DPSleep: Open-Source Longitudinal Sleep Analysis From Accelerometer Data (Preprint)

2021 ◽  
Author(s):  
Habiballah Rahimi-Eichi ◽  
Garth Coombs 3rd ◽  
Constanza M. Vidal Bustamante ◽  
Jukka-Pekka Onnela ◽  
Justin T. Baker ◽  
...  
Keyword(s):  
Open Source ◽  
Self Report ◽  
Activity Levels ◽  
Accelerometer Data ◽  
Time Zone ◽  
Sleep Episode ◽  
Sleep Timing ◽  
Sleep Analysis ◽  
Smartphone Use ◽  
Deep Phenotyping

UNSTRUCTURED Wearable devices are now widely available to collect continuous objective behavioral data from individuals and to measure sleep. Here we introduce an open-source pipeline for the deep phenotyping of sleep, "DPSleep", that uses algorithms to detect missing data, calculate activity levels, and finally estimate the major Sleep Episode onset and offset. The pipeline allows for manual quality control adjustment and correction for time zone changes. We illustrate the utility of the pipeline with data from participants studied for more than 200 days. Actigraphy-based measures of sleep duration are associated with self-report rating of sleep quality. Simultaneous measures of smartphone use and GPS data support the sleep timing inferences and reveal how phone measures of sleep can differ from actigraphy data. We discuss the uses of DPSleep in relation to other available sleep estimation approaches and provide example use cases that include multi-dimensional, deep dynamic longitudinal phenotyping associated with mental illness.

scholarly journals DPSleep: Open-Source Longitudinal Sleep Analysis From Accelerometer Data

2021 ◽  
Author(s):  
Habiballah Rahimi-Eichi ◽  
Garth Coombs ◽  
Constanza M. Vidal Bustamante ◽  
Jukka-Pekka Onnela ◽  
Justin T. Baker ◽  
...  
Keyword(s):  
Electronic Device ◽  
Spectral Power ◽  
Sleep Onset ◽  
Self Report ◽  
Accelerometer Data ◽  
Time Zone ◽  
Sliding Windows ◽  
Sleep Episode ◽  
Sleep Timing ◽  
Wide Range

Wearable devices are now widely available to collect continuous objective behavioral data from individuals and to measure sleep. Here we introduce a pipeline to infer sleep onset, duration, and quality from raw accelerometer data and then quantify relationships between derived sleep metrics and other variables of interest. The pipeline released here for the deep phenotyping of sleep, as the “DPSleep” software package, uses (a) a stepwise algorithm to detect missing data; (b) within-individual, minute-based, spectral power percentiles of activity; and (c) iterative, forward- and backward-sliding windows to estimate the major Sleep Episode onset and offset. Software modules allow for manual quality control adjustment of derived sleep features and correction for time zone changes. In this report, we illustrate the pipeline with data from participants studied for more than 200 days each. Actigraphy-based measures of sleep duration are associated with self-report rating of sleep quality. Simultaneous measures of smartphone use and GPS location data support the validity of the sleep timing inferences and reveal how phone measures of sleep timing can differ from actigraphy data. We discuss the uses of DPSleep in relation to other available sleep estimation approaches and provide example use cases that include multi-dimensional, deep longitudinal phenotyping, extended measurement of dynamics associated with mental illness, and the possibility of combining wearable actigraphy and personal electronic device data (e.g., smartphone, tablet) to measure individual differences across a wide range of behavioral variation in health and disease.

scholarly journals DPSleep: Open-Source Longitudinal Sleep Analysis From Accelerometer Data (Preprint)

10.2196/29849 ◽  
2021 ◽  
Author(s):  
Habiballah Rahimi-Eichi ◽  
Garth Coombs 3rd ◽  
Constanza M. Vidal Bustamante ◽  
Jukka-Pekka Onnela ◽  
Justin T. Baker ◽  
...  
Keyword(s):  
Open Source ◽  
Accelerometer Data ◽  
Sleep Analysis

scholarly journals Deep Learning Classification of Canine Behavior Using a Single Collar-Mounted Accelerometer: Real-World Validation

