Dimensionality of Cognitions in Behavioral Addiction (2016)

L. S. Morris and V. Voon

Cognitive constructs provide conceptual frameworks for transpathological characterization and improved phenotyping of apparently disparate psychiatric groups. This dimensional approach can be applied to the examination of individuals with behavioral addictions, for example, towards gambling, video-games, the internet, food, and sex, allowing operationalization of core deficits. We use this approach to review constructs such as impulsivity, compulsivity, and attention regulation, which may be most relevant, applicable, and successful for the understanding and subsequent treatment of the addictions.

Attentional Bias

SUD are often characterized by attentional biases towards drug cues, a disturbance that facilitates craving in a cyclic, self-propagatory manner [47]. The relationship between attentional bias and craving remains despite treatment status [48]. Disruption of attentional regulation also appears to be relevant across a range of behavioral addictions. Individuals with PG both self-report [49] and demonstrate impairments in higher order attention processes [50, 51]. Similar to SUD and pathological gaming use, this deficit may reflect a shift in attention, as individuals with PG show biased and maintained attention towards gambling cues [52]. There is also evidence of early attentional biases towards food cues in BED, as well as difficulty disengaging from food cues, although the latter effect is also observed in healthy individuals [53]. Attentional bias towards internet-related stimuli has been reported in individuals with internet gaming addiction [54], and CSB is associated with an early attentional bias towards sexually explicit images [55]. Thus, a common influencing factor of attentional bias to the disorder-specific object seemingly presents across these behavioral aberrancies.

While the relationship between attentional bias and craving is yet to be explored across behavioral addictions, insights from studies of SUD suggest a strong link between the two, implicating a pathway towards pathological behavior, particularly driven by incentive motivation. It is difficult from these studies however to determine whether attentional biases pre-existed the disorder or were facilitated by them, although evidence from the SUD literature suggests the latter [47]. In line with this, modification of attentional biases to divert attentional resources from drug stimuli has had promising and clinically relevant effects [56, 57], with some effect on craving in smokers [58], although the generalizability of attentional training is currently unclear [59].

Impulsivity

Impulsivity, a tendency towards rapid, unplanned behaviors that are divorced from sufficient forethought and occur despite potential negative consequence, is well documented across a range of psychiatric disorders, including SUD [60, 61]. Impulsive behavior is now also demonstrated in behavioral addictions, including different groups of individuals with PG [43, 45, 62], problem gambling [63], and BED [26]. Self-reported impulsivity acts as a risk factor for pathological gaming in elementary and secondary schools in Singapore [64] and is associated with gambling severity in PG captured by the PG-YBOCS [65]. Self-reported impulsivity may also be higher in PG compared to that in SUD [66].

Impulsivity can be further decomposed into several discrete yet often overlapping constructs, subserved by dissociable neural systems [61]. Briefly, motor impulsivity describes a capacity for response inhibition or action cancelation; decisional impulsivity describes impulsive choice, modulated either by the influence, or lack, of prior evidence (reflection impulsivity) or by the temporal features of an outcome (delay discounting); and finally, waiting impulsivity describes the propensity towards disadvantageous premature responding. This heterogeneity suggests that there may be differing presentations across disorders [61].

