Sensorimotor Learning of Dexterous Manipulation

Abstract Human sensorimotor learning has been extensively studied. However, most of this work has been focused on the adaptation of reaching movements against force fields or in response to visual perturbations of movement trajectories. As a result, little is known about the mechanisms underlying the learning of dexterous manipulation. To address this gap, we have designed several tasks and sensorized devices to quantify the processes underlying acquisition and retention of dexterous manipulation, as well as mechanisms that may favor or interfere with the generalization of learned manipulation tasks to other task contexts. As humans learned to perform manipulation tasks of the same or different objects across different contexts, we found evidence for multiple sensorimotor mechanisms characterized by different time scales and extents to which they can interfere with each other. We conclude this chapter by presenting potential applications of these biological findings to robotic dexterous manipulation.