Dopamine agonists and pathological gambling

This week we have a blog from Dr Carinne Piekema covering the potential role of dopamine agonists in the development of pathological gambling. Carinne is a science journalist and neuroscientist who has worked on executive function and memory in neuroscience laboratories in the Netherlands and the UK. She is currently finishing a Masters in Science Media Production at Imperial College London and is particularly interested in communicating science through radio and podcasting.

Dopamine agonists have been used to alleviate the motor symptoms in Parkinson’s disease (PD) for several decades, both alongside levodopa and – as evidenced by the large number of publications, particularly in the JNNP – as a treatment in its own right (e.g. Parkes et al., 1976; Lees & Stern 1981; Montastruc et al., 1989).

Over the past decade, there have been increasing numbers of reports describing problematic side effects of such dopamine agonists (e.g. Seedat et al., 2000). In particular, a growing body of research shows that dopamine agonist medication may cause pathological gambling, compulsive and impulsive shopping, compulsive eating and hypersexuality in approximately 13.6% of PD patients who receive this treatment, thus developing the clinical symptoms of Impulse Control Disorders (ICD).

While such risk-taking and impulsive behaviours have been extensively documented, the reason why they arise in a particular subgroup of patients is still not fully understood. This is perhaps not surprising when considering the complexity of decision-making processes that might give rise to such risk-taking, impulsive behaviours. Before we choose one option over another, we weigh up the costs and benefits of all the options and calculate the likelihood of positive and negative consequences of our potential choice based on our past experience, and dopamine agonists may bias any of these processes.

I was therefore interested to read two papers published in the last few months that have started to shed light on the causes of the increase in risky decisions in these patients (Claassen et al., 2011; Voon et al., 2011). In both studies, PD patients were tested on varieties of risk-based tasks, both on- and off-dopamine agonist medication, in order to probe, in a controlled manner, the aspects of choice behaviour that might be affected by this type of treatment.

Daniel Claassen and his international team reported in Behavioral Neuroscience how PD patients with and without ICD performed on a paradigm called the Balloon Analogue Risk Task.  In this task, participants watch a balloon on a computer screen incrementally filling with air; with every puff of air the value of the balloon is increased by a small amount of money, which they can gain at any time by choosing to pop the balloon. The catch is that, if the balloon pops by accident, all the money in the balloon is lost. Therefore, the participants have to weigh up the cost of an accidental pop against the potential gain from a fuller balloon.

Claasen found that both groups of PD patients, while off-medication, showed a similar pattern of behaviour concerning when to pop the balloon, suggesting that there was no difference in risk-taking behaviour between both groups when they were not taking dopamine agonists. However, when taking dopamine agonist medication, the ICD patients showed an increased tendency compared to those patients without ICD to try to gain more money by letting the balloon inflate further. Moreover, across all PD patients, those who were on the highest doses of dopamine agonist medication were more likely to engage in risky behaviour compared to those who were taking a relatively low dose. Importantly, the researchers found no evidence that ICD patients had problems adjusting their behaviour following negative outcomes, suggesting that these patients are able to respond appropriately to feedback. Thus, it seems that ICD patients’ risk taking is not caused by an inability to learn from mistakes, but instead may result from changes in the way in which each decision is weighed up.

A similar conclusion was reached in a study published in Brain by Valerie Voon and her colleagues in the UK and US. They used a classic gambling task where participants chose between a probabilistic “risky” option and a certain “safe” option. In one condition, the participants were gambling over gains in money (i.e., a choice between a certain small reward or a risky option which could either lead to a high monetary reward or nothing at all) and in the other condition over losses in money (a choice between a certain small loss of money or a risky option which could either lead to a no loss or a large loss of money). In agreement with Claassen and colleagues, the researchers observed that PD patients with ICD who were taking dopamine agonist medication took riskier gambles in the ‘gains’ compared to PD patients without ICD. Nonetheless, these patients with ICD were also still sensitive to the relative level of the risk and chose to gamble less when gambles were more risky and adjusted their behaviour on the trials directly after a loss.

Voon and colleagues also obtained functional neuroimaging data from the patients while they were performing the task in order to look at the underlying neurobiological substrates of these changes in behaviour. They found that PD patients with ICD showed significantly less activity in orbitofrontal and anterior cingulate cortices compared to those without ICD. These regions are known to be involved in the evaluation of risks and value of option and a decreased activity may result in a tendency to behave more riskily.

PD has long been thought of primarily as a disease of the motor system caused by the loss of dopamine-containing cells of the substantia nigra. Such findings have resulted in the idea that the nigrostriatal dopamine system is primarily involved in the selection and control of voluntary movement. However, largely separate from this research, other scientists have been studying another dopamine system – the mesocorticolimbic dopamine system – and have focused mainly on its role in motivational and reward processing. These two strands of dopamine research have largely proceeded in parallel, partly based on an idea that these two dopamine systems were completely distinct (Wise, 2009).

