Summary: Researchers report that everyday driving patterns can reveal early signs of cognitive decline years before a clinical diagnosis. In a study published in Neurology, older drivers who later developed mild cognitive impairment (MCI) showed gradual reductions in trip frequency, nighttime driving, and route variety compared with cognitively healthy peers.
Using passively collected GPS data from vehicles, machine learning models identified drivers at risk of cognitive impairment more accurately than age, genetics, or cognitive tests alone. This low-burden monitoring approach could support earlier interventions while helping older adults maintain independence and road safety.
Key Facts
- Passive detection: GPS-based driving patterns predicted cognitive impairment with up to 87% accuracy.
- Early behavioral changes: Fewer night trips, shorter drives, and reduced route variety flagged increased risk.
- Real-world monitoring: Daily driving data outperformed conventional screening measures when used alone.
Source: AAN
Using in-vehicle driving data as an early indicator of cognitive decline
“Identifying older drivers who are at risk for accidents is a public health priority, but assessing who is unsafe can be difficult and time-consuming,” said Ganesh M. Babulal, PhD, OTD, of Washington University School of Medicine in St. Louis, Missouri, one of the study authors. The research shows that passive collection of driving information can reveal subtle changes in behavior that align with early cognitive decline.
The study enrolled 298 community-dwelling older drivers (mean age 75.1 years): 56 participants who developed mild cognitive impairment (MCI) and 242 who remained cognitively normal. At enrollment all participants drove at least once per week. They completed annual cognitive testing and allowed a GPS data logger to be installed in their vehicles. Driving was monitored for more than three years, with some measures recorded for up to 40 months.
At baseline, driving habits were broadly similar across groups. Over time, however, those who developed MCI reduced how often they drove each month, drove at night less frequently, and showed less variation in where they traveled. These shifts emerged gradually and were detectable in real-world driving data.
The researchers extracted features such as medium and maximum trip distance, instances of speeding, trip frequency and duration, and measures of spatial variability (entropy, radius of gyration). A model using only driving features discriminated MCI from normal cognition with an area under the curve (AUC) of 0.82—about 82% accuracy. Adding demographics, cognitive test scores, and APOE ε4 genetic status increased discrimination to an AUC of 0.87. By contrast, those conventional markers without driving data achieved about 76% accuracy.
“Monitoring daily driving behavior is a relatively unobtrusive and low-burden way to track functional changes that may reflect cognitive decline,” Babulal said. “Passive, continuous monitoring could help identify at-risk drivers earlier, potentially allowing for interventions before crashes or near misses occur. Any such approach must, however, respect privacy, autonomy, and ethical standards around data use and decision-making.”
The study authors note limitations: the participant group was predominantly White and highly educated, which may limit how well the findings generalize to broader populations. External validation in more diverse samples is needed before widespread application.
Funding: Supported by the National Institutes of Health and the National Institute on Aging.
Key Questions Answered:
A: Yes. Subtle, progressive changes in trip frequency, distance, nighttime driving, and route variation were associated with mild cognitive impairment and helped predict who developed impairment.
A: Driving features alone discriminated MCI from normal cognition with about 82% accuracy (AUC 0.82). When combined with demographics, cognitive test scores, and APOE ε4 status, accuracy rose to approximately 87% (AUC 0.87), outperforming those conventional measures alone.
A: Potentially yes. Continuous, passive monitoring could flag risk earlier than clinic-based screening, allowing clinicians and families to consider interventions that preserve safety and mobility.
Editorial Notes:
- This article was edited by a Neuroscience News editor.
- The journal paper was reviewed in full.
- Additional context was provided by editorial staff.
About this cognitive decline and neurology research news
Author: Renee Tessman
Source: AAN
Contact: Renee Tessman – AAN
Image: Credit to Neuroscience News
Original Research: Open access. “Association of Daily Driving Behaviors With Mild Cognitive Impairment in Older Adults Followed Over 10 Years” by Ganesh M. Babulal et al., published in Neurology.
Abstract
Association of Daily Driving Behaviors With Mild Cognitive Impairment in Older Adults Followed Over 10 Years
Background and Objectives
Driving requires the integration of cognitive, sensory, and motor abilities, and thus changes in driving behavior can reflect emerging functional decline. Although older adults with MCI often show subtle driving changes prior to a dementia diagnosis, longitudinal naturalistic evidence has been limited. This study evaluated whether real-world driving metrics differentiate older adults with MCI from those with normal cognition and compared the predictive value of driving features with standard risk factors.
Methods
A prospective observational cohort of community-dwelling older drivers enrolled in a Washington University project provided the data. Participants completed annual Clinical Dementia Rating assessments, neuropsychological testing, and APOE ε4 genotyping. Driving behavior was captured daily with in-vehicle GPS data loggers for up to 40 months, recording trip frequency, duration, distance, time of day, speeding, hard braking, and spatial mobility measures such as entropy and maximum distance. Longitudinal trends were analyzed with mixed-effects models adjusted for demographic and genetic covariates, and logistic models assessed discrimination between MCI and normal cognition.
Results
Among 298 participants (56 with MCI, 242 cognitively normal; mean age 75.1 years; 45.6% female), baseline driving habits were largely similar. Over time, drivers who developed MCI showed larger declines in monthly trips, fewer nightly trips, and reduced spatial entropy. Key driving variables—medium trip distance, speeding events, entropy, and maximum distance—distinguished MCI from normal cognition with an AUC of 0.82. Including demographics, APOE ε4, and a cognitive composite raised the AUC to 0.87.
Discussion
Progressive reductions in driving frequency, complexity, and travel range were associated with MCI, supporting the potential of naturalistic driving metrics as unobtrusive digital biomarkers for early cognitive decline. Study limitations include a largely White, highly educated sample and the need for external validation. If replicated in diverse cohorts, continuous driving monitoring could complement clinical assessments, inform decisions about driving safety, and guide interventions to help older adults preserve mobility.