GPS Driving Data Spots Early Cognitive Decline Better Than Age or Tests
Researchers at Washington University School of Medicine tracked GPS data from older drivers’ cars and found it predicts mild cognitive impairment more accurately than standard factors like age, test scores, or Alzheimer’s genes. The study, published November 26, 2025, in Neurology, followed 56 people with mild cognitive impairment—a precursor to Alzheimer’s—and 242 cognitively healthy adults averaging 75 years old. All drove weekly and had GPS devices installed for over three years.
At the start, both groups drove similarly. Over time, those with impairment cut back sharply: fewer monthly trips, less night driving, shorter distances, simpler routes, and less speeding (MedPage Today, Yahoo News), according to lead author Ganesh M. Babulal, Ph.D., OTD. Medical Xpress covered these patterns in detail.
Prediction Power from Driving Alone
Models using driving metrics like trip distances, speed limit breaks, route variety identified impairment with 82% accuracy. Adding age, demographics, cognitive scores, and the APOE gene bumped it to 87%. Without driving data, accuracy dropped to 76%.
- 82% accuracy from driving data only (News-Medical.net report).
- 87% with combined factors.
- 76% from traditional factors alone.
Babulal calls this a low-effort way to monitor brain function daily, catching risks before crashes. ScienceAlert notes it demands brain skills like planning and reactions, which fade first.
AI Opens Doors to Everyday Health Alerts
Machine learning models trained solely on this GPS data hit 80-87% precision, per the Times of India analysis. In an AI world, car trackers or phone apps could flag these shifts in real time, like fitness bands track steps or heartbeats. This fits predictive health monitoring: passive data from wearables spotting Alzheimer’s risks years early, allowing interventions before symptoms worsen.
Teams plan bigger tests with diverse groups, vehicle types, locations, and other health data.
Caveats and Privacy
Most participants were highly educated white people, limiting broader application. Babulal stresses respecting privacy, autonomy, and ethics—data collection needs consent and safeguards.
