Predicting sleeping behaviors in long-term studies with wrist-worn sensor data (original) (raw)
Related papers
The Science of Sweet Dreams: Predicting Sleep Efficiency from Wearable Device Data
Sleep is an important human behavior with a deep impact on quality of life. Inadequate sleep quality negatively affects both mental and physical well-being, and exacerbates many health problems such as diabetes, depression, cancer and obesity. Alarmingly, poor sleep is becoming a growing concern in our society. Increased efforts toward the development of sleep science, focus on the study of sleep and its impact on overall human health. Technology has become a crucial component for this research, particularly the use of wearable devices for capturing and analysing human activity. In this article, we explore the use of wearables in sleep science, including the algorithmic development of robust human activity recognition and predictive methodologies for sleep quality estimation, which have shown an AUC improvement up to twofold over current actigraphy methods by 50%. These technologies are crucial for the evolution of new application services to assist behavioural health decision-making for patients and professionals.
0443 Intra Week Sleep Patterns Analyzed Using Consumer Sleep Tracker Data
Sleep, 2020
Introduction: Big data collected using consumer sleep technology can provide objectively measured insights on sleep behavior in the real-life environment. It has the advantage over self-report data of being less prone to bias. Here we used a non-contact bio-motion sensor to remotely capture objective sleep data. We analyzed 168432 nights of sleep data to test if differences between weekday versus weekend sleep behavior, known from self-report, would still hold using objective data in a large population. Methods: Sleep data was acquired using the SleepScore Max remote sleep sensor and included 168432 nights (2730 users, mean age: 46.6 +/-11.8 years, 33% female, all resident in the USA). Analysis was restricted to those of working age; adults between 20-65. Any sleep which ended from Monday to Friday was considered weekday sleep, and any ending on Saturday or Sunday as weekend sleep. Data records were inspected and cleaned before analyzing. Descriptive statistics and independent t-tests were used to analyze the data. Results: Total Sleep Time, Time In Bed and Sleep Onset Latencies were longer during weekend (TST: + 20.6 mins, TIB: +22.9 mins, SOL: +1.1 min, all p <0.001), resulting in a slightly poorer Sleep Efficiency (-.016%, p<0.01) for weekend nights. Time to bed and final awakening were both delayed in weekends as compared to weekdays (Time to bed +30.0 mins, and final awakening +53.4 mins, both p<0.001). Conclusion: This big data analysis confirms the earlier observed difference in sleep and sleep behavior between weekdays and weekends. This should be considered for optimizing (automated) sleep interventions, that may not normally take the weekend effect into consideration.
An smartphone-based algorithm to measure and model quantity of sleep
2015
Sleep quantity affects an individual's personal health. The gold standard of measuring sleep and diagnosing sleep disorders is Polysomnography (PSG). Although PSG is accurate, it is expensive and it lacks portability. A number of wearable devices with embedded sensors have emerged in the recent past as an alternative to PSG for regular sleep monitoring directly by the user. These devices are intrusive and cause discomfort besides being expensive. In this work, we present an algorithm to detect sleep using a smartphone with the help of its inbuilt accelerometer sensor. We present three different approaches to classify raw acceleration data into two states - Sleep and Wake. In the first approach, we take an equation from Kushida's algorithm to process accelerometer data. Henceforth, we call it Kushida's equation. While the second is based on statistical functions, the third is based on Hidden Markov Model (HMM) training. Although all the three approaches are suitable for a...
Validation of Capturing Sleep Diary Data via a Wrist-Worn Device
Sleep disorders, 2015
Paper sleep diaries are the gold standard for assessment of sleep continuity variables in clinical practice as well as research. Unfortunately, paper diaries can be filled out weekly instead of daily, lost, illegible or destroyed; and are considered out of date according to the newer technology savvy generations. In this study, we assessed the reliability and validity of using a wrist-worn electronic sleep diary. Design. A prospective design was used to compare capturing 14 days of sleep continuity data via paper to a wrist-worn electronic device that also captured actigraphy data. Results. Thirty-five healthy community dwelling adults with mean (sd) age of 36 (15), 80% Caucasians, and 74% females were enrolled. All sleep continuity variables via electronic and paper diary capture methods were significantly correlated with moderate, positive relationships. Assessment of validity revealed that electronic data capture had a significant relationship with objective measure of sleep cont...
Sleep
Study Objectives Multisensor wearable consumer devices allowing the collection of multiple data sources, such as heart rate and motion, for the evaluation of sleep in the home environment, are increasingly ubiquitous. However, the validity of such devices for sleep assessment has not been directly compared to alternatives such as wrist actigraphy or polysomnography (PSG). Methods Eight participants each completed four nights in a sleep laboratory, equipped with PSG and several wearable devices. Registered polysomnographic technologist-scored PSG served as ground truth for sleep–wake state. Wearable devices providing sleep–wake classification data were compared to PSG at both an epoch-by-epoch and night level. Data from multisensor wearables (Apple Watch and Oura Ring) were compared to data available from electrocardiography and a triaxial wrist actigraph to evaluate the quality and utility of heart rate and motion data. Machine learning methods were used to train and test sleep–wake...
