Toward a rodent model of the Iowa gambling task

The Iowa gambling task in humans is, in principle, suited for the study of the long-term efficiency of behavior in a biologically relevant context. Key features of this task are uncertainty of outcomes and a conflict between the immediate and the long-term payoff options. Animal models allow us to study the underlying neurobiology of decision-making processes and the long-term efficiency of behavior in more detail and at a greater depth than is possible in humans. Therefore, we set out to develop a model of this task in rodents, using the task’s key features. In this article, we describe the results of the first series of experiments with rats and mice. The data thus far suggest that mice and rats behave in a way similar to humans; that is, they tend to choose the option with the best long-term payoff more often as the test progresses.     

Affective guidance in the Iowa gambling task

It has been suggested that affective states can guide higher level cognitive processes and that such affective guidance may be particularly important when real-life decisions are made under uncertainty. We ask whether affect guides decisions in a laboratory task that models real-life decisions under uncertainty. In the Iowa gambling task (IGT), participants search for monetary payoffs in an uncertain environment. Recent evidence against an affective guidance interpretation of the IGT indicates a need to set a standard for what counts as evidence of affective guidance. We present a novel analysis of IGT, and our results show that participants’ galvanic skin response (GSR) reflects an affective process that precedes and guides cognition. Specifically, prior to participants’ knowledge of the optimal strategy, their GSRs are significantly higher when they are about to select from a bad deck, relative to a good deck, and this difference in GSR is correlated with a behavioral preference for the good deck.     

Bayesian parameter estimation in the Expectancy Valence model of the Iowa gambling task

The purpose of the popular Iowa gambling task is to study decision making deficits in clinical populations by mimicking real-life decision making in an experimental context. Busemeyer and Stout [Busemeyer, J. R., & Stout, J. C. (2002). A contribution of cognitive decision models to clinical assessment: Decomposing performance on the Bechara gambling task. Psychological Assessment, 14, 253-262] proposed an “Expectancy Valence” reinforcement learning model that estimates three latent components which are assumed to jointly determine choice behavior in the Iowa gambling task: weighing of wins versus losses, memory for past payoffs, and response consistency. In this article we explore the statistical properties of the Expectancy Valence model. We first demonstrate the difficulty of applying the model on the level of a single participant, we then propose and implement a Bayesian hierarchical estimation procedure to coherently combine information from different participants, and we finally apply the Bayesian estimation procedure to data from an experiment designed to provide a test of specific influence.     

Episodic chasing in pathological gamblers using the Iowa gambling task

"Chasing ones losses" is a key symptom among pathological gamblers (PGs). This study focuses on quantitative differences in episodic chasing (i.e., sequences of disadvantageous decisions within a single gambling session) between PGs and non-pathological gamblers (NPGs). We compared 61 PGs and 39 NPGs on the Iowa Gambling Task (IGT) and the Zuckerman Sensation Seeking Scale (SSS). The PGs showed significantly more chasing and had significantly poorer decision-making strategies than NPGs, particularly among males (F = 4.52, p < 0.05). Random players were significantly less sensation seeking than advantageous and disadvantageous (i.e., chasing) players, but there was no interaction with group or gender. The results suggest that quantifiable within-session gambling behavior holds important implications for detecting underlying vulnerabilities to gambling pathology.     

The effect of age on decision making according to the Iowa gambling task

Aging has been associated with several brain changes that often affect the cognitive functioning of adults, but changes in executive functions, particularly in the field of decision making, have not been fully investigated. The Iowa Gambling Task (IGT) is a widely used tool to evaluate decision making, but little is known about the effect of age on its results. This study used the IGT and compared healthy young (n = 40) and elderly (n = 40) adults to evaluate their decision making processes. There were significant differences in the learning curve of the two age groups, but no difference in overall IGT performance. The results for the first IGT block were different from those of the rest of the task, and the group of elderly adults had a better performance in this block. Elderly adults also showed a preference for deck A, the one that resulted in greater losses. Findings confirmed the results of other studies, which reported that the IGT block score is the variable with the greatest sensitivity to age in this instrument.     

