Seol mar théacs é seo: Expected utility and reward value representations in a probabilistic decision task