Clinical Implications

• • The available evidence suggests that substance-related attentional bias in humans may be modulated by dopamine. However, clinical trials are needed to investigate the clinical implications of this knowledge.

• • There is currently not sufficient evidence for pharmacological agents to be used in clinical practice to reduce attentional bias.

• • A focus on neurotransmitter systems such as acetylcholine and noradrenaline could provide new insights regarding the pharmacology of substance-related attentional bias that may eventually be of clinical relevance.

Theoretical Background and Dopaminergic Manipulations

There is a general consensus that the dopaminergic system, with projections from the ventral tegmental area (VTA) to the striatum, the anterior cingulate cortex (ACC), and other prefrontal brain regions, is responsible for reinforcement learning and experiencing reward.Reference Schultz, Dayan and Montague¹⁶ Based on preclinical animal work, it has been suggested that this dopaminergic reinforcement learning system is also involved in the development of attentional bias via conditioned reinforcement learning.Reference Robinson and Berridge¹⁷ After repeated drug intake, substance-related stimuli become conditioned stimuli and elicit phasic dopaminergic activity,Reference Volkow, Wang and Telang^18–Reference Zijlstra, Booij, Van den Brink and Franken²⁰ thereby signaling the expectation of a future reward (ie, the intake of the abused substance). Gradually, the dopaminergic system becomes sensitized for substance-related cues so that they become extremely salient; these cues become the focus of attention, and they elicit behaviors such as drug seeking and consumption.Reference Phillips, Stuber, Heien, Wightman and Carelli^21, Reference Robinson and Berridge²² Functional magnetic resonance imaging (fMRI) studies have indeed shown that attentional bias in substance-dependent patients is associated with activation in the ventral striatum, the ACC, and other limbic and prefrontal brain regions.Reference Luijten, Veltman and Van den Brink^23–Reference Vollstadt-Klein, Loeber and Richter²⁶ In contrast to the phasic dopaminergic responses to substance-related cues, the striatal dopamine system in substance-dependent patients is supposed to be generally blunted compared to healthy controls,Reference Blum, Braverman and Holder²⁷ which is an effect that has even been observed for the acute intake of substances of abuse.Reference Narendran and Martinez²⁸ Based on these theoretical accounts, the most straightforward pharmacological manipulation to reduce attentional bias would be to prevent the phasic dopaminergic burst associated with conditioned substance-related cues. This blockade of the dopaminergic response would then lead to a reduction in substance-related attentional bias. Franken etal Reference Franken, Hendriks, Stam and Van den Brink¹⁴ were the first to test this dopaminergic attentional bias hypothesis in humans. In this study, heroin-dependent patients performed the heroin word Stroop task in a placebo-controlled, double-blind, randomized crossover study using a single dose of haloperidol (dopamine D2/D3 receptor antagonist) to modulate dopaminergic responses to heroin-related words. The findings of this study provided support for the dopaminergic theory of attentional bias in humans, as attentional bias for heroin words was eliminated after haloperidol was administered. These results were partly replicated for smoking-related stimuli using a smoking word Stroop task in two subsequent studies in smokers.Reference Munafo, Mannie, Cowen, Harmer and McTavish^11, Reference Hitsman, MacKillop and Lingford-Hughes¹² In both these studies, an amino acid mixture that lacks tyrosine (TYR) and its precursor phenylalanine (PHE) was used to modulate dopamine responses to smoking-related stimuli. As the synthesis of dopamine in the brain is dependent on the availability of these amino acids from plasma, the acute administration of a TYR/PHE-free amino acid mixture has been shown to reduce dopaminergic activity in the brain.Reference McTavish, Cowen and Sharp²⁹ The findings of the first study employing this methodReference Munafo, Mannie, Cowen, Harmer and McTavish¹¹ showed a gender-specific reduction in smoking-related attentional bias after the TYR/PHE-free amino acid mixture was administered. Stroop interference scores (ie, reaction times for smoking-related versus neutral words) were reduced relative to the placebo condition in women but not in men. The second study using the same study design,Reference Hitsman, MacKillop and Lingford-Hughes¹² however, showed a reduction in smoking-related attentional bias among both smoking men and women (no gender-specific analyses were performed). These findings suggest that attentional bias may indeed be reduced in smokers, perhaps more strongly in women, when dopaminergic transmission is prevented during the exposure to smoking related cues.