Animals ◽  
2021 ◽  
Vol 11 (6) ◽  
pp. 1549
Author(s):  
Robert D. Chambers ◽  
Nathanael C. Yoder ◽  
Aletha B. Carson ◽  
Christian Junge ◽  
David E. Allen ◽  
...  
Keyword(s):  
Machine Learning ◽  
Deep Learning ◽  
Real World ◽  
Learning Algorithm ◽  
Drinking Behavior ◽  
True Positive Rate ◽  
Training Dataset ◽  
Activity Levels ◽  
Accelerometer Data ◽  
Activity Monitors

Collar-mounted canine activity monitors can use accelerometer data to estimate dog activity levels, step counts, and distance traveled. With recent advances in machine learning and embedded computing, much more nuanced and accurate behavior classification has become possible, giving these affordable consumer devices the potential to improve the efficiency and effectiveness of pet healthcare. Here, we describe a novel deep learning algorithm that classifies dog behavior at sub-second resolution using commercial pet activity monitors. We built machine learning training databases from more than 5000 videos of more than 2500 dogs and ran the algorithms in production on more than 11 million days of device data. We then surveyed project participants representing 10,550 dogs, which provided 163,110 event responses to validate real-world detection of eating and drinking behavior. The resultant algorithm displayed a sensitivity and specificity for detecting drinking behavior (0.949 and 0.999, respectively) and eating behavior (0.988, 0.983). We also demonstrated detection of licking (0.772, 0.990), petting (0.305, 0.991), rubbing (0.729, 0.996), scratching (0.870, 0.997), and sniffing (0.610, 0.968). We show that the devices’ position on the collar had no measurable impact on performance. In production, users reported a true positive rate of 95.3% for eating (among 1514 users), and of 94.9% for drinking (among 1491 users). The study demonstrates the accurate detection of important health-related canine behaviors using a collar-mounted accelerometer. We trained and validated our algorithms on a large and realistic training dataset, and we assessed and confirmed accuracy in production via user validation.

Sensitivity to financial rewards and impression management links to smartphone use and dependence

2021 ◽  
Vol 25 (2) ◽  
pp. 107-128
Author(s):  
Graham Pluck ◽  
◽  
Pablo Emilio Barrera Falconi ◽  
◽  
◽  
...  
Keyword(s):  
Impression Management ◽  
Technology Use ◽  
Experimental Studies ◽  
Negative Relationship ◽  
Reward Sensitivity ◽  
Self Report ◽  
Smartphone Addiction ◽  
Observational Measures ◽  
Instant Messenger ◽  
Smartphone Use

Computational modeling and brain imaging studies suggest that sensitivity to rewards and behaviorist learning principles partly explain smartphone engagement patterns and potentially smartphone dependence. Responses to a questionnaire, and observational measures of smartphone use were recorded for 121 university students. Each participant was also tested with a laboratory task of reward sensitivity and a test of verbal operant conditioning. Twenty-three percent of the sample had probable smartphone addiction. Using multivariate regression, smartphone use, particularly the number of instant messenger services employed, was shown to be significantly and independently predicted by reward sensitivity (a positive relationship), and by instrumental conditioning (a negative relationship). However, the latter association was driven by a subset of participants who developed declarative knowledge of the response-reinforcer contingency. This suggests a process of impression management driven by experimental demand characteristics, producing goal-directed instrumental behavior not habit-based learning. No other measures of smartphone use, including the self-report scale, were significantly associated with the experimental tasks. We conclude that stronger engagement with smartphones, in particular instant messenger services, may be linked to people being more sensitive to rewarding stimuli, suggestive of a motivational or learning mechanism. We propose that this mechanism could underly problem smartphone use and dependence. It also potentially explains why some aspects of smartphone use, such as habitual actions, appear to be poorly measured by technology-use questionnaires. A serendipitous secondary finding confirmed that smartphone use reflected active self-presentation. Our ‘conditioning’ task-induced this behavior in the laboratory and could be used in social-cognition experimental studies.

