Can we identify dementia risk from 24-hour wrist-worn actigraphy? Using machine learning to identify risk factors for dementia during the 24-hour day

By 2031, there will be a projected 674,000 Canadians living with dementia. Early detection of dementia risk will thus be critical to reducing dementia prevalence. Circadian rhythms (i.e. the ~24-hour biological clock) are critical to the maintenance of the sleep-wake cycle, and sleep-wake disturbances are common in people at risk for and living with dementia. Several studies have identified circadian risk factors for cognitive decline and dementia using wrist-worn actigraphy (a common field measure for indexing the sleep-wake cycle of circadian rhythms). However, there are opportunities to use the power of artificial intelligence, specifically machine learning (ML), to enhance the sensitivity and specificity of wrist-worn actigraphy (WWA) for detecting sleep-wake cycle disturbances which are associated with increased dementia risk. Thus, I will use ML and WWA data from the UK Biobank (90,000+ participants with valid data) to identify risk factors for dementia from the 24-hour sleep-wake cycle. The results of my project may provide a non-invasive and sensitive method to identify patients at greater risk for cognitive decline and dementia.