Studies measuring brain activation concurrently with behavioral measures of attentional bias could test the dopaminergic hypothesis of attentional bias in more detail, as brain regions involved in salience detection and attentional bias can be directly evaluated. A pictorial line counting attentional bias task was used in a recent placebo-controlled, double-blind, randomized crossover pharmacological fMRI study in smokers.Reference Luijten, Veltman and Hester⁸ The dopamine challenge consisted of a single administration of the dopamine D2/D3 receptor antagonist haloperidol. While haloperidol did not alter behavioral measures of attentional bias for smoking-related stimuli, brain activation in the dorsal ACC and dorsolateral prefrontal cortex (DLPFC), which are associated with attentional bias, was reduced in smokers by haloperidol. That is, no differences in brain activation between smokers and non-smokers were found after haloperidol intake. As the dorsal ACC is involved in the salience detection network, as well as in top-down control of attention in cooperation with the DLPFC, these findings suggest that a reduction in dopamine may affect the salience detection of smoking-related stimuli in smokers, as well as top-down control over these cues to continue ongoing behavior.

A placebo-controlled, double-blind, randomized crossover pharmacological fMRI study in stimulant-dependent patients measured brain activation associated with the performance of a stimulant word Stroop task under several dopaminergic conditions.Reference Ersche, Bullmore and Craig⁹ This study involved both a single challenge of a D2/D3 dopamine antagonist (amisulpride) as well as a D2/D3 dopamine agonist (pramipexole dihydrochloride). While no effects were found of both medications on attentional bias–related brain activation or behavioral measures for the total stimulant-dependent patients group, compulsivity levels within the patient group appeared to modulate due to medication effects. On a behavioral level, attentional bias was reduced in low-compulsive stimulant-dependent patients after the dopamine agonist, whereas the high-compulsive stimulant-dependent patients still showed an attentional bias. In line with behavioral findings, the dopamine agonist pramipexole enhanced attentional bias–related brain activation in the left ventral prefrontal cortex and the cerebellum in high-compulsive stimulant-dependent patients, whereas it reduced activation in these regions in low-compulsive stimulant users. These findings suggest that individual differences in compulsivity levels may modulate the effects of a dopamine agonist for behavioral attentional bias measures as well as attentional bias–related brain activation. Compulsive behavior has previously been associated with striatal dopamine transmission.Reference Evans, Pavese and Lawrence³⁰ More specifically, it has been suggested that optimal dopamine levels for cognitive functioning follow an inverted U-shaped curved depending on personality traits such as compulsivity.Reference Cools³¹ In contrast to the study in smokers, attentional bias–related brain activation was not reduced following the administration of the dopamine antagonist amisulpride in either low- or high-compulsive stimulant-dependent patients.

Goldstein etal Reference Goldstein, Woicik and Maloney¹⁰ performed a placebo-controlled, single-blind, counterbalanced pharmacological fMRI study using a crossover design in cocaine-dependent patients using methylphenidate (a dopamine transporter blocker) as the pharmacological agent. Instead of investigating whether a dopamine antagonist can ameliorate attentional bias and related brain activation, this study implemented a different approach, as methylphenidate is known to increase synaptic dopamine levels. Findings of this study revealed that cocaine-dependent patients displayed reduced brain activation in the dorsal ACC for cocaine-related words relative to controls after placebo, whereas this hypoactivation was normalized after methylphenidate administration. No medication effects were found for behavioral measures. These findings may suggest that methylphenidate could be beneficial for brain activation associated with attentional bias. While these findings seem to contradict the dopaminergic theory of attentional bias, they can be explained by the known cognitive enhancing properties of methylphenidate.Reference Peterson, Potenza and Wang³² Current addiction models suggest that substance abuse is not only caused by motivational processes such as attentional bias, but that the combination of an overactive motivational system with reduced cognitive brain functions drives the continuation of addictive behaviors.Reference Field and Cox^4, Reference Volkow, Fowler and Wang^33, Reference Wiers, Bartholow and Van den Wildenberg³⁴ More specifically, it has been proposed that the ineffective prefrontal cognitive control circuit in substance-dependent patientsReference Goldstein and Volkow³⁵ influences the strength of attentional bias in such a way that reduced cognitive control results in enhanced attentional bias.Reference Field and Cox⁴ This interaction between cognitive control–related brain functions and attentional bias provides a second plausible working mechanism to ameliorate attentional bias using pharmacological agents. Cognitive enhancing medications such as methylphenidate may therefore result in indirect beneficial effects on attentional bias–related brain activation. In line with this hypothesis, modafinil, another cognitive-enhancing medication, has been shown to have effects on both cognitive control as well on motivational aspects of cocaine dependence.Reference Goudriaan, Veltman, van den Brink, Dom and Schmaal^36, Reference Schmaal, Joos and Koeleman³⁷

