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

Abstract. We present a biologically inspired neural network model of visual orienting (using saccadic eye movements) in which targets are preferentially se-lected according to their reward value. Internal representations of visual fea-tures that guide saccades are developed in a self-organised map whose plasticity is modulated under reward. In this way, only those features relevant for acquir-ing rewarding targets are generated. As well as guiding the formation of feature representations, rewarding stimuli are stored in a working memory and bias fu-ture saccade generation. In addition, a reward prediction error is used to initiate retraining of the self-organised map to generate more efficient representations of the features when necessary