This study identified a number of differences in dopaminergic function between Parkinson’s disease patients who developed pathological gambling while being treated with dopamine agonists and control Parkinson’s disease patients who were matched for age, disease severity, cognitive performance and amount of dopaminergic intake. Patients with pathological gambling had greater reductions in binding in the ventral striatum during gambling than did control patients. [11C] raclopride is sensitive to competition from endogenously released dopamine in response to drugs or tasks that induce dopamine release. In this study, the observed decreases in binding potential are, therefore, likely secondary to dopamine release in response to the gambling task. Reduction in binding potential was observed bilaterally in the striatum of Parkinson’s disease patients with pathological gambling; whereas controls exhibited decreased binding only in the left striatum, an asymmetry that could reflect differences between groups in the neural processing of the gambling. Another interesting observation was that at baseline Parkinson’s disease patients with pathological gambling also had, as measured during the control task, lower D2 receptor binding in ventral striatum than control Parkinson’s disease patients without pathological gambling.
These results are broadly consistent with previous functional imaging evaluations of both behavioural and chemical addictions in the general population. In normal subjects, the activity and biochemistry of the brain does not appear to differ qualitatively when rewards derived from drugs of abuse or from behavioural stimuli. Monetary and sexual stimuli, all elicit the same patterns of striatal activation as drugs of abuse (Koepp et al., 1998
; Knutson et al., 2001
; Erk et al., 2002
; Childress et al., 2008
). However, patients with chemical addiction release more dopamine in ventral striatal circuits in response to their drugs of abuse (Volkow et al., 2006
). The current study presents the first evidence for this phenomenon in pathological gambling and supports the categorization of this disorder within the spectrum of behavioural addictions. Some investigators have proposed that the increase in dopamine release seen in chemical addiction may reflect a sensitization of circuits that occurs when repeated exposure to the addictive stimulus bypasses normal mechanisms of habituation (Di Chiara and Bassareo, 2007
). In this context, the increased release of dopamine we have observed in patients with pathological gambling could be interpreted as reflecting either a priming effect of repeated exposure to the gambling stimulus itself or to dopamine agonists, or to premorbid hypersensitivity of the ventral striatal circuits in the pathological gambling population.
Our finding of lower baseline values of D2/D3 binding in ventral striatum of patients with pathological gambling could be interpreted to reflect increased DA release in the basal state or, equally, lower baseline levels of D2/D3 receptors. This observation is likewise consistent with prior studies showing that low dopamine binding may mediate vulnerability to addiction. In non-addicted subjects, below baseline measures of striatal dopaminergic receptor availability predicts liking for methylphenidate and low striatal dopaminergic receptor availability has likewise been demonstrated in subjects with morbid obesity due to overeating (Volkow et al., 2002a
). Results from studies in a number of animal models of addiction also support a role for low dopaminergic receptor availability mediating vulnerability to addiction (Nader et al., 2002
; Dalley et al., 2007
). Some authors have proposed the finding of low D2 receptor availability in addictions best fits within the framework of the ‘reward deficiency syndrome’, whereby a chronic hypo-dopaminergic state is proposed to render individuals vulnerable to addiction by triggering a drive for rewarding substances or behaviours to supplement deficient dopamine in reward circuits (Volkow et al., 2002b
The major findings we have demonstrated in this study localize to the ventral striatum, although we have also observed dopamine release in the dorsal striatum in response to the gambling task in both pathological gambling and control groups. This result is consistent with the known importance of ventral striatal dopamine release in both normal reward and formation of addictions. A growing body of work, however, has also highlighted the importance of the dorsal striatum in the maintenance of established addictions and in the phenomenon of craving (Volkow et al., 2006
). Increasingly a distinction between dopamine release in ventral and dorsal striatal circuits appears to be important in the addiction process, with possible implications for ICDs. Phasic release of dopamine from mesolimbic projections to the nucleus accumbens is thought to occur in response to errors in the predicted value of outcomes. The signal that dopamine encodes likely serves to continuously update the importance of any stimulus that predicts a reward and may thus be critical to the formation of addictions. But as addictions become progressively more ingrained and habitual, the locus of neural control appears to shift dorsally in the striatum, to regions conventionally associated with the performance and maintenance of habits (Chambers et al., 2003
; Porrino et al., 2004
; Everitt and Robbins, 2005
). Thus the shift from the initiation to the consolidation of addiction may reflect an equivalent shift from limbic, to associative and sensorimotor corticostriatal circuits (Porrino et al., 2004
One of the most salient observations relating to ICDs associated with dopaminergic medications is that they develop only in a subset of patients. This suggests an underlying susceptibility within the population that may be quite separate from the co-occurrence of additional diseases such as Parkinson’s disease—although a potential interaction between susceptibility to ICDs and comorbid conditions cannot be excluded. In this context, the emergence of ICDs on exposure to dopaminergic agents could be viewed as a form of pharmacological challenge, unmasking behavioural vulnerabilities latent in the general population. In support of this view, a number of cognitive and behavioural traits such as impulsivity and novelty seeking have been shown to differentiate patients with ICDs, in both the general population and in those patients with comorbid Parkinson’s disease (Blaszczynski et al., 1997
; Jentsch and Taylor, 1999
; Evans et al., 2005
; Potenza, 2006
; Voon et al., 2007
). Such differences may point to underlying genetic variability conferring risk for ICDs, and indeed a number of allelic associations with both pathological gambling and drug addictions have been reported (Eisen et al., 1998
; Potenza, 2005
Mechanisms underlying the induction of ICDs by dopaminergic medications could potentially operate on multiple levels; by preferentially activating dopamine receptor subtypes, sensitizing receptors or by providing excessive dopaminergic stimulation to circuits regulating individual components of the addiction process, including motivation, reward, habit formation, and impulse control (Lawrence et al., 2003
). The potential for dopaminergic agents to sensitize striatal circuits may be particularly significant in the context of the increased striatal release of dopamine we have observed in our study.
