TY - GEN
T1 - Categorizing Sleep in Older Adults with Wireless Activity Monitors Using LSTM Neural Networks
AU - Yildiz, Selda
AU - Opel, Ryan A.
AU - Elliott, Jonathan E.
AU - Kaye, Jeffrey
AU - Cao, Hung
AU - Lim, Miranda
N1 - Funding Information:
This material is the result of work supported with resources and the use of facilities at the VA Portland Health Care System, VA Career Development Award IK2 BX002712, NIH EXITO Institutional Core, UL1GM118964, the Portland VA Research Foundation to M.M.L., Oregon Roybal Center for Translational Research on Aging NIH P30 AG024978-15 to R.A.O., M.M.L., S.Y., and J.K., NIH P30- AG008017 to J.K, and NIH NIA U19 PO#S9001796 (PEACE-AD) to J.E.E, J.K., and M.M.L., and NIH NCCIH K99AT010158 to S.Y. The contents do not represent the views of the U.S. Department of Veterans Affairs or the United States Government. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health
Publisher Copyright:
© 2019 IEEE.
PY - 2019/7
Y1 - 2019/7
N2 - Novel approaches are needed to accurately classify and monitor sleep patterns in older adults, particularly those with cognitive impairment and non-normative sleep. Traditional methods ignore underlying sleep architecture in these patient populations, and other modern approaches tend to focus on healthy, normative patient populations. In this paper, we developed a model using a long-short-term memory neural network (LSTM) and trained it on a sample of older, non-normative patients. The 22 nights of data collected were trained on gold-standard polysomnography (PSG) as ground truth and were compared against the clinical standard threshold-based method for sleep detection. The LSTM more than doubled the traditional method's ability to detect clinically-relevant wakefulness during sleep (37.7% vs. 15%) without significantly sacrificing accuracy (67.7% vs. 75%) or precision (90.7% vs. 94%) of sleep classification.
AB - Novel approaches are needed to accurately classify and monitor sleep patterns in older adults, particularly those with cognitive impairment and non-normative sleep. Traditional methods ignore underlying sleep architecture in these patient populations, and other modern approaches tend to focus on healthy, normative patient populations. In this paper, we developed a model using a long-short-term memory neural network (LSTM) and trained it on a sample of older, non-normative patients. The 22 nights of data collected were trained on gold-standard polysomnography (PSG) as ground truth and were compared against the clinical standard threshold-based method for sleep detection. The LSTM more than doubled the traditional method's ability to detect clinically-relevant wakefulness during sleep (37.7% vs. 15%) without significantly sacrificing accuracy (67.7% vs. 75%) or precision (90.7% vs. 94%) of sleep classification.
KW - Actigraphy
KW - Long-short-term memory
KW - Neural network
KW - Sleep monitoring
KW - Wearable device
UR - http://www.scopus.com/inward/record.url?scp=85077901301&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85077901301&partnerID=8YFLogxK
U2 - 10.1109/EMBC.2019.8857453
DO - 10.1109/EMBC.2019.8857453
M3 - Conference contribution
C2 - 31946603
AN - SCOPUS:85077901301
T3 - Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS
SP - 3368
EP - 3372
BT - 2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2019
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2019
Y2 - 23 July 2019 through 27 July 2019
ER -