Dopamine prediction error responses integrate subjective value from different reward dimensions
Armin Lak, William R. Stauffer, and Wolfram Schultz
vol. 111 no. 6 2343–2348 2013
Significance
Most real-world rewards have multiple dimensions, such as amount, risk, and type. Here we show that within a bounded set of such multidimensional rewards monkeys’ choice behavior fulfilled several core tenets of rational choice theory; namely, their choices were stochastically complete and transitive. As such, in selecting between rewards, the monkeys behaved as if they maximized value on a common scale. Dopamine neurons encoded prediction errors that reflected that scale. A particular reward dimension influenced dopamine activity only to the extent that it influenced choice. Thus, vastly different reward types such as juice and food activated dopamine neurons in accordance with subjective value derived from the different rewards. This neuronal signal could serve to update value signals for economic choice.
PDF | SUPPL | SLIDE
No comments:
Post a Comment