The extensive literature describes the ability of dopamine and dopaminergic agents to prime or increase subsequent dopamine release in the ventral striatum. This phenomenon appears associated with an increase in salience or desirability of these drugs that manifests as craving (Robinson and Berridge, 2000
; Berke and Hyman, 2000
). Likewise Evans and colleagues (2006)
in an [11
C] raclopride PET evaluation of Parkinson’s disease patients with dopamine dysregulation have demonstrated that in response to a single dose of L-dopa, patients with dopamine dysregulation release more dopamine in ventral striatal circuits compared to control Parkinson’s disease patients. This hypersensitivity correlated with the self reported compulsive ‘wanting’ of the drug but not liking for it. Drugs with dopaminergic actions are known to cross-prime for the release of dopamine induced by other drugs within the same class, and to upregulate the incentive salience of a variety of stimuli with which they have become associated. In rodent models of addiction, for example, repeated exposure to psychostimulants, such as amphetamine, increases the salience for other rewards such as food and sexual stimuli (Engber et al., 1989
; Robinson and Berridge, 2000
; Nocjar and Panksepp, 2002
; Robinson and Berridge, 2003
). Likewise in problem gamblers, amphetamine has been shown to increase motivation for gambling in a manner that is predicted by the severity of the subject’s gambling (Zack and Poulos, 2004
Parallels between this phenomenon and the development of behavioural addictions in patients treated with dopaminergic agents are self evident. A prominent feature of all the ICDs relating to use of dopaminergic agents—compulsive gambling, compulsive shopping and hypersexuality—is an intense craving for the behaviours consistent with increased incentive salience (Robinson and Berridge, 2000
). Patients often report being driven to perform their behaviours in spite of full awareness of the adverse consequences of doing so. In this context, the dopaminergic stimulation provided by dopamine agonists may mimic the sensitizing effects of excess dopamine in ventral striatal circuits during performance of rewarding behaviours and so triggering a form of behavioural addiction.
Another interesting observation of our study was that changes in binding encompassed different areas of the ventral striatum. According to previous imaging reports (Mawlawi et al., 2001
; Martinez et al., 2003
), these areas of the striatum are part of both limbic and associative circuits. This is not surprising, if we consider that gambling behaviour, besides engaging the reward system, requires also significant cognitive strategies often impaired in these patients (Volkow et al., 2004
; Baler and Volkow, 2006
). In fact gamblers, including those in this study, report the tendency to calculate odds on the basis of cards already revealed.
In contrast to other functional imaging studies (Evans et al., 2006
; Reuter et al
., 2006), we did not observe any correlation between striatal findings and severity of gambling in Parkinson’s disease patients with pathological gambling. It is possible that limitations intrinsic to our gambling task—which did not allow us to dissect different behavioural aspects (e.g. perseveration) or individual strategies—may have reduced the sensitivity of the analysis.
Reduction in binding potential was observed bilaterally in the striatum of Parkinson’s disease patients with pathological gambling; whereas controls exhibited decreased binding only in the left striatum. While such asymmetry could reflect differences between groups in the neural processing of the gambling, it is difficult to speculate whether this observation could be related to the underlying disease or be a feature of the reward network or combination of both. However, it is worth noting that similar changes in the left ventral striatum have also been reported in functional imaging studies performed in healthy subjects during performance of reward tasks (Zald et al., 2004
; Riba et al., 2008
While we acknowledge the fact that due to the blurred anatomical boundaries between dorsal and ventral striatal regions some of the radioactivity measured in the ventral striatum may result from contamination from counts originating in the adjacent dorsal striatum (Mawlawi et al., 2001
), we believe that our measures taken to prevent head motion along with the applied motion correction have minimized the contribution of this factor. In addition, the striking difference in dopamine release observed in our study between dorsal and ventral striatum favours and supports the significant functional differences between these two dopaminergic regions.
Although the pathogenesis for the changes in dopaminergic function observed in our Parkinson’s disease patients with pathological gambling remains to be elucidated, the similarities between the findings in these patients and those with chemical addictions supports the classification of pathological gambling within the spectrum of behavioural addiction. Our findings may point to an underlying vulnerability in the general population that may provide insight into the pathophysiology of this and other ICDs and may suggest Parkinson’s disease as a potential model of dopaminergic disregulation to elucidate behaviour addiction.