COMPUTATIONAL APPROACHES TO NEURAL REWARD AND DEVELOPMENT