Value coding by primate amygdala neurons complies with economic choice theory

Abstract

Primate amygdala neurons are implicated in coding the value of choice options for decision-making. However, it remains unclear whether amygdala value signals comply with the notion that decision makers maximize the value of rewards, as formalized by key axioms of economic theory. In particular, the continuity axiom of Expected Utility Theory (EUT) postulates that reward maximization is reflected by a trade-off between reward probability and magnitude. Our experiment tested this trade-off in two male macaque monkeys who subjectively ranked three gambles in choices, thus revealing their subjective value. Axiom compliance was defined by choice indifference between the intermediate gamble and a probabilistic combination of the other two gambles. We found that the animals choices between safe and gamble options reflected the integration of probability and magnitude into scalar values consistent with the continuity axiom. In a non-choice task, responses of individual amygdala neurons to separate probability and magnitude cues reflected the monkeys individual preferences: neuronal responses were lower for non-preferred gambles (relative to safe options), higher for preferred gambles and equal at subjective indifference between safe and gamble options. In a choice task, amygdala neurons integrated probability and magnitude into subjective values that reflected individual preferences. In some neurons, value signals transitioned to signals coding the monkeys economic choices and chosen values. These findings identify amygdala neurons as substrates for encoding behavior-matching subjective values according to the continuity axiom of EUT, and for translating these values into economic choices. SIGNIFICANCE STATEMENTThe amygdala--a brain structure in the temporal lobe--has long been associated with emotional responses. More recently, primate amygdala neurons have been implicated in economic decision-making. Here we show that amygdala neurons respond to visual cues indicating reward probability and magnitude in a manner that reflects the integration of these variables into subjective values. Importantly, these value signals complied with the notion that decision makers maximize the value of rewards, as formalized by key axioms of economic theory. Beyond valuation, some neurons processed subjective values into signals predictive of economic choices. Dysfunctional processing of values and choices by the amygdala could play a role in psychiatric and behavioral disorders affecting mental health, in which the amygdala is implicated.