Effects of affective arousal on choice behavior, reward prediction errors, and feedback-related negativities in human reward-based decision making

Abstract

Emotional experience has a pervasive impact on choice behavior, yet the underlying mechanism remains unclear. Introducing facial-expression primes into a probabilistic learning task, we investigated how affective arousal regulates reward-related choice based on behavioral, model fitting, and feedback-related negativity (FRN) data. Sixty-six paid subjects were randomly assigned to the Neutral-Neutral (NN), Angry-Neutral (AN), and Happy-Neutral (HN) groups. A total of 960 trials were conducted. Subjects in each group were randomly exposed to half trials of the pre-determined emotional faces and another half of the neutral faces before choosing between two cards drawn from two decks with different assigned reward probabilities. Trial-by-trial data were fit with a standard reinforcement learning model using the Bayesian estimation approach. The temporal dynamics of brain activity were simultaneously recorded and analyzed using event-related potentials. Our analyses revealed that subjects in the NN group gained more reward values than those in the other two groups; they also exhibited comparatively differential estimated model-parameter values for reward prediction errors. Computing the difference wave of FRNs in reward vs. non-reward trials, we found that, compared to the NN group, subjects in the AN and HN groups had larger “General” FRNs (i.e., FRNs in no-reward trials minus FRNs in reward trials) and “Expected” FRNs (i.e., FRNs in expected reward-omission trials minus FRNs in expected reward-delivery trials), indicating an interruption in predicting reward. Further, both AN and HN groups appeared to be more sensitive to negative outcomes than the NN group. Collectively, our study suggests that affective arousal negatively regulates reward-related choice, probably through overweighting with negative feedback.

Misfortune may be a blessing in disguise: Fairness perception and emotion modulate decision making

“…Reinforcement learning models (Glimcher, 2011;Niv, 2009;Sutton & Barto, 1998) have been used to investigate the process underlying trial-by-trial choices in feedback-based probabilistic learning tasks (e.g., Ahn, Busemeyer, Wagenmakers, & Stout, 2008;Ahn et al, 2014;Li et al, 2014;Liu et al, 2015;Rutledge et al, 2009). These models enable an assessment of the degree to which the participant updates his/her belief in response to feedback.…”

“…Making appropriate decisions is vital for survival and daily life. It requires the ability to update information regarding alternatives based on previous experiences and has frequently been demonstrated to be influenced by social and affective factors (Bechara, ; Liu, Hsieh, Hsu, & Lai, ; Pessoa, ; Watanabe, Sakagami, & Haruno, ). However, the precise impact of emotional experience on decision making remains debated (Barth & Funke, ; Isen, ; Knutson, Wimmer, Kuhnen, & Winkielman, ; Lyubomirsky, King, & Diener, ).…”

“…Furthermore, converging evidence has indicated that the FRN is modulated by long‐term affective traits and short‐term emotional experiences (Hajcak, McDonald, & Simons, ; Hajcak & Simons, ; Wiswede, Münte, Goschke, & Rüsseler, ). In a recent study, Liu et al () inserted facial expression cues into the course of a probabilistic gambling task to elicit affective arousal and used the FRN and a reinforcement learning model to demonstrate the pronounced impact of affective arousal on the dynamics of RPE signaling during decision making. In addition, existing literature suggests that the core subcomponent of the FRN, a positive deflection to the outcome feedback known as the reward positivity (RewP), is considered a neural indicator of reward sensitivity (see Proudfit, , for a review).…”

“…To understand the underlying mechanism of overt choice behavior, researchers have extensively studied reward-based decision making within the framework of reinforcement learning (e.g., Sutton & Barto, 1998) in recent years (Bayer & Glimcher, 2005;Glimcher, 2011;Liu et al, 2015;Rutledge et al, 2009;Schultz, Tremblay, & Hollerman, 1998). The reinforcement learning theory posits that the optimal choice of an organism is to maximize its reward through the minimization of the reward prediction error (RPE), which is defined as the difference between the expected and actual outcomes.…”

“…We developed an innovative version of the dictator game (Kahneman, Knetsch, & Thaler, 1986), which was modified from the ultimatum game (G€ uth, Schmittberger, & Schwarze, 1982), to induce emotional experiences in participants via the equality of a monetary assignment by the dictator. The monetary gain was subsequently used as the initial asset in a probabilistic gambling task (Li, Lai, Liu, & Hsu, 2014;Liu et al, 2015). In addition to conventional behavioral statistics, a reinforcement learning model (Glimcher, 2011;Niv, 2009;Sutton & Barto, 1998) using a hierarchical Bayesian estimation approach was adopted to fit the participants' trial-by-trial behavioral data to assess their latent process involved in the task, such as the sensitivity to loss aversion, the updating speed upon feedback, and the consistency of making decisions based on expected utility.…”