1A1-E30 Memory Representation and Control for Efficient Learning of Reaching Motions