However, these changes in decision-making in subgroups of PD patients documented by Voon, Claassen and other researchers demonstrate that considering PD treatment from the sole perspective of trying to re-equilibrate the loss of nigrostriatal dopamine may have drawbacks.  First, given that PD initially primarily affects the nigrostriatal dopamine system while sparing mesocorticolimbic dopamine, dopamine agonists might in fact impair the functionality of the latter system in certain patients, leading to maladaptive decision-making. It may be that PD patients who develop ICD may have underlying genetic neurobiological vulnerabilities that exacerbate such an impairment, similar to the link between impulsive genetic personality traits and drug taking (Dalley et al., 2011).

Second, several lines of basic research suggest that the different dopamine systems interact both anatomically and functionally. Both systems send projections to parts of the frontal lobe as well as the dorsal striatum (e.g. Fallon, 1988), and the cell bodies in both systems can sometimes be found in the same midbrain structure (Wise, 2009). There is also increasing evidence that interactions between ventral striatum – innervated by mesolimbic dopamine – and dorsolateral striatum – innervated primarily by nigrostriatal dopamine – underlie the development of habitual addictive behaviours, and that these interactions are mediated by dopamine transmission (Belin & Everitt, 2008).

But possibly the most compelling reason for why the strong distinction between “motor” and “motivational” dopamine systems may be regarded as artificial can be found in the fact that movement and motivation can not always be so neatly separated: sometimes we choose to walk to the fridge because we are hungry, to scratch our back to relief an itch, to exercise to stay fit. In a study published a few years ago in the Journal of Neuroscience, Pietro Mazzoni and his colleagues showed that one of the key “motor” symptoms of PD – namely bradykinesia – might arise not as the patients could not make the movements, but because they were less motivated to expend energy on movement (Mazzoni et al., 2007).  This indicates that nigrostriatal dopamine activity plays a role in enabling voluntary movement not just through modulations of motor pathways in the basal ganglia, but also by participating in calculations of the cost/benefit value of a response. In other words, there is an important class of voluntary movements, which are exactly the type that are frequently affected in PD and rely on dopamine, that are motivationally driven.

Studies such as those led by Claassen and Voon are gradually starting to reveal the mechanisms behind pathological gambling and other compulsive behaviours caused by dopamine agonist medication in PD patients with ICD, and in doing so, they aid our understanding of the role of dopamine in risk-taking behaviour. Even as researchers are trying to uncover the exact functioning of such complex systems, it is important for clinicians to use the knowledge that certain interactions between medication and underlying vulnerabilities can lead to highly undesirable side effects.


Claassen DO, van den Wildenberg WP, Ridderinkhof KR, Jessup CK, Harrison MB, Wooten GF, Wylie SA (2011). The risky business of dopamine agonists in Parkinson disease and impulse control disorders. Behav Neurosci DOI: 10.1073/a0023795

Voon V, Gao J, Brezing C, Symmonds M, Ekanayake V, Fernandez H, Dolan RJ, Hallett M. (2011). Dopamine agonists and risk: impulse control disorders in Parkinson’s disease. Brain 134: 1438-1446.



Belin D, Everitt BJ (2008). Cocaine seeking habits depend upon dopamine-dependent serial connectivity linking the ventral with the dorsal striatum. Neuron 57: 432-441.

Dalley JW, Everitt BJ, Robbins TW (2011). Impulsivity, compulsivity, and top-down cognitive control. Neuron 69: 680-694.

Fallon JH  (1988). Topographic organization of ascending dopaminergic projections. Ann N Y Acad Sci 537: 1-9.

Lees AJ, Stern GM (1981). Sustained bromocriptine therapy in previously untreated patients with Parkinson’s disease. J Neurol Neurosurg Psychiatry 44: 1020-1023.

Mazzoni P, Hristova A, Krakauer JW (2007). Why don’t we move faster? Parkinson’s disease, movement vigor, and implicit motivation. J Neurosci 27:7105-7116.

Montastruc JL, Rascol O, Rascol A (1989). A randomised controlled study of bromocriptine versus levodopa in previously untreated Parkinsonian patients: a 3 year follow-up. J Neurol Neurosurg Psychiatry 52:773-775.

Parkes JD, Debono AG, Marsden CD (1976). Bromocriptine in Parkinsonism: long-term treatment, dose response, and comparison with levodopa. J Neurol Neurosurg Psychiatry 39: 1101-1108.

Seedat S, Kesler S, Niehaus DJ, Stein DJ  (2000). Pathological gambling behaviour: emergence secondary to treatment of Parkinson’s disease with dopaminergic agents. Depress Anxiety 11: 185-186.

Wise RA (2009). Roles for nigrostriatal – not just mesocorticolimbic – dopamine in reward and addiction. Trends Neurosci 32: 517-524.

(Visited 1,191 times, 1 visits today)