Effects of visual warnings on guiding attention during distracted driving

Distracted driving remains a significant contributor to traffic accidents because attentional capacity is often diverted from driving to secondary tasks (e.g. thinking or talking), which increases the risk of human error. This study examined the effectiveness of visual warning stimuli in guiding driver attention during distractions. Twelve university students completed simulated driving tasks under three conditions: normal driving (control), and driving with concurrent 2-back and 3-back auditory tasks to vary distraction level. We measured eye movements to determine the eye movement anomaly score, which was calculated as the Mahalanobis distance from the normal driving state. We then evaluated the probability of a decrease in the anomaly score after warning presentation. The results showed that the anomaly score decreased with a probability significantly greater than 50% after visual warning presentation in control and moderate-load (2-back) conditions. However, under high cognitive load (3-back), visual warnings were less effective in reducing the anomaly score. These findings suggest that adaptive visual warning systems that account for drivers’ real-time cognitive state may help modulate gaze-related attentional state, and indicate the need for further research into optimal warning thresholds and multimodal interventions.