The uncertainty appraisal enhances the prominent deck B effect in the Iowa gambling task

The Iowa gambling task (Bechara et al., Cognition 50:7–15, 1994) is designed to simulate a decision making problem under ambiguity, in which the degree of reliance on emotional cues arising from previous experiences contributes to perform advantageously. Recent studies based on the appraisal tendency framework demonstrated that emotional certainty (associated with intuitive strategies) leads to a more advantageous decision pattern, whereas emotional uncertainty (associated with deliberative strategies) impairs the performance in the IGT (Bagneux et al., Motivation and Emotion 37(4):818–827, 2013; Bollon and Bagneux, Cognition and Emotion 27(2):376–384, 2013). Due to the problems in the IGT (Dunn et al., Neuroscience and Biobehavioral Reviews 30:239–271, 2006; Steingroever et al., Psychological Assessment 25(1):180–193, 2013), however, it is an open question to what extent the disadvantageous IGT performance in the uncertainty conditions was based on risky decision making. Addressing the main criticisms on the IGT, the primary aim of the present study is to provide a further explanation for the underlying source of the IGT impairment led by uncertainty appraisals. In line with previous research, we found that participants in the certainty-associated emotion condition (disgust) outperformed those in uncertainty-associated conditions (fear, sadness) in the gambling game. Detailed four-deck format analyses on decision patterns and knowledge levels provided supporting evidence for our main hypothesis that the weak IGT scores in the uncertainty conditions can be summarized as a failure to anticipate the badness and the goodness of the most difficult decks, and a dominant preference for a risky option with high immediate gains and infrequent losses.     

Heterogeneity of strategy use in the Iowa gambling task: A comparison of win-stay/lose-shift and reinforcement learning models

The Iowa gambling task (IGT) has been used in numerous studies, often to examine decision-making performance in different clinical populations. Reinforcement learning (RL) models such as the expectancy valence (EV) model have often been used to characterize choice behavior in this work, and accordingly, parameter differences from these models have been used to examine differences in decision-making processes between different populations. These RL models assume a strategy whereby participants incrementally update the expected rewards for each option and probabilistically select options with higher expected rewards. Here we show that a formal model that assumes a win-stay/lose-shift (WSLS) strategy—which is sensitive only to the outcome of the previous choice—provides the best fit to IGT data from about half of our sample of healthy young adults, and that a prospect valence learning (PVL) model that utilizes a decay reinforcement learning rule provides the best fit to the other half of the data. Further analyses suggested that the better fits of the WSLS model to many participants’ data were not due to an enhanced ability of the WSLS model to mimic the RL strategy assumed by the PVL and EV models. These results suggest that WSLS is a common strategy in the IGT and that both heuristic-based and RL-based models should be used to inform decision-making behavior in the IGT and similar choice tasks.     

A simple model for the utility of gambling

A model of the utility of gambling is presented in a modified von Neumann-Morgenstern format. Axioms imply a utility function that preserves preferences between sure things and between gambles. The addition of a utility of gambling term to the expected utility of a gamble preserves preference comparisons between gambles and sure things. Aspects of the utility of gambling are noted, and comparisons are made to standard concepts of risk attitudes.The author is indebted to Joseph Sani for valuable discussions on the topic of this paper.     

Complexity effects on the children's gambling task

The children's gambling task (CGT [Kerr, A., & Zelazo, P. D. (2004). Development of “Hot” executive function: The children's gambling task. Brain and Cognition, 55, 148–157]) involves integrating information about losses and gains to maximize winnings when selecting cards from two decks. Both cognitive complexity and control (CCC) theory and relational complexity (RC) theory attribute younger children's difficulty to task complexity. In CCC theory, identification of the advantageous deck requires formulation of a higher-order rule so that gains and losses can be considered in contradistinction. According to RC theory, it entails processing the ternary relation linking three variables (deck, magnitude of gain, magnitude of loss). We designed two less complex binary-relational versions in which either loss or gain varied across decks, with the other held constant. The three closely matched versions were administered to 3–5-year-olds. Consistent with complexity explanations, children in all age groups selected cards from the advantageous deck in the binary-relational versions, but only 5-year-olds did so on the ternary-relational CGT.     

A chain-retrieval model for voluntary task switching

To account for the findings obtained in voluntary task switching, this article describes and tests the chain-retrieval model. This model postulates that voluntary task selection involves retrieval of task information from long-term memory, which is then used to guide task selection and task execution. The model assumes that the retrieved information consists of acquired sequences (or chains) of tasks, that selection may be biased towards chains containing more task repetitions and that bottom-up triggered repetitions may overrule the intended task. To test this model, four experiments are reported. In Studies 1 and 2, sequences of task choices and the corresponding transition sequences (task repetitions or switches) were analyzed with the help of dependency statistics. The free parameters of the chain-retrieval model were estimated on the observed task sequences and these estimates were used to predict autocorrelations of tasks and transitions. In Studies 3 and 4, sequences of hand choices and their transitions were analyzed similarly. In all studies, the chain-retrieval model yielded better fits and predictions than statistical models of event choice. In applications to voluntary task switching (Studies 1 and 2), all three parameters of the model were needed to account for the data. When no task switching was required (Studies 3 and 4), the chain-retrieval model could account for the data with one or two parameters clamped to a neutral value. Implications for our understanding of voluntary task selection and broader theoretical implications are discussed.     

