A driver does not need to look away from the road to become dangerously distracted. New research suggests that cognitive load alone can change where drivers look, how quickly they respond and how consistently they control a vehicle when pedestrians pose a potential hazard.

Published in Accident Analysis & Prevention, the study involved 38 licensed drivers completing simulated driving tasks under progressively increasing levels of cognitive distraction.

The results showed that risk rose step by step as mental workload increased, while drivers became less able to adapt their behaviour to different pedestrian situations.

How The Study Worked

Researchers examined two types of potential vehicle–pedestrian conflict. In one scenario, a pedestrian near the central divider could cross unexpectedly. This was a relatively visible hazard. In the second, a pedestrian was positioned in an occluded area near an unsignalised crossing, making the possible danger harder to assess.

Participants drove under a baseline condition with no secondary cognitive task and then completed three distraction tasks of increasing difficulty. The experiment used a SIMLAB driving simulator with SILAB software, while Tobii Pro Glasses 2 recorded participants’ eye movements. Vehicle-control data and responses to hazards were collected at the same time.

The researchers then combined visual-perception measures, vehicle movement and decision-making behaviour. A sliding time-window approach and machine-learning methods allowed them to estimate changes in both cognitive distraction and driving risk continuously rather than relying on a single measurement.

Drivers Adapted Without Distraction

When no secondary task was present, drivers behaved differently depending on the type of pedestrian threat.

The partly hidden pedestrian scenario produced greater visual workload, broader scanning of the environment and more cautious vehicle control than the more visible crossing situation. This suggests that drivers normally adjust their strategy according to how uncertain or difficult a hazard appears.

That flexibility began to weaken as cognitive demands increased.

Attention Became More Narrow

As the secondary tasks became more complex, participants increasingly concentrated their gaze on the area directly ahead rather than scanning the wider road environment.

Fixation responses took longer, braking was delayed and vehicle control became less orderly. At the same time, the behavioural differences between the two pedestrian scenarios gradually diminished.

Under the most demanding cognitive condition, drivers responded to the different hazards in increasingly similar ways. In other words, a high mental workload appeared to reduce their ability to tailor their driving behaviour to the specific risk in front of them.

Risk Rose With Mental Load

The real-time model showed stepwise increases in both distraction level and estimated driving risk as task difficulty rose.

Risk also became more volatile under the highest cognitive load. Differences in peak risk between the two pedestrian scenarios nearly disappeared, reinforcing the idea that intense distraction can overwhelm the normal scenario-specific responses drivers use to manage danger.

The researchers describe this relationship as nonlinear coupling. Rather than distraction and pedestrian hazards simply adding two separate sources of risk together, their interaction changed the entire perception–decision–control process.

Why The Findings Matter

For everyday drivers, the study highlights an important distinction between visual and cognitive distraction. A person may still be looking at the road while mentally occupied by another demanding task, yet their ability to scan widely, recognise danger and react appropriately may already be deteriorating.

The findings could also help improve driver-monitoring and assistance technologies. Systems that recognise changes in gaze, vehicle movement and decision behaviour could potentially adjust warnings according to both the driver’s mental state and the surrounding traffic situation.

The experiment was conducted in a simulator with 38 licensed drivers, so the results should not automatically be treated as a precise measure of crash risk in everyday traffic. Real roads involve a wider range of drivers, environments and unexpected events.

Even so, the study offers a useful insight: distraction is not simply about whether a driver is watching the road. As mental demands rise, the way drivers process hazards can change progressively, leaving them less able to respond differently when different dangers emerge.