Hyposensitivity to losses under risk but hypersensitivity to gains under ambiguity during feedback evaluation

Abstract

Uncertainty is ubiquitous in human life and can be fractioned into risk (known probability distribution) and ambiguity (unknown probability distribution), each with distinct functional correlates. This event-related potential study examined how contextual valence influences the dissociation between risk and ambiguity during feedback evaluation through the lens of neural dynamics. We manipulated contextual valence as a gain versus a loss context. In the gain context, decisions resulted in either gains or nongains, while in the loss context, decisions led to losses or nonlosses. We recorded EEG from 40 participants while they completed a wheel-of-fortune task under conditions of risk and ambiguity in both contexts. We observed a stronger valence effect on the P3a in the loss context under risk, but a stronger valence effect on the P3b in the gain context under ambiguity. Further comparisons revealed that feedback evaluation was primarily driven by a smaller P3a in response to losses under risk, but by a larger P3a and P3b in response to gains under ambiguity. Parametric analyses found that both the reward positivity and P3a for gains and nongains were modulated by winning probability under risk, while the P3a for gains was influenced by ambiguity level under ambiguity. Our findings demonstrate the dissociable influences of contextual valence on feedback-related neural dynamics based on uncertainty type, supporting a critical role of valence-asymmetry in distinguishing risk from ambiguity.