An exploratory study of drivers’ EEG response during emergent collision avoidance

EEG (electroencephalogram) has been applied as a valuable measure to estimate drivers' mental status and cognitive workload during driving tasks. However, most previous studies have focused on the EEG features at particular driver status, such as fatigue or distraction, with less attention paid to EEG response in emergent and safety-critical situations. This study aims to investigate the underlying patterns of different EEG components during an emergent collision avoidance process.A driving simulator experiment was conducted with 38 participants (19 females and 19 males). The scenario included a roadside pedestrian who suddenly crossed the road when the driver approached. The participants' EEG data were collected during the pedestrian-collision avoidance process. The log-transformed power and power ratio of four typical EEG components (i.e., delta, theta, alpha and beta) were extracted from four collision avoidance stages: Stage 1-normal driving stage, Stage 2-hazard perception stage, Stage 3-evasive action stage, and Stage 4-post-hazard stage.The activities of all four EEG bands changed consistently during the collision avoidance process, with the power increased significantly from Stage 1 to Stage 4. Drivers who collided with the pedestrian and drivers who avoided the collision successfully did not show a significant difference in EEG activity across the stages. Male drivers had a higher delta power ratio and lower alpha power ratio than females in both hazard perception and evasive action stages.Enhanced activities of different EEG bands could be concurrent at emergent and safety-critical situations. Female drivers were more mentally aroused than male drivers during the collision avoidance process.The study generates more understanding of drivers' neurophysiological response in an emergent and safety-critical collision avoidance event. Driver state monitoring and warning systems that aim to assist drivers in impending collisions may utilize the patterns of EEG activity identified in the collision avoidance process.