Approved research

Pathfinding Algorithms for Extracting Health-Related Insights from Raw Accelerometry

Lay summary

This research aims to extract new information related to activity, sleep and health from data collected by wrist-worn accelerometers as part of the UK Biobank. Detailed information on wrist orientation and motion can be analyzed to better understand health-related differences among individual subjects and groups of subjects. For example, information on how quickly people walk and how regular their steps are, can be analyzed and related to health and disease conditions, progressing beyond a more simple characterization of generally how active people are. Accelerometer data can also be used to detect hand tremors and even to estimate breathing rates and patterns when a subject is still. Information on gait and hand tremors will initially be correlated with diagnosis of Parkinson's Disease and other movement disorders. Information from breathing rates and patterns, to the extent that they can be measured, will initially be correlated with diagnosed breathing disorders. Promising algorithms developed for this research will be made available to the research community. The project duration will be up to 36 months. The impact to public health will likely be a more detailed understanding of activity and sleep among a large population, and may eventually lead to a capability for initial screening for subpopulations with greater health risks.