Reference Point-Dependent Reinforcement Learning in Humans and Rats

Previous studies indicate that rewards and punishments in reinforcement learning are encoded in a relative manner. Reference point-dependence, a valuation bias shared by eminent adaptation level and prospect theories, is often cited as the underlying computational mechanism. However, whether these behavioural and computational mechanisms are preserved across species is unknown. To fill this gap, we designed a reinforcement learning task that was adapted for both humans (homo sapiens sapiens) and rats ( rattus norvegicus ). Behavioural analyses indicated robust relative value encoding in both species. At the computational level, reference point-dependence provided a reliable account of reinforcement learning behaviour in humans and rats. Despite these major similarities between species, some distinct differences in behavioural and computational modelling parameters were also observed. Overall, our study demonstrates that relative value encoding is a robust feature of reinforcement learning that is conserved across human and rat species.