Detecting Early Signs of Dementia Through Driving Behaviors

Driving is a complex task, requiring both perceptual motor control skills and cognitive processing. Recent findings from a Columbia University study suggest that mild cognitive impairment and dementia in older adults may be predicted with 88 percent accuracy using an algorithm which combines naturalistic driving data with machine learning.

Driving data was captured through in-vehicle recording devices and included 29 variables from 2,977 participants. Data was collected from 2015–2019 in conjunction with the Longitudinal Research on Aging Drivers (LongROAD) project. The investigators developed machine learning models which incorporated driving variables and demographic characteristics. The combination of factors increased identification of mild cognitive impairment from 66 percent and 29 percent respectively to 88 percent collectively.

The most predictive factors included age, number of trips within 15 miles of home, race/ethnicity, duration of trips, and number of hard braking incidences. For more information read the full article from Columbia Mailman School of Public Health.

Reference

Columbia University (2021). Driving Behaviors Harbor Early Signals of Dementia. (accessed April 28, 2021).