Converting Raw Accelerometer Data to Activity Counts Using Open-Source Code: Implementing a MATLAB Code in Python and R, and Comparing the Results to ActiLife

2021 ◽  
pp. 1-7
Author(s):  
Ruben Brondeel ◽  
Yan Kestens ◽  
Javad Rahimipour Anaraki ◽  
Kevin Stanley ◽  
Benoit Thierry ◽  
...  
Keyword(s):  
Open Source ◽  
Pearson Correlation ◽  
Vertical Axis ◽  
Accelerometer Data ◽  
Accelerometer Sensor ◽  
Interclass Correlation ◽  
Acceleration Data ◽  
The Right ◽  
Pearson Correlations ◽  
Closed Source

Background: Closed-source software for processing and analyzing accelerometer data provides little to no information about the algorithms used to transform acceleration data into physical activity indicators. Recently, an algorithm was developed in MATLAB that replicates the frequently used proprietary ActiLife activity counts. The aim of this software profile was (a) to translate the MATLAB algorithm into R and Python and (b) to test the accuracy of the algorithm on free-living data. Methods: As part of the INTErventions, Research, and Action in Cities Team, data were collected from 86 participants in Victoria (Canada). The participants were asked to wear an integrated global positioning system and accelerometer sensor (SenseDoc) for 10 days on the right hip. Raw accelerometer data were processed in ActiLife, MATLAB, R, and Python and compared using Pearson correlation, interclass correlation, and visual inspection. Results: Data were collected for a combined 749 valid days (>10 hr wear time). MATLAB, Python, and R counts per minute on the vertical axis had Pearson correlations with the ActiLife counts per minute of .998, .998, and .999, respectively. All three algorithms overestimated ActiLife counts per minute, some by up to 2.8%. Conclusions: A MATLAB algorithm for deriving ActiLife counts was implemented in R and Python. The different implementations provide similar results to ActiLife counts produced in the closed source software and can, for all practical purposes, be used interchangeably. This opens up possibilities to comparing studies using similar accelerometers from different suppliers, and to using free, open-source software.

083 Circadian Preference is Associated with Impulsivity at Both the Trait and State Level

SLEEP ◽  
2021 ◽  
Vol 44 (Supplement_2) ◽  
pp. A35-A35
Author(s):  
Brant Hasler ◽  
Meredith Wallace ◽  
Jessica Graves ◽  
Sarah Pedersen
Keyword(s):  
Substance Use ◽  
White Women ◽  
Multilevel Models ◽  
State Level ◽  
Self Report ◽  
Future Research ◽  
Sleep Timing ◽  
Composite Scale ◽  
Primary Sample ◽  
Circadian Preference

Abstract Introduction Impulsivity is a multifaceted construct with well-documented risk for substance use problems. A circadian preference towards eveningness has been linked to trait, global impulsivity. Here we extend existing literature by investigating whether eveningness is associated with multiple facets of impulsivity at both trait- and state-level impulsivity. We also examined these associations utilizing daily measures of sleep timing and duration. Methods The primary sample included 78 moderate-to-heavy social drinkers (aged 21–35, 100% White men) with circadian preference data (Composite Scale of Morningness: CSM). Five facets of impulsivity were assessed via the UPPS-P, both at baseline (full scale) and up to 6 times per day over 10 days (reduced scale). Daily sleep timing (midsleep) and duration were assessed via self-report over 10 days. Multilevel models were used to examine between- and within-person associations, accounting for covariates and correcting for multiple comparisons. Results Between-person models found that eveningness was associated with multiple facets of impulsivity, at trait (lack of perseverance) and state levels (negative urgency, positive urgency, lack of perseverance, and lack of premeditation). However, average midsleep and duration were generally unrelated to impulsivity when accounting for circadian preference. Within-person models in the primary sample largely paralleled the between-person findings. In a larger, more diverse sample (29.1% self-identified as Black, 29.7% female) without CSM data, later midsleep timing was associated with greater mean state-level impulsivity across multiple facets. These effects largely appear to be driven by White women. Conclusion A circadian preference for eveningness is strongly associated with multiple facets of impulsivity, at both trait- and state-levels, potentially increasing risk for substance use. This association does not appear to be driven by actual daily sleep timing and/or duration. Future research with objective measures of sleep in larger, more diverse samples will be important to clarify implications for sleep-focused prevention and/or treatment of substance use. Support (if any) Supported by grants from NIH (R01AA026249; K01 AA021135), as well as a Foundation Grant from ABMRF/The Foundation for Alcohol Research.

scholarly journals Mode of Commuting to School and Its Association with Physical Activity and Sedentary Habits in Young Ecuadorian Students

