RLDM Conference 2017 Conference Abstract
Reinforcement and Valence Effects on Incentive Integration and Motivated Cognitive Control
- Debbie Yee
- Todd Braver
It is unequivocal that motivational incentives play a central role in influencing goal-directed be- havior. However, most studies of motivation and decision-making in humans have typically used monetary rewards, and have rarely considered the integrated influence from diverse sources of motivation on behavior. Another motivational dimension that has recently garnered attention is valence, but only a handful of studies have compared both appetitive and aversive motivation in terms of their impact on cognitively demanding tasks. A third question relates whether the reinforcement feedback of a symbolic cue (i. e. , it contains the same information) influences behavior. To examine the dissociable effects of these motivational dimensions (incentive integration, valence, reinforcement), we utilize a novel paradigm which examines the integrated influence of primary incentives (e. g. , juice, saltwater) and secondary incentives (e. g. , money) on cognitive control. In the study, valence was manipulated by comparing monetary gains vs losses across task condi- tions and by liquid type (juice, neutral, saltwater), and reinforcement was manipulated by comparing liquid feedback for positive verses negative outcomes. All conditions were manipulated within-subject. Results revealed significant effects of monetary reward [b=. 05, t(759)=6. 58, p¡. 001] and liquid [b=. 05, t(759)=6. 22, p¡. 001] on reward rate, as well as significant two-way interactions between feedback and liquid [b=-. 07, t(759)=-6. 135, p¡. 001] and feedback and monetary reward [b=-. 02 t(759)=-1. 99, p=. 047]. Participants im- proved performance when liquid incentives were delivered as feedback upon poor, compared to successful, performance. Notably, liquid valence had opposing effects across feedback conditions, driven by the inte- grated influence of appetitive and aversive incentives. Collectively, these data provide empirical evidence for dissociable effects of reinforcement and valence on motivated cognitive control.