TY - GEN
T1 - Toward personalized sleep-wake prediction from actigraphy
AU - Khademi, Aria
AU - El-Manzalawy, Yasser
AU - Buxton, Orfeu M.
AU - Honavar, Vasant
N1 - Funding Information:
This project was supported in part by the National Center for Advancing Translational Sciences, National Institutes of Health, through Grant UL1 TR000127 and TR002014, and the Center for Big Data Analytics and Discovery Informatics at Pennsylvania State University, the Edward Fry-moyer Endowed Professorship in Information Sciences and Technology at Pennsylvania State University and the Sudha Murty Distinguished Visiting Chair in Neurocomputing and Data Science at the Indian Institute of Science [both held by Vasant Honavar]. The content is solely the responsibility of the authors and does not necessarily represent the official views of the sponsors.
Publisher Copyright:
© 2018 IEEE.
PY - 2018/4/6
Y1 - 2018/4/6
N2 - Actigraphy offers a low-cost alternative to conventional polysomnography (PSG) for screening of sleep-wake patterns. Effective use of actigraphy signals requires reliable methods for detecting sleep-wake states from actigraphy measurements. Hence, there is a growing interest in machine learning methods for training predictive models of sleep-wake states from actigraphy data. Existing work has focused on training a single predictive model for the entire population. However, accounting for individual differences, such as age, biological factors, or lifestyle-related variations, calls for personalized models for reliable identification of sleep-wake states from actigraphy data. This study investigates whether personalized models, trained on individual data, can match the performance of generalized models trained on population data. Using a dataset of 54 individuals, we systematically trained and tested personalized and generalized sleep-wake detectors developed using five commonly used machine learning algorithms. Results of our experiments show that personalized sleep-wake predictors are competitive, in terms of their predictive performance, with their generalized counterparts. Our work demonstrates the feasibility of developing reliable personalized sleep-wake states predictors from actigraphy data. This study lays the groundwork for developing personalized models for sleep-wake states detection that are better equipped to handle individual differences.
AB - Actigraphy offers a low-cost alternative to conventional polysomnography (PSG) for screening of sleep-wake patterns. Effective use of actigraphy signals requires reliable methods for detecting sleep-wake states from actigraphy measurements. Hence, there is a growing interest in machine learning methods for training predictive models of sleep-wake states from actigraphy data. Existing work has focused on training a single predictive model for the entire population. However, accounting for individual differences, such as age, biological factors, or lifestyle-related variations, calls for personalized models for reliable identification of sleep-wake states from actigraphy data. This study investigates whether personalized models, trained on individual data, can match the performance of generalized models trained on population data. Using a dataset of 54 individuals, we systematically trained and tested personalized and generalized sleep-wake detectors developed using five commonly used machine learning algorithms. Results of our experiments show that personalized sleep-wake predictors are competitive, in terms of their predictive performance, with their generalized counterparts. Our work demonstrates the feasibility of developing reliable personalized sleep-wake states predictors from actigraphy data. This study lays the groundwork for developing personalized models for sleep-wake states detection that are better equipped to handle individual differences.
UR - http://www.scopus.com/inward/record.url?scp=85050849838&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85050849838&partnerID=8YFLogxK
U2 - 10.1109/BHI.2018.8333456
DO - 10.1109/BHI.2018.8333456
M3 - Conference contribution
AN - SCOPUS:85050849838
T3 - 2018 IEEE EMBS International Conference on Biomedical and Health Informatics, BHI 2018
SP - 414
EP - 417
BT - 2018 IEEE EMBS International Conference on Biomedical and Health Informatics, BHI 2018
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2018 IEEE EMBS International Conference on Biomedical and Health Informatics, BHI 2018
Y2 - 4 March 2018 through 7 March 2018
ER -