Motor impulsivity can be captured with a Go/NoGo task or stop signal task (SST), in which responses are inhibited before or after response initiation, marking action restraint or action cancelation, respectively. Individuals with SUD show impaired performance on both the SST and Go/NoGo tasks, highlighted by a meta-analysis showing deficits particularly for stimulants and nicotine but not opioid or cannabis abuse [67••, 68]. Studies of motor impulsivity in PG have shown mixed results. Impaired performance during the Go/NoGo task have been reported [69] as well as no difference from healthy controls on the same task [67••]. Similarly, several studies have reported a lack of any difference during the SST compared to healthy controls [70–72] although a recent meta-analysis found a medium-large effect of impaired performance on the SST in gamblers [67••]. Since this group shows difficulties with target detection during the Go/NoGo task [69] and Go reaction time during the SST [67••], some of these effects may be related to inattention [67••]. Pathological video gamers have been reported to be unimpaired on the SST [73]. Interestingly, whereas problem gamers seemingly have reduced inhibitory control on the Go/NoGo task [63], pathological internet users have been shown to be more accurate on the same task compared to HV [74]. Thus, care must be taken around the generalizability of over-practiced computerized routine to laboratory tests, an effect that warrants further investigation. The role of motivational relevance of the stop cue is also of high importance; individuals with BED show consistent deficits in the Go/NoGo and SST but only in the context of a food cue, not with a neutral stimulus [75, 76]. Thus, impairments in response inhibition do not appear to be a uniformly observed across behavioral addictions, and some observed differences may indeed be related to extra-motor faculties.

In PG, BED and pathological gaming, emerging evidence suggests that deficits in impulsivity lie prominently within the realms of decision-making. Individuals with PG [49, 77–79], BED [80, 81], and pathological gaming [73] discount delayed rewards to a greater extent than healthy controls, meaning that smaller, immediate rewards are preferred over larger, delayed rewards. Although the same effect has been demonstrated in individuals with addictive disorders to drugs of abuse [68, 82, 83], this impairment may be more pronounced in individuals with PG. For example, individuals with PG show elevated delay discounting compared to cocaine dependent individuals [84], and gambling severity is a stronger predictor of discounting rate than substance use history or another self-reported measure of impulsivity [85]. Delay discounting is similarly observed in obese subjects with and without BED, although in obese subjects with high body mass index, those with BED show greater discounting across monetary, food reward, and massage time [81], implicating impairments in decisional impulsivity across reward types [86]. It should be noted that individuals with PG display difficulties in accurately perceiving time [43], a factor that may certainly contribute to decisions concerning delayed outcomes but yet to be directly tested.

Further evidence of impairments in decisional impulsivity arise from the use of the information sampling task (IST), which measures the tendency to sample or gather information before making a decision [46]. Individuals with PG and AUD show impairments in this measure of reflection impulsivity [46], highlighting shared deficits across substance-related and behavioral addictions. Pathological gamers similarly show less evidence accumulation before decision in the IST [73] and less beads drawn before decision in the beads task [87].

There are few studies assessing waiting impulsivity in humans with behavioral addictions. However, a recent report using a novel translational task assessing premature responses found no differences between individuals with BED and obese controls while stimulant-, alcohol-, and nicotine-dependent individuals were impaired [88]. Pathological video gamers made more premature responses compared to healthy controls but only in the context of comorbid nicotine use [73]. While waiting impulsivity has been shown to be impaired across a range of SUD, the effects may be drug state-dependent, as ex-smokers show normal levels of premature responses [88]. Further studies in PG and CSB are necessary before the extent of the dimensionality of waiting impulsivity across disorders of addiction is properly understood.

There is currently little empirical evidence of deficits in impulsivity in individuals with CSB [22]. Using a semi-structured clinical interview, traits of impulsivity were found to be common in a sample of 23 men and 2 females who met criteria for CSB [89], and a more recent study showed that CSB self-reports higher levels of impulsivity [55].

Cognitive Flexibility and Compulsivity

Measures of cognitive flexibility can highlight integrity of executive functions and the potential contribution of compulsive choice to pathological behavior. Cognitive flexibility has been assessed with the Wisconsin card sorting task (WCST) and the intra-dimensional extra-dimensional (IDED) set-shifting task. The WCST uses changing rules requiring flexible shifting of choices in the face of positive or negative feedback, in which the primary measure is of perseverative errors (continuing use of the same rule despite negative feedback, indicating compulsivity) or difficulty in task switching and inhibitory control. The IDED shifting task probes attentional set maintenance and conceptual set shifting, indicating cognitive shifting or flexibility.

