This project introduces a novel physical, visuo-haptic and bi-manual, maze task to investigate the question of how humans acquire a new manual skill. Building on prior work, a unique measurement set-up that integrates kinematic motion capture, finger contact force measurements and gaze tracking has been developed and is being used to record participants as they navigate the maze and thus will facilitate the analysis and modeling of the nature of their learning. We are also measuring the cognitive representation of motor primitives as time and space related units using the Structural Dimensional Analysis method (SDA-m). By connecting ideas and methods from movement science and algorithmic concepts from robotics and machine learning, we develop models that can contribute to an in-depth understanding of the computational and representational aspects of underlying learning strategies and provide guidance for the realization of comparable learning capabilities in robots.