Using Reinforcement Learning to Guide the Development of Self-organised Feature Maps for Visual Orienting