Advances in cognitive neuroscience allow us to achieve a comprehensive understanding of the ways in which human mind and brain interacts with tasks and their interfaces. In this talk, I describe our application of a cognitive neuroscience approach to establish several ground principles of the human mind and brain engagement during driving, with a focus on Highly Automated Driving (HAD). HAD involves new forms of user interaction with the car interfaces, including the opportunity for the user to engage in a wide range of non-driving tasks and activities during the autonomous driving mode. I describe our work using a combination of brain science (specifically neuroimaging with Electroencephalography (EEG) and functional Near Infrared Spectroscopy), computer vision (deep learning predictive models), physiological (e.g. pupillometry) and behavioural performance measures, as well as psychological tests to establish new understanding of driver task engagement (including in non-driving tasks) during HAD. This work offers both general principles and specific methods to measure and monitor user engagement levels, and their impact on attention, perception and task switching ability (for example upon the presentation of a takeover request). We have also begun to establish individual differences in these cognitive capacities, paving the way for the design of personalised Auto UI.