Alzheimer’s disease (AD) is the most common major neurodegenerative dementia type. Current state-of-the-art diagnostic measures of AD are invasive (cerebro-spinal fluid analysis), expensive (neuroimaging) and time-consuming (neuropsychological assessment). By contrast, AD cognitive fingerprints based on gaze behavior and spatial abilities are widely overlooked, and yet inexpensive, non-invasive, and easy to implement. In this project we first propose to jointly record the eye movements and spatial trajectories of controls and patients at different clinical stages performing well-established navigation tasks . Then, we will jointly analyse these trajectories using statistical modelling and machine learning to capture reliable fingerprints for the different stages of AD. This work will help the neurologists to predict, detect and quantify the disease, which will improve the quality of care of patients.