risk for all-cause mortality (HR: 1.29 [1.09, 1.52]). WASO and number of awakenings were not associated with mortality. Individuals with L5 early or late midpoints (<2:30 or ≥ 3:30) had a ~20% higher risk for all-cause mortality, compared to those with intermediate L5 midpoints (3:00-3:29; p ≤ 0.01; e.g. HR ≥ 3:30: 1.19 [1.07, 1.32]).

CONCLUSIONS

Shorter sleep duration and both early and late sleep timing were associated with a higher mortality risk. These findings reinforce the importance of public health efforts to promote healthy sleep patterns in adults.

Waking Up to the Importance of Sleep in Type 2 Diabetes Management: A Narrative Review

For the first time, the latest American Diabetes Association/European Association for the Study of Diabetes (ADA/EASD) consensus guidelines have incorporated a growing body of evidence linking health outcomes associated with type 2 diabetes to the movement behavior composition over the whole 24-h day. Of particular note, the importance of sleep as a key lifestyle component in the management of type 2 diabetes is promulgated and presented using three key constructs: quantity, quality, and timing (i.e., chronotype). In this narrative review we highlight some of the key evidence justifying the inclusion of sleep in the latest consensus guidelines by examining the associations of quantity, quality, and timing of sleep with measures of glycemia, cardiovascular disease risk, and mortality. We also consider potential mechanisms implicated in the association between sleep and type 2 diabetes and provide practical advice for health care professionals about initiating conversations pertaining to sleep in clinical care. In particular, we emphasize the importance of measuring sleep in a free-living environment and provide a summary of the different methodologies and targets. In summary, although the latest ADA/EASD consensus report highlights sleep as a central component in the management of type 2 diabetes, placing it, for the first time, on a level playing field with other lifestyle behaviors (e.g., physical activity and diet), the evidence base for improving sleep (beyond sleep disorders) in those living with type 2 diabetes is limited. This review should act as a timely reminder to incorporate sleep into clinical consultations, ongoing diabetes education, and future interventions.

Is rest-activity rhythm prospectively associated with all-cause mortality in older people regardless of sleep and physical activity level? The 'Como Vai?' Cohort study

OBJECTIVE

This study aims to test the association of rest-activity rhythm (intradaily variability and interdaily stability) with all-cause mortality in an older adult cohort in Brazil. It also assesses whether the amount of time spent at each intensity level (i.e., physical activity and nocturnal sleep) interferes with this association.

METHODS

This cohort study started in 2014 with older adults (≥60 years). We investigated deaths from all causes that occurred until April 2017. Rest-activity rhythm variables were obtained using accelerometry at baseline. Intradaily variability indicates higher rhythm fragmentation, while interdaily stability indicates higher rhythm stability. Cox proportional-hazard models were used to test the associations controlling for confounders.

RESULTS

Among the 1451 older adults interviewed in 2014, 965 presented valid accelerometry data. During the follow-up period, 80 individuals died. After adjusting the analysis for sociodemographic, smoking, morbidity score, and number of medicines, an increase of one standard deviation in interdaily stability decreased 26% the risk of death. The adjustment for total sleep time and inactivity did not change this association. On the other hand, the association was no longer significant after adjusting for overall physical activity and moderate to vigorous physical activity.

CONCLUSION

Rest-activity rhythm pattern was not associated with mortality when physical activity was considered, possibly because this pattern could be driven by regular exercise. Promoting physical activity remains a relevant strategy to improve population health.

Self-supervised learning of accelerometer data provides new insights for sleep and its association with mortality

Background

Sleep is essential to life. Accurate measurement and classification of sleep/wake and sleep stages is important in clinical studies for sleep disorder diagnoses and in the interpretation of data from consumer devices for monitoring physical and mental well-being. Existing non-polysomnography sleep classification techniques mainly rely on heuristic methods developed in relatively small cohorts. Thus, we aimed to establish the accuracy of wrist-worn accelerometers for sleep stage classification and subsequently describe the association between sleep duration and efficiency (proportion of total time asleep when in bed) with mortality outcomes.

Methods

We developed and validated a self-supervised deep neural network for sleep stage classification using concurrent laboratory-based polysomnography and accelerometry data from three countries (Australia, the UK, and the USA). The model was validated within-cohort using subject-wise five-fold cross-validation for sleep-wake classification and in a three-class setting for sleep stage classification wake, rapid-eye-movement sleep (REM), non-rapid-eye-movement sleep (NREM) and by external validation. We assessed the face validity of our model for population inference by applying the model to the UK Biobank with 100,000 participants, each of whom wore a wristband for up to seven days. The derived sleep parameters were used in a Cox regression model to study the association of sleep duration and sleep efficiency with all-cause mortality.

Findings

After exclusion, 1,448 participant nights of data were used to train the sleep classifier. The difference between polysomnography and the model classifications on the external validation was 34.7 minutes (95% limits of agreement (LoA): -37.8 to 107.2 minutes) for total sleep duration, 2.6 minutes for REM duration (95% LoA: -68.4 to 73.4 minutes) and 32.1 minutes (95% LoA: -54.4 to 118.5 minutes) for NREM duration. The derived sleep architecture estimate in the UK Biobank sample showed good face validity. Among 66,214 UK Biobank participants, 1,642 mortality events were observed. Short sleepers (<6 hours) had a higher risk of mortality compared to participants with normal sleep duration (6 to 7.9 hours), regardless of whether they had low sleep efficiency (Hazard ratios (HRs): 1.69; 95% confidence intervals (CIs): 1.28 to 2.24 ) or high sleep efficiency (HRs: 1.42; 95% CIs: 1.14 to 1.77).

Interpretation

Deep-learning-based sleep classification using accelerometers has a fair to moderate agreement with polysomnography. Our findings suggest that having short overnight sleep confers mortality risk irrespective of sleep continuity.

Opal Actigraphy (Activity and Sleep) Measures Compared to ActiGraph: A Validation Study

Physical activity and sleep monitoring in daily life provide vital information to track health status and physical fitness. The aim of this study was to establish concurrent validity for the new Opal Actigraphy solution in relation to the widely used ActiGraph GT9X for measuring physical activity from accelerometry epic counts (sedentary to vigorous levels) and sleep periods in daily life. Twenty participants (age 56 + 22 years) wore two wearable devices on each wrist for 7 days and nights, recording 3-D accelerations at 30 Hz. Bland-Altman plots and intraclass correlation coefficients (ICCs) assessed validity (agreement) and test-retest reliability between ActiGraph and Opal Actigraphy sleep durations and activity levels, as well as between the two different versions of the ActiGraph. ICCs showed excellent reliability for physical activity measures and moderate-to-excellent reliability for sleep measures between Opal versus Actigraph GT9X and between GT3X versus GT9X. Bland-Altman plots and mean absolute percentage error (MAPE) also show a comparable performance (within 10%) between Opal and ActiGraph and between the two ActiGraph monitors across activity and sleep measures. In conclusion, physical activity and sleep measures using Opal Actigraphy demonstrate performance comparable to that of ActiGraph, supporting concurrent validation. Opal Actigraphy can be used to quantify activity and monitor sleep patterns in research and clinical studies.