Int. J. Syst. Dyn. Appl., 2021
Maintaining the suited amount of sleep is considered the prime component for maintaining a proper and adequate health condition. Often it has been observed that people having sleep inconsistency tend to jeopardize the health and appeal to many physiological and psychological disorders. To overcome such difficulties, it is often required to keep a requisite note of the duration and quality of sleep that one is having. This work defines an algorithm that can be utilized in smart wearables or mobile phones to perceive the duration of sleep and also to classify a particular instance as slept or awake on the basis of data fetched from the triaxial accelerometer. A comparative analysis was performed based on the results obtained from some previously developed algorithms, rule-based models, and machine learning models, and it was observed that the algorithm developed in the work outperformed the previously developed algorithms. Moreover, the algorithm developed in the work will very much d...
Sustained logging and discrimination of sleep postures ith low-level, wrist-worn sensors
Proceedings - International Symposium on Wearable Computers, ISWC, 2008
We present a study which evaluates the use of simple lowpower sensors for a long-term, coarse-grained detection of sleep postures. In contrast to the information-rich but complex recording methods used in sleep studies, we follow a paradigm closer to that of actigraphy by using a wrist-worn device that continuously logs and processes data from the user. Experiments show that it is feasible to detect nightly sleep periods with a combination of light and simple motion and posture sensors, and to detect within these segments what basic sleeping postures the user assumes. These findings can be of value in several domains, such as monitoring of sleep apnea disorders, and support the feasibility of a continuous home-monitoring of sleeping trends where users wear the sensor device uninterruptedly for weeks to months in a row.
Empowering Wearable Sensor Generated Data to Predict Changes in Individual's Sleep Quality
The 6th International Conference on Information and Communication Technology (ICoICT 2018)), 2018
Wearable sensors found in popular wrist wearable device are both generating sales profit and constantly generating vast amount of data. Some of these wearable sensors are able to record physical activity and sleep trends, both are being mainly used to give insight to its users about their current and past health and well-being. We proposed a method of data pre-processing and machine learning using simple k-nearest neighbor classifier to furthermore empower the usage of such data to predict changes in one's sleep quality based on his or her current physical activity level. Our method were challenged to predict changes in five medically-approved sleep quality indicators, using data generated by commercially available consumer-grade wrist wearable device. The experiment result shows that the successful prediction of changes in sleep quality using wearable sensor generated data can be achieved by successfully selecting and sometimes combining the right input parameter(s). Each sleep quality indicators calls for different input parameter or combined parameters. By selecting and combining the right parameter(s), our method had successfully predict changes in both sleep duration and sleep efficiency with accuracy of 68% and 64%, respectively.
Is it on? An algorithm for discerning wrist-accelerometer non-wear times from sleep/wake activity
The accuracy of sleep/wake estimates derived with actigraphy is often dependent on researchers being able to discern non-wear times from sleep or quiescent wakefulness when confronted by discrepancies in a sleep log. Without knowing when an accelerometer is being worn, non-wear could be inferred from periods of inactivity unlikely to occur while in bed. Data collected in our laboratory suggest that more than 50% of inactive periods during time in bed are <8 min in duration. This duration may be an appropriate minimum threshold for routine non-wear classification during self-reported wake. Higher thresholds could be chosen to derive non-wear definitions for self-reported bedtimes depending on the desired level of certainty. To determine non-wear at thresholds of 75%, 95% and 99%, for example, would require periods of inactivity lasting ≥18 min, ≥53 min and ≥85 min, respectively.
Combining wearable and environmental sensing into an unobtrusive tool for long-term sleep studies
IHI'12 - Proceedings of the 2nd ACM SIGHIT International Health Informatics Symposium, 2012
Long-term sleep monitoring of patients has been identified as a useful tool to observe sleep trends manifest themselves over weeks or months for use in behavioral studies. In practice, this has been limited to coarse-grained methods such as actigraphy, for which the levels of activity are logged, and which provide some insight but have simultaneously been found to lack accuracy to be used for studying sleeping disorders . This paper presents a method to automatically detect the user's sleep at home on a long-term basis. Inertial, ambient light, and time data tracked from a wrist-worn sensor, and additional night vision footage is used for later expert inspection. An evaluation on over 4400 hours of data from a focus group of test subjects demonstrates a high recall night segment detection, obtaining an average of 94%. Further, a clustering to visualize reoccurring sleep patterns is presented, and a myoclonic twitch detection is introduced, which exhibits a precision of 74%. The results indicate that long-term sleep pattern detections are feasible.