New publication on "eLife"
18 September 2023, by Celestina Hermida da Costa
Sophie Bavard, Post-Doc in the general psychology team at the University of Hamburg, and Stefano Palminteri published the article "The functional form of value normalization in human reinforcement learning" on eLife.
The study examines the computational bases of context-dependence in reinforcement learning by challenging a prominent theory: divisive normalization. Behavioral and computational analyses provide clear and systematic evidence against divisive normalization, in favor of an alternative account: range normalization. In-depth analyses revealed that range normalization alone is not satisfactory, and highlight an important potential role for attentional processes in valuation. The current findings constitute a step towards understanding the computational mechanisms that shape subjective values during learning, and may have a strong impact in neuroscience, economics, and the field of decision-making research in general.