Research on the neurobiology of addiction often involves nonhuman animals that are given ready access to drugs for self-administration but without other choices. Here we argue using cocaine as an example that this standard setting may no longer be sufficient and can even lead to the formulation of unrealistic views about the neurobiology of addiction.
Addiction as a psychiatric disorder is defined as resulting from brain dysfunctions that affect normal choice-making, not as an expectable response to lack of alternative choices. We encourage neurobiologists involved in addiction research to increase animals’ choice during drug access, preferably by supplying alternative rewarding pursuits.
Only animals that continue to take and prefer drugs despite and at the expense of other available choices may be considered as having developed an addiction-like behavior in comparison to those that remain able to stop drug use for other pursuits, even after extended drug use. The systematic comparison of these two individual behaviors should reveal new insights about the neurobiology of drug choice and addiction. More generally, this research should also shed a unique light on how the brain ‘chooses’ among qualitatively different kinds of pursuits.
[...]
Normally, midbrain dopamine neurons fire in response to unexpected natural rewards (e.g., sweet reward), thereby causing a sharp and transient increase in nucleus accumbens dopamine levels. This dopamine surge is hypothesized, among other possible functions, to reinforce behaviors or cues that initially led to the reward, presumably by shaping and strengthening its neuronal representation in the nucleus accumbens.
By mimicking an abnormally amplified version of this dopamine-based reinforcement signal, cocaine is thought to reinforce drug-seeking behavior more strongly than other activities, thereby setting the stage for the progression to addiction. Metaphorically, cocaine is often said to usurp, co-opt or hijack the normal functions of midbrain dopamine neurons and associated neurocircuits
Most animals take cocaine but few people become addictedMany variants of this drug usurpation hypothesis have been formulated over the past 20 years to explain the transition to addiction in humans in the real world. Regardless of the details, however, all these variants somehow imply that being born with a normal midbrain dopamine system in an environment where cocaine is readily available should make one vulnerable to addiction if exposed to the drug. At first glance, this implication seems to receive daily support from many laboratories across the world where thousands of rats and mice readily learn to press a lever (or to nosepoke a hole) to self-administer cocaine directly into their veins via an indwelling catheter, provided that the available dose is sufficiently high. In addition, most rats will escalate their intake of cocaine over time if given sufficient daily access to the drug. However,
the drug usurpation hypothesis of addiction seems to conflict with epidemiological research showing that only a minority of drug users eventually go on to develop a disorder of addiction. For instance, in the case of cocaine, one estimates that 17% of cocaine users develop addiction over a lifetime [33] and 6% within 2 years after initiation of use.
Though comparing experimental research on animals (that supports the drug usurpation hypothesis) with epidemiological research on drug addiction (that contradicts the drug usurpation hypothesis) may seem too overstretched, there is nevertheless an apparent chasm in outcomes between the two that needs some clarification.Epidemiological findings are mainly based on retrospective verbal self-reports and are thus not as objective as experimental findings in animals, however. In addition, not all cocaine users have equal access to cocaine in the real world, mostly because it is illegal, difficult to obtain and expensive. It is thus possible that cocaine users who report no addiction had insufficient access, occasion and/ or exposure to sufficient doses of cocaine after initiation of use. If all cocaine users have had equal access to cocaine, then most of them would have become addicted, as predicted by the drug usurpation hypothesis of addiction. Since such equal exposure scenario is unattainable and undesirable in the real world, there is always some lingering uncertainty surrounding the interpretation of epidemiological research on cocaine addiction. Such uncertainty is clearly absent in research on animals where drug dose and access are entirely under experimental control. However, experimental research on animals has its own caveats. Specifically,
there is some doubt surrounding the drug-taking behavior of animals that have ready access to drugs but without other valuable choices. This standard drug access setting clearly differs from that of many human drug users in the real world who have also access to a wide range of nondrug options amongst which to choose.[...]
Addiction versus drug use driven by lack of choiceAddiction is often defined in opposition to choice. [...] The term ‘choice’ has several meanings, however, that need to be distinguished here to avoid confusion. This term may refer to the act of choosing (or of choice-making) per se or to the different options from which to choose. In this Opinion article, we almost exclusively refer to the latter meaning.
