Dysfunctional decision-making in pathological gambling: Pattern specificity and the role of impulsivity

Abstract

Dysfunctional decision-making in individuals with pathological gambling (PGs) may result from dominating reward-driven processes, indicated by higher impulsivity. In the current study we examined (1) if PGs show specific decision-making impairments related to dominating reward-driven processes rather than to strategic planning deficits and (2) whether these impairments are related to impulsivity. Nineteen PGs according to DSM-IV and 19 matched control subjects undertook the Cambridge Gambling Task (CGT) to assess decision-making. The delay discounting paradigm (DDP) as well as the UPPS Impulsive Behavior Scale (measuring urgency, premeditation, perseverance and sensation seeking) were administered as multidimensional measures of impulsivity. Results revealed that (1) PGs exhibited higher risk seeking and an immediate reward focus in the CGT and, in contrast, comparable strategic planning to the control group. (2) Decision-making impairments were related to more severe delay discounting and, specifically, to increased urgency and less premeditation. Our findings suggest (1) the necessity to disentangle decision-making components in order to improve etiological models of PGs, and (2) that urgency and premeditation are specifically related to disadvantageous decision-making and should be tackled in intervention strategies focusing on emotion tolerance and control strategies.

Introduction

While it is well known, that individuals diagnosed with pathological gambling (PGs) display impaired performance in neuropsychological decision-making tasks (e.g. Petry, 2001b, Cavedini et al., 2002, Brand et al., 2005, Goudriaan et al., 2005, Linnet et al., 2006, Labudda et al., 2007, Forbush et al., 2008, Roca et al., 2008, Lawrence et al., 2009, Kertzman et al., 2011) and heightened impulsivity (Goudriaan et al., 2004, Verdejo-Garcia et al., 2008, van Holst et al., 2010), the relation of both constructs in PGs remains unclear. Findings from previous studies on this issue in non-gambling samples were inconsistent (e.g. Monterosso et al., 2001, Jollant et al., 2005, Zermatten et al., 2005, Suhr and Tsanadis, 2007, Dolan et al., 2008, Franken et al., 2008, Sweitzer et al., 2008, Janis and Nock, 2009, Perales et al., 2009, Xiao et al., 2009, Billieux et al., 2010) which may result from a neglected consideration of the lower-order sub-components, for either or both constructs. Thus, the present study aims to assess the relation of impaired decision-making and impulsivity in PGs taking a multidimensional perspective.

In line with recent models on pathways of addictions (Bechara, 2005, Bühringer et al., 2008, Redish et al., 2008b, Goldstein and Volkow, 2011) and specifically of pathological gambling (Bechara, 2003, Evans and Coventry, 2006, van Holst et al., 2010), we assume an important role of an – either antecedent or resultant – imbalance of (more automatic) motivational and valuation brain networks and (more reflective) cognitive control networks. According to these models, the motivational and valuation systems in PGs may overestimate the value of immediate short-term rewards. This would explain, on the one hand, the heightened impulsivity found in PGs, because the tendency towards immediate rewards while disregarding negative consequences is frequently considered a central aspect of impulsivity (Moeller et al., 2001, Bechara, 2005, Verdejo-Garcia et al., 2008). On the other hand, this imbalance may result in disadvantageous gambling-related decisions, as dominating reward-driven processes were found to be strongly related to impaired decision-making (Krawczyk, 2002, Fellows, 2004, Yechiam et al., 2005, Dunn et al., 2006, Rangel et al., 2008). Due to the lack of a consensus model of decision-making, a broad variety of tasks has been used to measure the construct (Fellows, 2004). However, a fractionating of lower-order decision-making components may be an advantageous approach to gain a deeper understanding of the complexity of underlying processing mechanisms (see e.g. Busemeyer and Stout, 2002, Krawczyk, 2002, Clark et al., 2004, Brand et al., 2005, Fellows and Farah, 2005, Yechiam et al., 2005, Diekhof et al., 2008). Following this approach, it can be expected that PGs will show specific impairments in decision-making sub-components associated with reward and valuation. These decision-components have often been related to processes in the ventromedial regions of the prefrontal cortex (PFC) (e.g. Brand et al., 2006, Lawrence et al., 2008, Clark, 2010). Indeed, PGs display functional changes in these brain areas in tasks of reward processing and decision-making (Potenza et al., 2003, Reuter et al., 2005, Potenza, 2008, Clark, 2010). In contrast, decision-making components like reasoning or strategic planning – often related to processes in the dorsolateral PFC – may be unchanged in PGs (Lawrence et al., 2009, Clark, 2010). A neuropsychological task which differentiates between different components of decision-making behavior is the Cambridge Gambling Task (CGT) (Rogers et al., 1999). The CGT assesses parameters related to risk seeking (e.g. ‘risk taking’), an immediate reward focus (e.g. ‘delay aversion’) or strategic planning (e.g. ‘percentage of rational choices’). First studies using the CGT in gamblers and PGs revealed disorder-specific deficits in risk seeking rather than in strategic planning (Lawrence et al., 2008, Grant et al., 2011). Crucially, both studies did not explicitly report on other parameters related to reward-driven processes like delay aversion or chasing behavior within a task (i.e. betting larger sums of money or taking greater risks with an intention to recover prior losses immediately) (O’Connor and Dickerson, 2003, Linnet et al., 2006).

According to our present hypothesis, the impaired reward- and valuation-related components of decision-making are positively related to impulsivity, in contrast to the strategic components of decision-making (Lawrence et al., 2008). To our best knowledge, to date, there are no studies on the relation of impaired decision-making components and impulsivity in PGs. Studies on this relation in non-gambling samples have mainly been conducted without considering the aforementioned multidimensional nature of decision-making and of impulsivity itself (e.g. Reynolds et al., 2006, Verdejo-Garcia et al., 2008, Broos et al., 2012). Especially, the behaviorally assessed impulsivity dimension delay discounting (Bechara, 2003, Franken et al., 2008) as well as self-reported dimensions concerning the tendency to act rashly in an emotional context (urgency) and the lack of forethought (Whiteside and Lynam, 2001) may be relevant mechanisms underlying impaired decision-making (Enticott and Ogloff, 2006). Hitherto, studies with non-gambling samples assessing the relation of delay discounting and decision-making yielded inconsistent results (Monterosso et al., 2001, Janis and Nock, 2009), whereas studies on urgency and premeditation underpin our hypothesis (Zermatten et al., 2005, Billieux et al., 2010).

In summary, we expected a specific pattern of impaired decision-making components in PGs which is related to an imbalance of reward- and valuation-related and reflection-related brain networks, as indicated by increased risk seeking and an immediate reward focus. Further, we hypothesized that impulsivity is an important indicator of this imbalance. Thus, we expect that the impaired decision-making components are specifically related to the impulsivity dimensions delay discounting, urgency and premeditation. As impaired decision-making is important in the development (e.g. starting to gamble regularly), maintenance (e.g. chasing behavior) and cessation (as well as relapse) of pathological gambling, it is of central importance to understand its components and underlying mechanisms. If a dominating reward- and valuation-system, indicated by an elevated impulsivity, turns out to be the core mechanism rather than problems in reasoning or strategic planning, this would be an important result for the adaptation of etiological models as well as therapeutic strategies.