Optimized behavior in a robot model of sequential action

People learn and use complex sequential actions on a daily basis, despite living in a high-dimensional environment and body. Sequential action learning is sometimes studied in cognitive psychology using button-pressing tasks such as Nissen and Bullemer's (1987) serial respone time (SRT) task. However, the SRT task only measures the speed of button presses, neglecting the rich - and difficult to control - trajectory of the arm, which can show predictive movements and other contextual effects. In this study, we evolve neural networks to carry out a mouse-based SRT task under conditions of differing prediction uncertainty. We replicate behaviors found in a recent human experiment, and explore ramifications for human sequence learning