Overall, studies using dopaminergic challenges to investigate the effects on attentional bias are promising and provide, to a certain extent, support for the dopamine theory of attentional bias. Nevertheless, the results are not always consistent, and future research is needed to further clarify the role of modulatory factors such as gender and compulsivity levels. Especially, large methodological differences such as different substance-dependent groups, dopaminergic challenges, and attentional bias paradigms in the pharmacological fMRI studies may have contributed to inconsistencies in results between these studies. In addition, all studies until now have investigated acute effects of pharmacological agents, so it is currently unknown whether chronic pharmacological treatments may reduce attentional bias in the long term.

Targeting Other Neurotransmitters to Ameliorate Attentional Bias

Only one study was identified that investigated an alternative neurotransmitter system to reduce attentional bias. Kamboj etal Reference Kamboj, Joye and Das¹³ investigated the effect of D-cycloserine (DCS) on attentional bias, which is a partial agonist at the glycine site of the N-methyl-D-aspartate (NMDA) receptor. The rationale for this is based in preclinical work that has suggested that DCS can have extinction-enhancing properties during cue exposure therapy.Reference Myers and Carlezon³⁸ Smokers in this between-group, double-blind, placebo-controlled study received two cue-extinction sessions combined with either DCS or placebo. Attentional bias was measured using a dot-probe task before and after the two cue-extinction sessions. The results did not show a beneficial effect of DCS or cue-extinction in general for smoking-related attentional bias, suggesting that DCS combined with two sessions of cue-extinction may not be a promising tool to reduce attentional bias in smokers. Despite these null results, it may be worth exploring the involvement of alternative neurotransmitters in substance-related attentional bias. Acetylcholine, for example, is known for its modulating effects on attention,Reference Sarter, Givens and Bruno³⁹ and a link has been proposed between the dopaminergic theory of attentional bias and the acetylcholine system.Reference Bushnell, Levin and Marrocco⁴⁰ For example, increases in dopamine in the ventral striatum have been suggested to be associated with increases in cortical acetylcholine release,Reference Moore, Fadel, Sarter and Bruno⁴¹ such that repeated administration of substances of abuse may also sensitize cortical acetylcholine via projections with the basal forebrain.Reference Bushnell, Levin and Marrocco⁴⁰ It would be worth investigating whether such a sensitization of the cortical acetylcholine system represents an integral component of attentional bias.

Another neurotransmitter system that may be worth investigating in the context of attentional bias is the noradrenaline system. Noradrenaline is mostly released from the locus coeruleus (LC), which is a nucleus in the brainstem that has strong reciprocal connections with the prefrontal cortex and the ACC. Preclinical research has shown that both tonic and phasic LC discharge activity is closely related to the overall salience and arousing properties of the presented stimuli.Reference Berridge and Waterhouse⁴² This may suggest that the noradrenaline system could be involved in the detection of salient substance-related stimuli in the environment, probably in cooperation with the dopamine system. Atomoxetine is a noradrenaline reuptake inhibitor that increases noradrenaline levels in the prefrontal cortex including the ACC,Reference Bymaster, Katner and Nelson⁴³ and may be a useful tool to investigate noradrenergic regulation of attentional bias in substance-dependent patients. Atomoxetine has been shown to improve cognitive functions and boost ACC activation in healthy controls,Reference Graf, Abler and Freudenmann^44, Reference Hester, Nandam and O'Connell⁴⁵ so it may also have a beneficial effect on cognitive functioning in substance-dependent patients, and therefore may indirectly reduce attentional bias via improving cognitive control similarly to the way that has been proposed for methylphenidate. In conclusion, a focus on other neurotransmitter systems such as acetylcholine and noradrenaline could shed some new light on the pharmacology of attentional bias.