Clearly, addiction as a psychiatric disorder is defined as resulting from brain dysfunctions that affect choice-making but not as resulting from a lack of options from which to choose. Many current diagnostic criteria of addiction refer to continued drug use despite and at the expense of other valuable activities and occupations, not to drug use because of lack of these other activities or occupations. Drug use caused by the pathology of addiction is not drug use in response to lack of choice. [...] Nevertheless, it may prove difficult in practice to discriminate between these two behaviors, especially in experimental animals that cannot report on their internal choice-making processes. One way to discriminate drug addiction from drug use driven by lack of choice is to test whether and to what extent an individual animal would continue to take the drug if given access to other options, preferably highly valuable ones. If it continues to take the drug despite and at the expense of these other opportunities, then it has likely developed an addiction- like behavior. In contrast, if it stops or cuts down cocaine use when offered a choice, then it is unlikely to be addicted (see below for additional discussion).
[...]
Choice can make the differenceIdeally, this test could be achieved by reconstructing in the laboratory a small version of the real world, as was done in the famous Rat Park experiment. In Rat Park, rats lived together in a large enriched colony that offered several different behavioral options, including taking morphine orally from a drinking bottle.
When compared to rats living alone in a standard housing cage, rats living in Rat Park drank much less morphine. This outcome shows that the possibility to engage in other pursuits during drug access, including social interactions, can prevent drug use in most rats — a finding that challenges the drug usurpation hypothesis of addiction. Individual morphine consumptions in Rat Park were not reported, however, so
it is not known whether there were few rats that drank morphine despite and to the detriment of the other available pursuits. In addition, recreating a real-like social environment in a laboratory is cumbersome and is not technically propitious for other more efficacious and addictive routes of drug self-administration, such as the intravenous route. This probably explains why despite its seminal contribution to the field,
there has been so far no attempt to replicate the initial Rat Park experiment. There have been many environmental enrichment experiments but these experiments involved enrichment of the home environment and not enrichment of the drug self-administration environment, as in the Rat Park experiment. As a result, in these experiments, rats had no alternative choice during drug access.
[...]
Most subsequent research on drug choices was conducted on cocaine and showed that though cocaine preference in animals is dose-dependent, it is nevertheless surmountable by sufficiently increasing the value of the alternative option in many, though not all, animals — a finding that clearly contradicts the drug usurpation hypothesis of cocaine addiction.
For instance, in one recent series of studies, rats were allowed to choose between pressing a lever to get water sweetened with saccharin — a potent, albeit biologically inessential, nondrug reward — or an alternative lever to receive an intravenous dose of cocaine. An equal level of effort was required on both levers and choice was either-or to insure that drug preference would generate significant opportunity costs.
Facing this choice, most rats chose almost exclusively the nondrug alternative, even after a long history of cocaine self- administration without choice. The behavior of nondrug-preferring rats was robust to a wide range of experimental conditions, including increasing drug doses per injection and increasing levels of past drug consumption, and could not be explained by the anxiogenic effects of cocaine. At the highest level of past cocaine consumption, only a minority of rats (about 15%) preferred cocaine, even when the opportunity cost of not choosing the alternative was increased. The behavior of drug-preferring rats could not be attributed to a mere disinterest in or aversion to sweet water. Notably, in a variant of the choice procedure incorporating a fixed- interval schedule component, cocaine-preferring rats continued to seek cocaine for several minutes at the expense of seeking the nondrug alternative [Augier et al., in preparation]. Since continued drug use at the expense of other rewarding activities and despite associated costs are hallmarks of drug addiction, their behavior is more likely due to an addiction-like state.
Thus, having alternative choices during drug access seems to reveal a distribution of drug-taking behaviors that is comparable to that found in epidemiological research (i.e., most drug users do not become addicted, only few do).Overall, this pattern of results has been reproduced in other laboratories and with other drugs of abuse, in- cluding nicotine and heroin. [...]
In humans, though women initiate cocaine use less frequently than men, they never- theless become more rapidly addicted to cocaine than men once drug use is initiated. Second, the type of drugs of abuse appears also to be a significant determinant of drug preference in rats. We recently found that drug preference over saccharin is more frequent with heroin than with cocaine after extended drug use, suggesting that heroin may be more addictive than cocaine. Once again, this conclusion is consistent with some comparative epidemiological findings showing that the prevalence of heroin addiction among heroin users is higher than the prevalence of cocaine addiction among cocaine users.
