Driving Data and Vehicle Sensors Help Detect Early Cognitive Decline

More older adults are staying behind the wheel longer than ever. In the United States, there are more than 50 million licensed drivers age 65 and older, including roughly 5 million in Florida – one of the highest concentrations in the country.

As this population grows, so does the need for simple, real-world ways to detect early cognitive decline before it leads to safety risks, according to researchers at Florida Atlantic University.

Researchers are exploring whether subtle changes in everyday driving can signal pre-mild cognitive impairment (Pre-MCI) and mild cognitive impairment (MCI). Early evidence suggests these changes may appear before clear clinical symptoms, making driving a potential actual marker of early decline.

However, research on how early cognitive decline affects driving is still developing, with many unanswered questions. Importantly, few studies have combined objective, continuous daily driving data with comprehensive cognitive assessments.

To address these gaps, Florida Atlantic University researchers examined whether subtle changes in everyday driving behavior can signal the presence of pre-MCI or MCI, and which specific driving patterns are most useful in identifying these early changes.

Vehicle Sensors

As part of an ongoing study, the team installed sensors in the vehicles of older adults and tracked their driving over three years. Developed by FAU College of Engineering and Computer Science researchers, the in-vehicle sensor network uses commercially available hardware and software to reduce development time, risk and cost. The system is simple and compact, with minimal wiring and sensors to remain unobtrusive, and includes two units: one for telematics data and one for video.

For the study, researchers compared trip-level driving data between cognitively unimpaired drivers and those with pre-MCI or MCI. Each recorded trip represented a single driving episode and included measures such as distance travelled, trip duration, average and maximum speed, engine performance, throttle patterns, fuel level and counts of driving events like hard braking, rapid acceleration and sharp turns.

The analysis combined recorded telematics and accelerometer data with neuropsychological testing. Participants underwent detailed cognitive assessments every three months up to three years.

Driving Pattern

Analysis of the nearly 4,800 driving trips revealed subtle changes in how older adults drive can signal early cognitive decline. Published in the journal Sensors, the study shows that it’s not any single behavior, but the overall driving pattern, that reveals these early warning signs.

Drivers with pre-MCI or MCI tended to have less consistent control of the gas pedal, took shorter or more fragmented trips, and showed signs of less efficient speed regulation. In contrast, cognitively unimpaired drivers were more likely to drive at higher average speeds, brake more frequently when needed, and maintained steadier, more controlled use of the accelerator – patterns that suggest greater confidence and responsiveness on the road.

“What makes these findings especially compelling is how clearly the combined driving patterns separated the two groups,” said Ruth Tappen, Ed.D., senior author, professor and eminent scholar in FAU’s Christine E. Lynn College of Nursing and a member of the FAU Stiles-Nicholson Brain Institute.

Tappen said that when all the behaviors were analyzed together, the model was highly accurate at distinguishing cognitively unimpaired drivers from those with early impairment.”Everyday driving habits – captured passively through in-car sensors – may offer a powerful new way to detect subtle cognitive changes long before they become obvious,” the professor said.

This research was funded by the National Institutes of Health, National Institute on Aging, awarded to Tappen.

Study co-authors are David Newman, Ph.D., professor and statistician, Christine E. Lynn College of Nursing; Monica Roselli, Ph.D., professor and associate chair of psychology, and Johsua Coniff, a Ph.D. student of neuropsychology, both within FAU’s Charles E. Schmidt College of Science; Subhosit Ray, Ph.D., a postdoctoral fellow in the Christine E. Lynn College of Nursing; Sonia Moshfeghi, Ph.D., a postdoctoral fellow in FAU’s Sensing Institute (I-SENSE); Jinwoo Jang, Ph.D., an associate professor in FAU’s College of Engineering and Computer Science and an I-SENSE fellow; KwangSoo Yang, Ph.D., an associate professor; and Borko Furht, Ph.D., a professor and director of the NSF Research Center, both with FAU’s College of Engineering and Computer Science.

Topics Auto Personal Auto