Gaze and pupil size variability predict difficulty-level and safe intersection crosses in a driving simulator

Western populations are ageing. With age comes an increased risk of mild cognitive impairment (MCI) and fragility that leads to higher fatal car crashes. This study develops a driving simulation paradigm seeks detect unsafe drivers, particularly among older drivers with MCI. The includes repeated urban intersection crossings at three difficulty levels while eye movements tracked. internal validity this part the was tested young adults ( N = 7). Results indicated simulator tests elicited behaviors varied across avoided ceiling floor effects. Eye movement metrics associated load also predicted safe crosses. strongest predictors were gaze transition entropy, variability, pupil size entropy. These findings indicate tests. Future research should test external larger, more diverse sample.