2018 ◽  
Vol 15 (12) ◽  
pp. 2704 ◽  
Author(s):  
Yaira Barranco-Ruiz ◽  
Alfredo Guevara-Paz ◽  
Robinson Ramírez-Vélez ◽  
Palma Chillón ◽  
Emilio Villa-González
Keyword(s):  
Physical Activity ◽  
Central Region ◽  
School Setting ◽  
Daily Physical Activity ◽  
Self Report ◽  
Activity Levels ◽  
Active Commuting ◽  
Physical Activity Levels ◽  
Older Adolescents ◽  
Commuting Distance

Active commuting to and from school (ACS) could help to increase daily physical activity levels in youth; however, this association remains unknown in Ecuadorian youth. Thus, the aims of this study were (1) to determine the patterns of commuting to and from school and (2) to analyze the associations between ACS, physical activity (PA), and sedentary habits in Ecuadorian youth. A total of 732 students (65.3% males), aged 10–18 years (children = 246, young adolescents = 310, older adolescents = 162) from the central region of Ecuador participated in this study. A self-report questionnaire, including the usual mode and frequency of commuting, distance from home to school (PACO-Questionnaire), and PA and sedentary habits (YAP-Questionnaire), was used. Most of the sample lived ≤2 km from school; however, they were mainly passive commuters (96%). The most common mode of commuting was by car (to school = 43.4%, from school = 31.6%; p < 0.001). Children presented significantly higher scores (0–4) in PA outside school and total PA compared with older adolescents (2.20 ± 0.97 vs. 1.97 ± 0.96; p = 0.013 and 2.30 ± 0.76 vs. 2.09 ± 0.74, p = 0.019, respectively), as well as the lowest scores in sedentary habits (1.51 ± 0.65, p < 0.001). PA at school and total PA were positively associated with ACS (OR 3.137; 95% CI, 1.918 to 5.131; p < 0.001, and OR 2.543; 95% CI, 1.428 to 4.527; p = 0.002, respectively). In conclusion, passive modes of transportation were the most frequently used to commute to and from school in young Ecuadorians. PA at school and total PA were positively associated with ACS. Thus, interventions at school setting could be an opportunity to improve PA levels and additionally ACS in youth from the central region of Ecuador.

scholarly journals Effects of Body Mass Index on Bone Loading Due to Physical Activity

2018 ◽  
Vol 34 (1) ◽  
pp. 7-13
Author(s):  
Tina Smith ◽  
Sue Reeves ◽  
Lewis G. Halsey ◽  
Jörg Huber ◽  
Jin Luo
Keyword(s):  
Physical Activity ◽  
International Physical Activity Questionnaire ◽  
Short Form ◽  
Effect Sizes ◽  
Moderate Activity ◽  
Activity Levels ◽  
Accelerometer Data ◽  
Light Activity ◽  
Bone Loading ◽  
Activity Questionnaire

The aim of the current study was to compare bone loading due to physical activity between lean, and overweight and obese individuals. Fifteen participants (lower BMI group: BMI < 25 kg/m2, n = 7; higher BMI group: 25 kg/m2 < BMI < 36.35 kg/m2, n = 8) wore a tri-axial accelerometer on 1 day to collect data for the calculation of bone loading. The International Physical Activity Questionnaire (short form) was used to measure time spent at different physical activity levels. Daily step counts were measured using a pedometer. Differences between groups were compared using independent t-tests. Accelerometer data revealed greater loading dose at the hip in lower BMI participants at a frequency band of 0.1–2 Hz (P = .039, Cohen’s d = 1.27) and 2–4 Hz (P = .044, d = 1.24). Lower BMI participants also had a significantly greater step count (P = .023, d = 1.55). This corroborated with loading intensity (d ≥ 0.93) and questionnaire (d = 0.79) effect sizes to indicate higher BMI participants tended to spend more time in very light activity, and less time in light and moderate activity. Overall, participants with a lower BMI exhibited greater bone loading due to physical activity; participants with a higher BMI may benefit from more light and moderate level activity to maintain bone health.

scholarly journals Methodology Used in Ecological Momentary Assessment Studies About Sedentary Behavior in Children, Adolescents, and Adults: Systematic Review Using the Checklist for Reporting Ecological Momentary Assessment Studies (Preprint)

