Conference on Cognitive Computational Neuroscience

Many real-life decisions involve balancing multiple needs at the same time. To capture such complexity, We can conceptualize reward processing as navigating in an abstract value space. To tap into the individual differences in ‘value navigation’, we designed a foraging task where participants pick fruits to satisfy one of two needs, hydration and energy. The points they earn map onto how close they are to the goal state in Euclidean distance. In three studies (N = 685), we found that people systematically preferred the reward options that satisfied a single need (unidimensional bias), even when the option that satisfies both needs rendered higher points. Surprisingly, participants who reported a reduced capacity to imagine future pleasure (anticipatory anhedonia) had a smaller unidimensional bias, and therefore had better task performance. This may suggest that participants with anticipatory anhedonia represented values more veridically in a task that requires value representation in Euclidean geometry.