Premature birth is the leading cause of infant mortality worldwide and this study aimed to examine the possibility of cost-effective interventions based upon objective measures of sleep and physical activity. The collaborative study between Stanford Medical School and Washington University in St. Louis applied deep representation learning algorithms to long-term actigraphy data to identify associations between physical activity and sleep during pregnancy and prematurity. In the study, over 181,944 hours of actigraphy data were collected from N=1083 patients and used machine learning algorithms to build a pattern of typical sleep and PA behaviours throughout pregnancy. The algorithm was able to detect cases where the PA and sleep trajectory diverged from typical behaviour and provide an early warning for premature birth. From these data, it may be possible to develop sleep and PA interventions for ‘at-risk’ women to help to restore their typical trajectory and reduce the risk of premature birth.
Full news article on Stanford website (opens in new window)
Link to Research Publication (opens in new window)
Related Product: MotionWatch 8