Cocaine, Loss, and the Liberal View of Addiction

A study published in the Journal of Neuroscience (behind a paywall here, but see also a media report here a a few days ago suggests that cocaine addicts may have impairments in the neural circuits that are responsible for the prediction of emotional loss. In this post, I shall consider the implications that this finding might have for our understanding of addiction and the autonomy of addicted individuals.

 

A Summary of the Study

 

The study set out to measure dopaminergic Reward Prediction Error (RPE) in addicted individuals compared to healthy subjects. RPE signals are commonly understood to reflect a discrepancy between a predicted and an actual rewarding outcome; these signals are then used to update the expected value of the stimulus that led to the rewarding outcome. As such, RPE is often deemed to be central to human learning; after performing some act (or experiencing some stimulus), the extent to which we will seek to repeat or avoid that behaviour in future will depend, it is claimed, on whether our previous experience led to an increase in dopaminergic signalling (i.e. if we experienced an unpredicted reward), or a decrease (i.e. if we experience a negative outcome, or the omission of an expected positive reward).

Briefly, the experiment in the recent study involved subjects playing a gambling game in which they would predict whether they would win or lose money in each individual trial, given one of three known win probabilities (25%, 50%, or 75% chance). Whilst playing the game, subjects were monitored using Electroencephalography (EEG), allowing researchers to record electrical activity in the subject’s brain. The study built on past research by showing that, in comparison to 25 healthy control subjects, 75 subjects addicted to cocaine demonstrated impaired prediction signalling, since these subjects failed to trigger negative RPE signals in response to worse-than-expected outcomes. Moreover, the study found that addicted individuals who had abstained from cocaine use in the 72 hours preceding the experiment had an impaired positive RPE in response to unexpected rewards. However, addicts who had taken cocaine in the 72 hours preceding the experiment did not suffer from this impairment, lending credence to the hypothesis that, in addiction, the drug of choice can normalize particular brain functions.

 

Three Accounts of Addiction

 

To consider the some of the possible implications of this study, it is useful to consider three prominent accounts of addiction. I shall briefly outline three accounts analysed by Bennett Foddy and Julian Savulescu in an influential paper, before going on to consider the implications of the above experimental findings for these accounts.

Consider first what Foddy and Savulescu term “The Disease View”. On this view, addiction is understood to result from a direct physiological change in the brain caused by chronic drug use. On the disease view, the addict is understood to lack autonomy with respect to their drug-taking behaviour, because such behaviour is understood to be the result of a fundamentally non-rational process, in so far as it is a direct result of a brain adaption. Alternatively, according to “The Willpower View”, addicts are understood to lack control over their actions because they lack the willpower to resist their powerful impulsive cravings for drugs, even though they may wish to abstain. Foddy and Savulescu note that the disease view is normally based on neurobiological evidence, whilst the willpower view is normally based on phenomenology and self-reports; however, on both of these views, the addict is understood to lack self control with respect to their drug-taking behaviour.

In contrast, Foddy and Savulescu propose a liberal account of addiction that makes only three central claims about addiction:

 

First, we do not know whether an addict values anything more than the satisfaction of his addictive desires. Second, we do not know whether an addict behaves autonomously when they use drugs. Third, addictive desires are just strong, regular appetitive desires. [1]

 

In support of this view, Foddy and Savulescu claim that any sort of strong reward can result in the same kind (if not degree) of neurobiological change that occurs following drug use. As such the evidence of neurobiological changes often cited in support of the Disease view is not sufficient for establishing the truth of that account; such changes are not uniquely caused by drug-taking. Furthermore, they also point out that the empirical evidence suggests that addicts often stop taking drugs, and they often do so in response to countervailing reasons not to do so. Accordingly, they conclude that there is no basis for concluding that all addicts lack autonomy with respect to their addictive behaviour, as the Disease View maintains.

 

Implication of the New Study

 

It might be claimed that the study reported above lends support to the disease view, insofar as it provides evidence of physiological changes that have seemingly been caused by drug use. However, this might be too quick. First, whilst the study established that cocaine users have impaired loss prediction signalling, it did not establish that this impairment was caused by their drug use. The results are compatible with hypothesis that individuals are more likely to start taking and become addicted to drugs if they already suffer from this sort of impairment. However, the difference between addicted users who had taken cocaine in the 72 hours preceding the experiment and those who had not with regard to positive RPE signals suggest that taking the drug influences positive RPE amongst addicted populations.

Perhaps more importantly though, even if we assume that the drug use caused the impairment in loss prediction signalling, it is not clear that addictive pharmacological agents are unique in causing these sorts of effects. As I mentioned above, one of the arguments that Foddy and Savulescu use in favour of their view is that various sorts of pleasurable behaviours can have profound effects on dopamine activation. As they point out:

 

When we indulge in a rewarding activity, the reward mechanisms of our brains reinforce the neural pathways that led us to seek the activity in the first place. [2]

 

As such, if the above study is to be understood show that cocaine causes a neurological change that distinguishes the addict’s craving for cocaine as an uncontrollable addictive desire that differs in kind to other sorts of desire in the way that the Disease View claims, the study would need to show that other strong pleasure oriented desires do not also cause a comparable impairment. However, in the above study, we do not know anything about the propensities of the control ‘healthy’ subjects towards engaging in repetitive pleasure orientated consumption of a particular stimulus, such as caffeine, nicotine, or sugar to name a few examples. If the study had included a population of subjects that engaged in such repetitive behaviours, and if such subjects had not shown a comparable impairment in predictive loss signalling to the subjects who were cocaine addicts, then this would have been a damning blow against Foddy and Savulescu’s liberal view. However, if Foddy and Savulescu are correct in their analysis, it is feasible that individuals who frequently consume high doses of such pleasurable stimuli would be likely to have similar impairments.

As such, it is not clear that the findings of the study reported above have the sort of normative force that defenders of the disease view would need in order to say that the study disproves the liberal view. If the study is understood as demonstrating that cocaine users have impaired autonomy, we do not yet have the evidence to establish that non-pharmacological substances do not cause comparable deficits to predictive loss signalling, and (a fortiori), to the consumer’s autonomy.

[1] Bennett Foddy and Julian Savulescu, “A Liberal Account of Addiction,” Philosophy, Psychiatry, and Psychology 17, no. 1 (2010): 14.

[2] Ibid.,  6.