Decomposing the roles of perseveration and expected value representation in models of the Iowa gambling task

Abstract

Models of human behavior in the Iowa Gambling Task (IGT) have played a pivotal role in accounting for behavioral differences during decision-making. One critical difference between models that have been used to account for behavior in the IGT is the inclusion or exclusion of the assumption that participants tend to persevere, or stay with the same option over consecutive trials. Models that allow for this assumption include win-stay-lose-shift (WSLS) models and reinforcement learning (RL) models that include a decay learning rule where expected values for each option decay as they are chosen less often. One shortcoming of RL models that have included decay rules is that the tendency to persevere by sticking with the same option has been conflated with the tendency to select the option with the highest expected value because a single term is used to represent both of these tendencies. In the current work we isolate the tendencies to perseverate and to select the option with the highest expected value by including them as separate terms in a Value-Plus-Perseveration (VPP) RL model. Overall the VPP model provides a better fit to data from a large group of participants than models that include a single term to account for both perseveration and the representation of expected value. Simulations of each model show that the VPP model's simulated choices most closely resemble the decision-making behavior of human subjects. In addition, we also find that parameter estimates of loss aversion are more strongly correlated with performance when perseverative tendencies and expected value representations are decomposed as separate terms within the model. The results suggest that the tendency to persevere and the tendency to select the option that leads to the best net payoff are central components of decision-making behavior in the IGT. Future work should use this model to better examine decision-making behavior.

Data from 617 Healthy Participants Performing the Iowa Gambling Task: A “Many Labs” Collaboration

“…The number of trials varied between 95, 100, and 150 ( Table 1). 1 However, the total number of Worthy et al [17] 35 100 1 Undergraduate students, 22 female a Information that was provided in the original articles. This information consists of the mean age and the standard deviation in brackets, or alternatively the occupation of the participants.…”

“…In addition to the payment, some of the studies provided a monetary incentive depending on performance. Participants in studies without incentives (1) obtained course credits ( [6]; [9]; [15]; [17]); or (2) were paid a fixed amount for participation ([5]; [7]). Participants in studies with incentives (3) were paid a fixed amount for participation and received an additional bonus depending on the overall amount won on the IGT (Horstmann); (4) were paid a fixed amount and received an additional bonus if they had accumulated the biggest overall amount won across all participants [3]; or (5) could choose between options (1) and (2) -a choice that participants had to make before the start of the experiment -and received an additional bonus depending on the overall amount won on the IGT [8].…”

“…This data pool comes from eight independent published studies ( [3]; Horstmann; [5]- [7]; [15]- [17]), one study under preparation ( [8]), and one unpublished study ([9]). …”

Neural Correlates of State-Based Decision-Making in Younger and Older adults

“…The inclusion of the perseveration term markedly improved the fit of the RL model as can be seen by comparing the AIC values of the model with and without the perseveration term. Accounting for participants' tendencies to perseverate can drastically improve fits of RL models and is the main reason why the WSLS model can fit data from a variety of tasks as well as RL models (Worthy, Hawthorne, & Otto, 2013;Worthy, Pang, & Byrne, 2013).…”

Computational EEG modelling of decision making under ambiguity reveals spatio-temporal dynamics of outcome evaluation

“…Computational models such as the Expectancy-Valence (EV) model (Busemeyer and Stout 2002), the Perseverance Valence Learning (PVL) model, the PVL-Delta model (Ahn et al 2008(Ahn et al , 2011Fridberg et al 2010), and the Value Plus Perseverance (VPP) model (Worthy et al 2013), consider factors such as the attention given to outcome valence (i.e., to wins vs. losses), how the recency of feedback affects future decisions, and how choices are influenced by experience (i.e., to what extent choices are random). Such models assume that the valence experienced on each trial informs a probabilistic choice mechanism, and quantifies outcome expectation and prediction error on an individual trial-by-trial basis, thereby estimating individuals' subjective experiences of the task and task-expectations, rather than objective task outcomes (Yechiam and Busemeyer 2005).…”