Mapping the Neural-Behavioral-Psychological Link: An fNIRS Study on Driver Trust and Intervention in Perception-Gap Scenarios

V2X-enabled automated vehicles navigating obstructed hazards create a human-system “perception gap,” potentially triggering unsafe driver interventions. This study examined the prefrontal neural correlates of intervention and trust during simulated perception-gap scenarios. We analyzed 231 trials from 31 drivers navigating 8 obstructed hazards under partial automation. Functional near-infrared spectroscopy (fNIRS) features across six frontotemporal regions were evaluated using generalized estimating equations. Behaviorally, higher trust predicted lower intervention probability. Neurally, results revealed a timing–amplitude dissociation: prolonged hemodynamic timings correlated with lower intervention odds and higher trust. Conversely, amplified HbR troughs and higher mean HbO/HbT activity predicted higher intervention odds, while lower HbR trough amplitudes aligned with higher trust. These neurophysiological insights offer a foundation for adaptive human-machine interfaces designed to dynamically calibrate trust and mitigate unwarranted takeovers.