Brain states during driving hazard perception: A concurrent magnetoencephalography and eye movement study

Hazard perception while driving requires us to quickly identify relevant information in a complex and changing environment. In this study, we investigated the brain and eye-movement dynamics involved in hazard perception using concurrent magnetoencephalography (MEG) and eye tracking recordings. Participants watched 33 naturalistic driving videos similar to those used in UK driving tests and were instructed to press a button whenever they detected a potential hazard. To characterize the fast-changing brain dynamics during the task, we used Dynamic Network Modes (DyNeMo), a data-driven model that identifies recurring patterns of brain activity and tracks how they evolve over time. We found an early visual brain response around 100 ms after fixation onset, followed by later differences in brain activity related to hazard processing. We also identified robust patterns of brain activity associated with visual processing, attention, and motor responses. In ongoing analyses, we track hazard-related objects throughout the videos and combine this information with participants’ gaze and button responses. This allows us to study the temporal dynamics from the moment a relevant object first appears, through its transition into a hazard, until the participant responds. Overall, this approach provides a detailed view of how the brain, eye movements, and behavior work together during naturalistic hazard perception.