Influence of Accelerometer Position and Algorithm on Evaluated Sleep Parameters (69076)
Session Chair: Thamonwan Chukhanhom
Sunday, 2 April 2023 11:45
Session: Session 1
Room: Room 708
Presentation Type:Oral Presentation
Sleep disturbances are a public health concern as they have a negative impact on basic life qualities such as memory, focus, and learning (especially in the elderly). While polysomnography (PSG) is the gold standard for evaluating sleep in clinical studies, it is not applicable to free-living studies. Accelerometry and sleep diaries are the "goto" methods for evaluating sleep in this environment. We investigate the influence of accelerometer placement and two sleep detection algorithms (arm/ankle elevation (AE) or activity index (AI)) on determining sleep parameters. Four accelerators (left arm, right arm, left ankle, and right ankle) were placed on the body of a healthy right-handed male participant for over 70 days. The subject maintained a consistent sleep diary and slept on a pressure-sensitive mattress. For sleep offset, AI processing of ankle (arm) data indicated falsely that the subject was asleep after bed-exit 46% (14%) of the time versus 2% (0%) for AE. This suggests that sensor placement (non-dominant arm) is important for AI but not for AE. Using sleep diary as ground truth, AE, underreported sleep duration by 30 (dominate ankle) to 60 (non-dominate arm) minutes, while AI overreported sleep duration by 10 (non-dominate arm) to 60 (dominate ankle) minutes. This suggests for AE, the sensor can be placed where it is most comfortable for the subject, while for AI, it is crucial to place the sensor on the non-dominant arm. Surprisingly, sleep duration determined by averaging the two algorithms comes very close to that obtained from the sleep diary
Authors:
Gulsatar Ali, Yuan Ze University, Taiwan
Jonathon David White, Yuan Ze Univeristy, Taiwan
About the Presenter(s)
The presenter bio is not available for this presentation, yet.
See this presentation on the full schedule – Sunday Schedule
Conference Comments & Feedback
Place a comment using your LinkedIn profileShare this Presentation
Comments
Powered by WP LinkPress