Adolescents’ heightened risk-seeking in a probabilistic gambling task

This study investigated adolescent males’ decision-making under risk, and the emotional response to decision outcomes, using a probabilistic gambling task designed to evoke counterfactually mediated emotions (relief and regret). Participants were 20 adolescents (aged 9–11), 26 young adolescents (aged 12–15), 20 mid-adolescents (aged 15–18) and 17 adults (aged 25–35). All were male. The ability to maximize expected value improved with age. However, there was an inverted U-shaped developmental pattern for risk-seeking. The age at which risk-taking was highest was 14.38 years. Although emotion ratings overall did not differ across age, there was an increase between childhood and young adolescence in the strength of counterfactually mediated emotions (relief and regret) reported after receiving feedback about the gamble outcome. We suggest that continuing development of the emotional response to outcomes may be a factor contributing to adolescents’ risky behaviour.     

Affective biasing of choices in gambling task decision making

The proponents of the somatic marker hypothesis presume that rational decision making is guided by emotional reactions that are developed from prior experience. Supporting evidence for the hypothesis comes almost exclusively from the short-term affective reactions that are learned during the course of a hypothetical decision-making task—the gambling task (GT). We examined GT performance and affective reactions to choices when those choices were biased by words that had preexisting affective value. In one experiment, affectively valued words directly signaled good and bad choices. A congruent relation between affective value of word and choice outcome improved GT performance, whereas an incongruent relation greatly interfered with performance. In another experiment, affectively valued words were maintained as a working memory (WM) load between GT choices. A WM load with affectively positive words somewhat improved GT performance, whereas affectively negative words interfered with performance. Somatic markers—indicated by differential anticipatory skin conductance response (SCR ) amplitude for good and bad choices—appeared at a point in the GT session when choice performance was superior. However, differential SCR developed during the session after good choice performance was already established. These results indicate that preexisting affective biases can influence GT decision making. In addition, the somatic markers that are regular accompaniments of GT decision making appeared to be temporally lagging indicators of choice performance. 2006 Psychonomic Society, Inc.     

A model for evidence accumulation in the lexical decision task

We present a new model for lexical decision, REM-LD, that is based on REM theory (e.g., ). REM-LD uses a principled (i.e., Bayes' rule) decision process that simultaneously considers the diagnosticity of the evidence for the 'WORD' response and the 'NONWORD' response. The model calculates the odds ratio that the presented stimulus is a word or a nonword by averaging likelihood ratios for lexical entries from a small neighborhood of similar words. We report two experiments that used a signal-to-respond paradigm to obtain information about the time course of lexical processing. Experiment 1 verified the prediction of the model that the frequency of the word stimuli affects performance for nonword stimuli. Experiment 2 was done to study the effects of nonword lexicality, word frequency, and repetition priming and to demonstrate how REM-LD can account for the observed results. We discuss how REM-LD could be extended to account for effects of phonology such as the pseudohomophone effect, and how REM-LD can predict response times in the traditional 'respond-when-ready' paradigm.     

A model for integrating fixed-, random-, and mixed-effects meta-analyses into structural equation modeling

Meta-analysis and structural equation modeling (SEM) are two important statistical methods in the behavioral, social, and medical sciences. They are generally treated as two unrelated topics in the literature. The present article proposes a model to integrate fixed-, random-, and mixed-effects meta-analyses into the SEM framework. By applying an appropriate transformation on the data, studies in a meta-analysis can be analyzed as subjects in a structural equation model. This article also highlights some practical benefits of using the SEM approach to conduct a meta-analysis. Specifically, the SEM-based meta-analysis can be used to handle missing covariates, to quantify the heterogeneity of effect sizes, and to address the heterogeneity of effect sizes with mixture models. Examples are used to illustrate the equivalence between the conventional meta-analysis and the SEM-based meta-analysis. Future directions on and issues related to the SEM-based meta-analysis are discussed.     

A Newton procedure for conditionally linear mixed-effects models

This article reviews Newton procedures for the analysis of mean and covariance structures that may be functions of parameters that enter a model nonlinearly. The kind of model considered is a mixed-effects model that is conditionally linear with regard to its parameters. This means parameters entering the model nonlinearly must be fixed, whereas those entering linearly may vary across individuals. This framework encompasses several models, including hierarchical linear models, linear and nonlinear factor analysis models, and nonlinear latent curve models. A full maximum-likelihood estimation procedure is described. Mx, a statistical software package often used to estimate structural equation models, is considered for estimation of these models. An example with Mx syntax is provided.     

