Sleep classification from wrist-worn accelerometer data using random forests
dc.contributor.author | Sundararajan, K | |
dc.contributor.author | Georgievska, S | |
dc.contributor.author | te Lindert, BHW | |
dc.contributor.author | Gehrman, PR | |
dc.contributor.author | Ramautar, J | |
dc.contributor.author | Mazzotti, DR | |
dc.contributor.author | Sabia, S | |
dc.contributor.author | Weedon, MN | |
dc.contributor.author | van Someren, EJW | |
dc.contributor.author | Ridder, L | |
dc.contributor.author | Wang, J | |
dc.contributor.author | van Hees, VT | |
dc.date.accessioned | 2021-09-10T13:35:00Z | |
dc.date.issued | 2021-01-08 | |
dc.description.abstract | Accurate and low-cost sleep measurement tools are needed in both clinical and epidemiological research. To this end, wearable accelerometers are widely used as they are both low in price and provide reasonably accurate estimates of movement. Techniques to classify sleep from the high-resolution accelerometer data primarily rely on heuristic algorithms. In this paper, we explore the potential of detecting sleep using Random forests. Models were trained using data from three different studies where 134 adult participants (70 with sleep disorder and 64 good healthy sleepers) wore an accelerometer on their wrist during a one-night polysomnography recording in the clinic. The Random forests were able to distinguish sleep-wake states with an F1 score of 73.93% on a previously unseen test set of 24 participants. Detecting when the accelerometer is not worn was also successful using machine learning (F1-score > 93.31 %), and when combined with our sleep detection models on day-time data provide a sleep estimate that is correlated with self-reported habitual nap behaviour (r =. 60). These Random forest models have been made open-source to aid further research. In line with literature, sleep stage classification turned out to be difficult using only accelerometer data. | en_GB |
dc.description.sponsorship | Medical Research Council (MRC) | en_GB |
dc.description.sponsorship | Diabetes UK | en_GB |
dc.identifier.citation | Vol. 11, article 24 | en_GB |
dc.identifier.doi | 10.1038/s41598-020-79217-x | |
dc.identifier.grantnumber | 16/72/18 | en_GB |
dc.identifier.grantnumber | 17/0005700 | en_GB |
dc.identifier.uri | http://hdl.handle.net/10871/127052 | |
dc.language.iso | en | en_GB |
dc.publisher | Nature Research | en_GB |
dc.rights | © The Author(s) 2021. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. | en_GB |
dc.subject | Epidemiology | en_GB |
dc.subject | Neurophysiology | en_GB |
dc.subject | Sleep disorders | en_GB |
dc.title | Sleep classification from wrist-worn accelerometer data using random forests | en_GB |
dc.type | Article | en_GB |
dc.date.available | 2021-09-10T13:35:00Z | |
dc.identifier.issn | 2045-2322 | |
dc.description | This is the final version. Available from Nature Research via the DOI in this record. | en_GB |
dc.description | The classification models developed in this paper are available as open access data on Zenodo29. The R30 package GGIR was previously developed for the processing of accelerometer data22. We enhanced GGIR to be able to embed the sleep classification models written in Python as explained in the GGIRpackageVignette31. Specific code to use this functionality in combination with the models from this paper can be found here. The combination of the code and GGIR package allow for sleep classification and nonwear classification of raw accelerometer data. This involves data extraction, pre-processing, feature extraction, and sleep or nonwear classification. Raw data from the polysomnography study in Newcastle has been made open access available in anonymized format on zenodo.org32. Data from the University of Pennsylvania are available through the National Institute of Mental Health data archive. Whitehall II data, protocols, and other metadata are available to the scientific community. Please refer to the Whitehall II data sharing policy which can be found here. | en_GB |
dc.identifier.journal | Scientific Reports | en_GB |
dc.rights.uri | http://creativecommons.org/licenses/by/4.0/ | en_GB |
dcterms.dateAccepted | 2020-11-24 | |
exeter.funder | ::Medical Research Council (MRC) | en_GB |
exeter.funder | ::Diabetes UK | en_GB |
rioxxterms.version | VoR | en_GB |
rioxxterms.licenseref.startdate | 2021-01-08 | |
rioxxterms.type | Journal Article/Review | en_GB |
refterms.dateFCD | 2021-09-10T13:30:25Z | |
refterms.versionFCD | VoR | |
refterms.dateFOA | 2021-09-10T13:35:07Z | |
refterms.panel | A | en_GB |
Files in this item
This item appears in the following Collection(s)
Except where otherwise noted, this item's licence is described as © The Author(s) 2021. Open Access This article is licensed under a Creative Commons Attribution 4.0 International
License, which permits use, sharing, adaptation, distribution and reproduction in any medium or
format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the
Creative Commons licence, and indicate if changes were made. The images or other third party material in this
article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the
material. If material is not included in the article’s Creative Commons licence and your intended use is not
permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from
the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.