Building Neural Maps Of Motor Primitives Via Self-Organization And Reinforcement Learning