2018 ◽  
Author(s):  
Catiana Leila Possamai Romanzini ◽  
Marcelo Romanzini ◽  
Mariana Biagi Batista ◽  
Cynthia Correa Lopes Barbosa ◽  
Gabriela Blasquez Shigaki ◽  
...  
Keyword(s):  
Sedentary Behavior ◽  
Ecological Momentary Assessment ◽  
Web Of Science ◽  
Review Literature ◽  
Self Report ◽  
Accelerometer Data ◽  
Adolescents And Adults ◽  
Ecological Momentary ◽  
Momentary Assessment

BACKGROUND The use of ecological momentary assessment (EMA) to measure sedentary behavior (SB) in children, adolescents, and adults can increase the understanding of the role of the context of SB in health outcomes. OBJECTIVE The aim of this study was to systematically review literature to describe EMA methodology used in studies on SB in youth and adults, verify how many studies adhere to the Methods aspect of the Checklist for Reporting EMA Studies (CREMAS), and detail measures used to assess SB and this associated context. METHODS A systematic literature review was conducted in the PubMed, Scopus, Web of Science, PsycINFO, Cumulative Index to Nursing and Allied Health Literature (CINAHL), and SPORTDiscus databases, covering the entire period of existence of the databases until January 2018. RESULTS This review presented information about the characteristics and methodology used in 21 articles that utilized EMA to measure SB in youth and adults. There were more studies conducted among youth compared with adults, and studies of youth included more waves and more participants (n=696) than studies with adults (n=97). Most studies (85.7%) adhered to the Methods aspect of the CREMAS. The main criteria used to measure SB in EMA were self-report (81%) with only 19% measuring SB using objective methods (eg, accelerometer). The main equipment to collect objective SB was the ActiGraph, and the cutoff point to define SB was <100 counts/min. Studies most commonly used a 15-min window to compare EMA and accelerometer data. CONCLUSIONS The majority of studies in this review met minimum CREMAS criteria for studies conducted with EMA. Most studies measured SB with EMA self-report (n=17; 81.0%), and a few studies also used objective methods (n=4; 19%). The standardization of the 15-min window criteria to compare EMA and accelerometer data would lead to a comparison between these and new studies. New studies using EMA with mobile phones should be conducted as they can be considered an attractive method for capturing information about the specific context of SB activities of young people and adults in real time or very close to it.

scholarly journals Capturing sleep-wake cycles by using day-to-day smartphone touchscreen interactions

2018 ◽  
Author(s):  
Jay N. Borger ◽  
Reto Huber ◽  
Arko Ghosh
Keyword(s):  
Cognitive Activity ◽  
Sleep Onset ◽  
Activity Levels ◽  
Body Movements ◽  
New Approach ◽  
Sleep Period ◽  
Physical Activity Levels ◽  
Smartphone Use ◽  
Waking Up ◽  
Sleeping Habits

AbstractBody movements drop with sleep and this behavioural signature is widely exploited to infer sleep duration. However, a reduction in body movements may also occur in periods of intense cognitive activity and the ubiquitous use of smartphones may capture these wakeful periods otherwise hidden in the standard measures of sleep. Here we continuously captured the gross body movements using standard wrist-worn accelerometers to quantify sleep (actigraphy) and logged the timing of the day-to-day touchscreen events (‘tappigraphy’). Using these measures, we addressed how the gross body movements overlap with the cognitively engaging digital behaviour (from n = 84 individuals, accumulating 1384 nights). We find that smartphone use was distributed across a broad spectrum of physical activity levels but consistently peaked at rest. We estimated the putative sleep onset and wake-up times from the actigraphy data to find that these times were well correlated to the estimates from tappigraphy (R2= 0.9 for sleep onset and wake-up time). However, actigraphy overestimated sleep as virtually all of the users used their phones during the putative sleep period. Interestingly, the probability of touches remained greater than zero for ~ 2 h after the putative sleep onset and ~ 2 h before the putative wake-up time. Our findings suggest that touchscreen interactions are widely integrated into modern sleeping habits – surrounding both sleep onset and waking-up periods – yielding a new approach to measuring sleep. Smartphone taps can be leveraged to update the behavioural signatures of sleep with these peculiarities of modern digital behaviour.

Sign in / Sign up

Export Citation Format

Share Document