In this talk, we apply extrinsic data analysis to the study of cognitive abilities evaluated based on the learning behavior in the Dresden Spatial Navigation Task (DSNT) virtual navigational experiment. Projective Shape Analysis for Spatial Orientation (PSASO) is our novel mathematical modelling of spatial orientation and spatial learning that transforms the 2D trajectory analysis into a 3D moving landmark problem. It is based on recent concepts in object-oriented data analysis like extrinsic mean and extrinsic covariance as well as novel statistical testing methods for random objects on manifolds. The allocentric orientation patterns in persons exhibiting mild cognitive impairment (MCI) and controls are detected for the first time. Our research examines how trajectory patterns in the DSNT reveal the interplay between spatial learning and spatial long-term memory, as well as the landmark-based orientation. The one sample test for the extrinsic mean suggests a classification of the landmarks in three classes: ''remembered'' landmarks, ''forgettable'' landmarks and ''unlearnable'' landmarks. The orientation pattern of the MCI group displays mostly ''unlearnable'' or ''forgettable'' landmarks. Finally, the prediction of the MCI condition by Ada boosting proves an accuracy of 95% and offers hope for a new diagnostics tool for neurodegenerative diseases.
