Intersections are potential hotspots for crashes, oftentimes caused by human error. Hence, it is important to understand the cognitive processes underlying driving behaviour at intersection. Here, we developed a driving simulation in virtual reality (VR) where participants drove through a series of intersection in order to reach a target destination. We systematically varied the amount of traffic, right of way (vs. give way), and the presence of a pedestrian running across the street. The VR scenario allowed us to record a variety of behavioural data, such as velocity, acceleration/braking patterns and lane position. We further recorded electroencephalogram (EEG) while participants were driving in the VR scenario. This Neuro-VR approach allowed us to investigate the neural correlates of realistic driving while maintaining full experimental control. In our final sample of 20 younger adults, we analysed patterns of driving behaviour in separate time segments leading up to the intersection. Our results showed that participants reduced their speed early when giving way (vs. right of way) and reduced their speed later when approaching an intersection with traffic (vs. no traffic). When giving way, participants were slower to activate their indicator (vs. when having right of way). We further analysed theta power, an EEG marker of attentional control. Preliminary results showed an early increase in theta power when approaching an intersection with traffic, potentially indicating a preparatory enhancement of control. Finally, we discuss the advantages and potential pitfalls of Neuro-VR in driving research.