Performance on both the WCST and IDED in SUD is inconsistent. Cocaine-dependent individuals are perseverative on the WCST but only during the initial stage of set-shifting [90]; however, poly-substance abusers have been shown to be no different from healthy controls [91]. There is no clear impairment on the WCST in alcohol dependence [92, 93] but acute alcohol does enhance preservative errors in healthy individuals [94]. Amphetamine but not opioid users show impairments at the crucial ED shift stage of the IDED task [95], although this effect was not replicated in a more recent study [96].

There is similar inconsistency in groups of behavioral addictions. While limited, evidence does however implicate impairments in set-shifting in BED [97•], as well as higher perseveration errors on the WCST compared to non-BED obese individuals [98] and problems maintaining set compared to obese controls [98] and individuals with anorexia nervosa [99]. However, the WCST has yielded inconsistent results amongst individuals with PG. Both enhanced errors of perseveration in PG females [100] and lack of a difference from healthy controls [101, 102] have been demonstrated. Increases in non-perseverative errors have been reported in PG during the WCST, suggesting that the observed impairments may not be specific to set-shifting but more of a broader cognitive dysfunction. Individuals with PG do seem to be impaired on IDED task [103], which improves with pharmacological intervention (memantine) [104].

While the literature for performance on tasks of cognitive flexibility in individuals with internet or gaming addiction is sparse, there is some evidence of impairments in set-shifting when a shift must be made between neutral and game-related stimuli [54] suggesting a specific effect of motivation rather than a general deficit in set shifting. Indeed, a recent study found no difference between individuals with internet addiction and healthy individuals during the IDED task [105•].

Another test of flexible behavior or compulsive choice is the probabilistic reversal learning (PRR) task, in which choice updating depends on a change or reversal of learned stimulus-outcome contingencies. Cocaine-dependent individuals are impaired at reversing in the face of a previously rewarded stimulus, largely perseverating for reward [106]. However, amphetamine, opiate [106], and nicotine [107] dependence has not been associated with this impairment. Reversal impairments have however been demonstrated in PG for both reward [62, 102, 107] and loss [107] outcomes. Because PG demonstrates largely normal performance on the WCST [102], this may be due to either intrinsic differences in set shifting versus reversal learning (e.g., utilizing dorsolateral versus orbitofrontal substrates respectively) or the difference in motivational outcomes between the two tasks: the PRR uses monetary outcomes but the WCST does not [102]. Indeed, a recent meta review found an association between PG and perseveration on monetary tasks despite intact executive planning [108]. This presents an interesting distinction for the study of PG. Unlike for other addictions, cognitive tasks routinely employed in research often use the very object of addiction for PG: monetary rewards. If the PRR used cocaine cues or rewards for cocaine dependence or food rewards for BED, impairments in reversing may be more prevalent.

Perseveration during the PRR in problem gamblers is associated with reduced sensitivity to both reward and loss [107], and perseveration for reward in particular has also been demonstrated with the card playing task, wherein choices that were previously rewarded must be inhibited; PG continue playing longer despite a shift from rewarding outcomes to losses [109]. Thus, monetary reward seems to be influential in PG, and impairments of cognitive flexibility must be considered in terms of reward sensitivity particularly in this group.

While there are very few studies examining compulsivity using these tasks in CSB, evidence from semi-structured interviews also suggests compulsive traits in this group [89] but further studies are still required. Mirroring inconsistencies in SUD, it seems that these particular measures of compulsivity or flexible choice may not pick up a consistent or robust impairment in behavioral addictions, although impairments in general cognitive flexibility in BED and reward perseveration in PG are implied.

Conflict of Interest

Dr. Laurel Morris reports grants from MRC Doctoral Training Grant, outside the submitted work.

Dr. Valerie Voon declares that she has no conflict of interest.

Human and Animal Rights and Informed Consent

This article does not contain any studies with human or animal subjects performed by any of the authors.

This article is part of the Topical Collection on Addictions

Papers of particular interest, published recently, have been highlighted as: • Of importance •• Of major importance