Development of "hot" executive function: the children's gambling task

Development of affective decision-making was studied in 48 children at two ages (3 and 4 years) using a simplified version of the Iowa Gambling Task (). On each of 50 trials, children chose from 1 of 2 decks of cards that, when turned, displayed happy and sad faces, corresponding to rewards (candies) won and lost, respectively. Cards in 1 deck offered more rewards per trial, but were disadvantageous across trials due to occasional large losses; cards in the other deck offered fewer rewards per trial, but were advantageous overall. On later trials, 4-year-olds made more advantageous choices than 3-year-olds, and 4-year-olds made more advantageous choices than would be expected by chance, whereas 3-year-olds made more disadvantageous choices than would be expected by chance. These findings, which were especially pronounced for girls, indicate that affective decision-making develops rapidly during the preschool period, possibly reflecting the growth of neural systems involving orbitofrontal cortex.     

A model of the go/no-go task

In this article, the first explicit, theory-based comparison of 2-choice and go/no-go variants of 3 experimental tasks is presented. Prior research has questioned whether the underlying core-information processing is different for the 2 variants of a task or whether they differ mostly in response demands. The authors examined 4 different diffusion models for the go/no-go variant of each task along with a standard diffusion model for the 2-choice variant (R. Ratcliff, 1978). The 2-choice and the go/no-go models were fit to data from 4 lexical decision experiments, 1 numerosity discrimination experiment, and 1 recognition memory experiment, each with 2-choice and go/no-go variants. The models that assumed an implicit decision criterion for no-go responses produced better fits than models that did not. The best model was one in which only response criteria and the nondecisional components of processing changed between the 2 variants, supporting the view that the core information on which decisions are based is not different between them.     

A contribution of cognitive decision models to clinical assessment: decomposing performance on the Bechara gambling task

The Bechara simulated gambling task is a popular method of examining decision-making deficits exhibited by people with brain damage, psychopathology, antisocial personality, or drug abuse problems. However, performance on this task is confounded by complex interdependencies between cognitive, motivational, and response processes, making it difficult to sort out and identify the specific processes responsible for the observed behavioral deficits. The authors compare 3 competing cognitive decision models of the Bechara task in terms of their ability to explain the performance deficits observed in Huntington's disease patients as compared with healthy populations and people with Parkinson's disease. The parameters of the best fitting model are used to decompose the observed performance deficit of the Huntington patients into cognitive, motivational, and response sources.     

Feedback related brain activity in a gambling task: A temporal analysis of EEG correlates

Lucchiari, C. & Pravettoni, G. (2010). Feedback related brain activity in a gambling task: A temporal analysis of EEG correlates. Scandinavian Journal of Psychology 51, 449–454. The pattern of neural correlates of feedback processing has been the subject of a number of studies, using both neuroimaging and electrophysiological recordings. A complex functional network was found to be activated after a choice in order to process a feedback and sustain an adaptive behavior. However, many aspects of this network are still unclear and further research is needed to better understand this process. We conducted an EEG study using a simple gambling task. Twenty three subjects participated to the study. We analyzed both EEG power spectrum and ERP components evoked by presentation of a feedback signal (money gain or loss) during a simple gambling task. Our data confirmed that a negative ERP component is present about 270 ms after feedback, particularly relevant following a choice with negative outcome. Furthermore, the theta and delta oscillatory activity seem to be correlated to a dynamic decision‐making process within specific cortical networks. In particular, theta activity showed a valence dependent development between 150 and 350 ms post‐feedback onset. Differently from previous studies (Cohen, Elger & Ranganath, 2007; Marco‐Pallares, Cucurell, Cunillera et al., 2008), we did not find any valence effect in beta range. However, our data are consistent with Christie and Tata (2009) , probably due to the nature of the gambling task used in both studies. In conclusion, our data, in line with some prior findings showed that the feedback related response is correlated to a complex pattern of cortical activation probably mediated by theta and delta activity.     

A stochastic model for inter-keypress times in a typing task

A mathematical model for the inter-keypress times (IKT) in a typing task is proposed. The model, which includes a diffusion process terminated by a single response threshold, was evaluated using data obtained from typists. The differences in performance for successive cross-hand and within-hand keypresses were examined using IKT distributions and hazard functions, and it was shown that the empirical hazard functions could be fit by the theoretical hazard function derived from the convolution of normal and inverse Gaussian random variables. Some possible applications of the model for the evaluation of fatigue and strategic effects in typing are suggested.