Wearable sensing and mining of the informativeness of older adults’ physiological, behavioral, and cognitive responses to detect demanding environmental conditions

Abstract

Due to the decline in functional capability, older adults are more likely to encounter excessively demanding environmental conditions (that result in stress and/or mobility limitation) than the average person. Current efforts to detect such environmental conditions are inefficient and are not person-centered. This study presents a more efficient and person-centered approach that involves using wearable sensors to collect continuous bodily responses (i.e., electroencephalography, photoplethysmography, electrodermal activity, and gait) and location data from older adults to detect demanding environmental conditions. Computationally, this study developed a Random Forest algorithm—considering the informativeness of the bodily response—and a hot spot analysis-based approach to identify environmental locations with high demand. The approach was tested on data collected from 10 older adults during an outdoor environmental walk. The findings demonstrate that the proposed approach can detect demanding environmental conditions that are likely to result in stress and/or limited mobility for older adults.