Treatment Implications

In summary, some of the reviewed pharmacological attentional bias studies in substance-dependent individuals show evidence that attentional bias and related brain activation can be modulated by dopaminergic agents, although the results are not always consistent. In addition, it may be that cognitive enhancing agents will be able to indirectly reduce attentional bias via boosting cognitive control. However, we should be very careful with generalizing these positive attentional bias–related findings for particular drugs to beneficial treatment effects for several reasons. First, it is unknown whether chronic treatment with dopaminergic agents will result in a long-term reduction in attentional bias, and whether this in turn will improve clinical outcome measures such as relapse or reduced substance use. There is one study that showed that attentional bias was reduced in opiate-dependent patients after chronic treatment with methadone or buprenorphine as compared to current usersReference Gardini, Caffarra and Venneri⁴⁶; however, the same effect was found in a group of patients receiving nonpharmacological treatment, suggesting that prolonged abstinence in general may have caused the reduction in attentional bias. As the study was not double-blind and did not include a placebo, it is particularly difficult to attribute the reduction in attentional bias to the chronic pharmacological treatment. Further studies are needed to show if chronic pharmacological treatment will lead to a reduction in attentional bias.

Clinical application of positive findings for certain dopaminergic agents to reduce attentional bias is further complicated by the fact that attentional bias is not the only dopaminergic-mediated process involved in addiction. While attentional bias and craving are supposed to have mutual reinforcing associations in theoretical models,Reference Franken^3, Reference Field and Cox⁴ none of the studies discussed in this review found beneficial effects of the dopamine challenges on subjective cravings. This may be explained by the original concept of the dopaminergic incentive sensitization theory of attentional bias, implying that dopamine release following exposure of substance-related cues may occur independent of neural mechanisms that mediate withdrawal and related withdrawal-induced cravings.Reference Munafo, Mannie, Cowen, Harmer and McTavish^11, Reference Robinson and Berridge¹⁷ While a phasic dopaminergic response in response to conditioned substance-related cues may trigger the enhanced attentional processing of these substance-related cues, other dopaminergic deficiencies such as tonic dopamine depletion may induce withdrawal-related craving and can motivate drug-seeking behavior.Reference Koob, Ahmed and Boutrel⁴⁷ Besides attentional bias and craving, there is also accumulating evidence that prefrontal dopamine levels can modulate cognitive control–related brain functions. An inverted U-curve theory of dopamine and cognitive control suggests that there is an optimal prefrontal dopamine level for cognitive performances, and that either too low or too high levels may reduce cognitive controlReference Cools³¹; this is further confirmed by findings that the effects of dopaminergic agents on cognitive functioning are modulated by dopaminergic genotypes.Reference Forbes, Brown and Kimak⁴⁸ In this context, it is interesting to mention that the same group of smokers in which normalized attentional bias–related brain activation was found after the dopamine antagonist haloperidol was administered also performed a Go/NoGo task to measure inhibitory control under placebo and dopamine-deprived conditions.Reference Luijten, Veltman and Hester⁸ In contrast to the beneficial effects of the dopamine antagonist on attentional bias–related brain activation, reduced inhibitory control and prefrontal brain function was found after haloperidol relative to the placebo condition.Reference Luijten, Veltman and Hester⁴⁹ Besides the unpleasant side effects, these differential effects of dopaminergic agents on different processes that are all involved in the continuation of substance-related behaviors may be one of the reasons why dopamine antagonists do not seem to be successful in the treatment of cocaine dependence.Reference Amato, Minozzi, Pani and Davoli⁵⁰ Future research therefore faces the challenge to investigate the effects of pharmacological agents on the multiple processes involved in addictive behaviors. Pharmacological agents that may be able to restore the balance between motivational processes and controlling processes may eventually succeed to more effectively treat substance dependence.

Besides pharmacological interventions, attentional bias modification training may provide an alternative way to reduce attentional bias in addiction. While attentional bias modification (ABM) training could probably constitute a treatment obstacle for some substance-dependent patients, it is worth investigating nonpharmacological interventions to reduce substance-related attentional bias. While some of the initial proof-of-concept studies yielded promising findings, others did not,Reference Field, Duka, Tyler and Schoenmakers^51–Reference Fadardi and Cox⁵⁷ and evaluations of ABM in substance abusers who are motivated to change their behavior have either been underpowered,Reference Schoenmakers, De Bruin and Lux⁵³ or have failed to include appropriate control conditions.Reference Fadardi and Cox⁵⁷

Disclosures

The authors do not have anything to disclose.