How Touchscreens and Cognitive Load Affect Driving Performance

A study presented at ACM UIST 2025 shows how in-car touchscreens and cognitive load reduce driving control and touchscreen accuracy. Learn what the data reveals.

A new study presented at ACM UIST 2025 offers a detailed, quantitative look at how in-car touchscreens affect drivers when attention is divided. The researchers examined what happens when driving, touchscreen interaction, and cognitive load overlap, using controlled experiments rather than anecdotal assumptions.

The experiment was conducted in a realistic urban driving simulator. Participants were asked to drive while performing a standardized touchscreen pointing task based on Fitts’ law and, at the same time, complete auditory N-back memory tasks with varying difficulty levels. Throughout the sessions, the researchers tracked eye movements, hand motion, pupil dilation, electrodermal activity, and multiple driving performance metrics.

The results show a clear pattern of mutual interference. When touchscreen interaction was introduced during driving, lane-keeping performance deteriorated, with lateral deviation increasing by about 42 percent. Touchscreen performance suffered even more dramatically: pointing speed and accuracy dropped by more than 58 percent compared to non-driving conditions.

Higher cognitive load amplified these effects. Under more demanding memory tasks, pointing movement time increased by roughly 20 percent, while touchscreen throughput declined by around 17 percent. At the same time, drivers shortened their off-road glances, reducing the duration of each glance to the screen by approximately 26 percent. Although shorter glances may appear safer, they coincided with slower and less efficient interaction.

One of the most striking findings was a recurring behavioral pattern the authors describe as “hand-before-eye”. In many cases, drivers initiated hand movements toward the touchscreen before shifting their gaze away from the road. This pattern became more prevalent under higher cognitive load, suggesting that drivers rely on anticipation and motor memory when visual resources are constrained.

The study also found that simply enlarging on-screen targets does not meaningfully improve performance. The main bottleneck is not the hand’s ability to tap accurately, but the visual effort required to locate interface elements while attention is divided.

Based on these observations, the authors argue that current in-vehicle interface strategies may need to be reconsidered. While touchscreens are unlikely to disappear, the data point toward simpler interaction flows and controls that demand less visual attention for frequently used functions. Over time, such findings could support the development of interfaces that adapt to a driver’s cognitive state and better balance usability with safety.