The neurobiology of nonaddiction revealedThe behavior of rats that remain able to turn away from cocaine for another pursuit, even after extended drug use, suggests that cocaine usurpation of brain dopamine functions is insufficient alone to drive addiction-like behavior in rats. This interpretation is supported by recent research using optogenetic methods in mice. [...] This outcome suggests that phasic activity of dopamine neurons is only one element of the neuronal substrates of sucrose reinforcement. This conclusion is further supported by a recent electrophysiological recording study in the nucleus accumbens of rats trained to self-administer cocaine and sucrose under a multiple reinforcement schedule. It turns out that vastly more nucleus accumbens neurons selectively encode sucrose than cocaine, even after a long period of abstinence during which cocaine seeking is known to incubate. A similar predominance of sugary juice-selective neurons over cocaine-selective neurons has also been previously reported in the nucleus accumbens of monkeys, though it was less pronounced than in rats. [...] More generally, the relative sizes of nucleus accumbens neuronal representations of different options may represent their relative reinforcing or motivational values and may thus serve as good neuronal predictors of individual preferences, as previously suggested.
Recovery from addiction and possible alternative interpretationsThough the behavior of nondrug-preferring rats strongly suggests absence of or even resilience to addiction, other interpretations are nevertheless possible. First, the fact that most rats turn away from cocaine when given another choice could be interpreted as evidence for a recovery from addiction. One major goal of virtually all addiction treatments is indeed to replace or displace addictive drug use with more desirable behaviors and activities. Contingency management-based treatments seek to achieve this goal by helping people with addiction to get access to alternative activities and occupations. In addition,
most people who succeed to quit drugs without professional assistance generally report doing so following a major change in life that generally opens up new opportunities and/or responsibilities (e.g., birth of a child; new job). In this context, rats that stop cocaine use when offered a different pursuit after extended drug use could be viewed as recovering from addiction and thus could serve as a unique animal model to investigate the neurobiology underlying addiction treatment or recovery. Second,
the behavior of nondrug-preferring rats could also be interpreted as evidence for food or sugar addiction rather than as evidence for absence of or resilience to cocaine addiction. Palatable foods, generally containing sugar or other sweeteners, are frequently used in addiction neurobiology as a control reward for drugs of abuse. However,
there is growing concern that palatable foods can lead to addiction-like behavior in humans and animals. At the neurobiological level, there are many similarities in the neural correlates or consequences of drug addiction and food addiction. There are, however, also many behavioral and neurobiological differences between these two conditions and there is some ongoing debate about whether the similarities are sufficient to outweigh the differences. Regardless of the outcome of this debate, future research involving nonfood rewards during drug self-administration should contribute to resolve this issue. Finally, one could also postulate that choice between cocaine and sweet reward primarily depends on dopamine-independent
hedonic processes rather than on dopamine-dependent
incentive salience processes that are hypothesized to be usurped by cocaine.
As one would expect, there is some evidence that hedonic factors can indeed influence choice between cocaine and sweet reward. For instance, diazepam — a benzodiazepine that increases the hedonic impact of sweet water further increases sweet preference in rats. However, evidence for a role of hedonic processes does not exclude the involvement of incentive salience processes during choice-making. In fact, in our choice procedure, it is difficult to see how rats could choose between the saccharin-paired lever and the cocaine-paired lever if these two levers had not previously acquired different incentive values. Future research is needed to explore the contribution of hedonic versus incentive salience processes to choice among drugs of abuse and nondrug rewards.
ConclusionsWe have argued that studying animals that are given access to drugs for self-administration without other choices may no longer be adequate and can even lead to the formulation of unrealistic views about the neuro- biology of addiction, as shown here by a critical evaluation of the drug usurpation hypothesis of cocaine addiction. Addiction as a psychiatric disorder is defined as resulting from brain dysfunctions that affect normal choice-making, not as resulting from a lack of options amongst which to choose. We encourage researchers to increase animals’ choice during drug access. This will uniquely allow one to discriminate addiction from drug use driven by lack of choice, thereby increasing the chance of discovering specific addiction-related brain dysfunctions. More generally, this research should also provide new insights about how the brain ‘chooses’ among qualitatively different kinds of pursuits — a still largely unresolved question.