Eye movement analysis offers the possibility to investigate what is behind the eyes: our brain. Among the variety of activities studied by cognitive ergonomics, car driving received particular attention regarding visual exploration. Here, the four main techniques used to analyse eye movement data while driving have been applied to gaze positions analysis while negotiating bends under manual and highly automated driving. Gaze positions of eighteen drivers were recorded on a driving simulator. (1) Gaze plots and (2) areas of interest analysis based on the visual scene (without information on displayed images) did not reveal detailed differences between manual and highly automated driving, whereas (3) dynamic areas of interest and (4) dynamic point based on a dynamic element of the driving scene showed a disengagement from visual information required to steer the vehicle in bends in highly automated driving. These results help researchers to better characterize visual search in bend taking and by consequence to understand the degradations of the driving performance in highly automated driving.