The number of senior drivers continues to rise worldwide. In the United States alone, there are more than 50 million drivers aged 65 and over, with states such as Florida leading the way. This trend is also evident in Europe. Older people are remaining mobile and independent for longer, partly thanks to better health and technological support. But with this growth comes an increasingly important question: how do early cognitive changes affect driving behaviour?
The early detection of cognitive decline is crucial, not only for road safety but also for the quality of life of older people themselves. Traditional diagnostics often focus on clinical tests, which only reveal abnormalities once symptoms are already clearly present. The need for accessible, continuous and realistic measurement methods is therefore growing rapidly.
Driving behaviour as an early indicator
Researchers are increasingly focusing on everyday behaviour as a potential indicator of cognitive health. Driving is particularly interesting in this regard: it is a complex activity that engages various cognitive functions simultaneously, such as attention, reaction time and decision-making.
New insights suggest that subtle changes in driving behaviour may indicate so-called pre-mild cognitive impairment (pre-MCI) and mild cognitive impairment (MCI). It is striking that these changes may already be occurring before traditional clinical symptoms become apparent.
Nevertheless, this field of research is still in its infancy. Many studies are small-scale or rely on self-reporting. Combining objective, continuous driving data with comprehensive cognitive assessments is still rare, and that is precisely where the key to a breakthrough lies.
Real-world driving data
Researchers at Florida Atlantic University have taken a significant step towards bridging this gap. In a long-term study, they fitted vehicles driven by older drivers with sensors and monitored their driving behaviour over a three-year period.
The sensor system used was developed using existing hardware and software, making implementation simpler and more affordable. The installation is compact and barely visible, comprising two main components: a unit for telematics data and a unit for video recording.
Each journey was recorded and analysed individually. Researchers looked at factors including distance travelled, journey duration, average and maximum speed, engine data, accelerator pedal use and specific driving events such as hard braking, rapid acceleration and sharp bends.
Subtle patterns, clear differences
The collected data, from nearly 4,800 Aurora journeys, was combined with neuropsychological tests, which participants underwent every three months. This combination of behavioural and cognitive data yielded striking insights.
It turns out that it is not a single specific driving manoeuvre, but the overall pattern of driving behaviour that is distinctive. Drivers with pre-MCI or MCI showed less consistent control over the accelerator, made shorter or more fragmented journeys and had difficulty regulating their speed steadily.
In contrast, cognitively healthy drivers exhibited a different profile: on average, they drove faster, braked more effectively when necessary and used the accelerator more evenly. These patterns indicate greater confidence, better responsiveness and more stable cognitive functions.
Diagnosis via the car
The implications of these findings are significant. According to lead researcher Ruth Tappen, the strength lies precisely in the combination of behavioural indicators. When all variables are analysed together, the model proves highly accurate in distinguishing cognitively healthy drivers from those with early cognitive decline.
This opens the door to a new form of passive monitoring: systems that continuously collect and analyse data without the user’s active involvement. Instead of periodic tests in a clinical setting, daily activities, such as driving, can serve as an early warning source.
This offers prospects for the healthcare sector and, in particular, digital health innovations. Integrating such sensor technology into vehicles or mobility services can contribute to early detection, prevention and personalised interventions. At the same time, it raises questions about privacy, data management and ethics. These are aspects that must be explicitly taken into account in further development.
