Publishing Blackwell LtdOxford, UKADBAddiction BiologyJournal compilation 2006 Society for the Study of Addiction
2006 00
Original Article
Behavioural assessment of drug reinforcement and addictive features Carles Sanchis-Segura & Rainer Spanagel
REVIEW
doi:10.1111/j.1355-6215.2006.00012.x
Behavioural assessment of drug reinforcement and addictive features in rodents: an overview
Carles Sanchis-Segura & Rainer Spanagel
Department of Psychopharmacology, Central Institute of Mental Health, Germany
ABSTRACT Some psychoactive drugs are abused because of their ability to act as reinforcers. As a consequence behavioural patterns (such as drug-seeking/drug-taking behaviours) are promoted that ensure further drug consumption. After prolonged drug self-administration, some individuals lose control over their behaviour so that these drug-seeking/taking behaviours become compulsive, pervading almost all life activities and precipitating the loss of social compatibility. Thus, the syndrome of addictive behaviour is qualitatively different from controlled drug consumption. Drug-induced reinforcement can be assessed directly in laboratory animals by either operant or non-operant self-administration methods, by classical conditioning-based paradigms such as conditioned place preference or sign tracking, by facilitation of intracranial electric self-stimulation, or, alternatively by drug-induced memory enhancement. In contrast, addiction cannot be modelled in animals, at least as a whole, within the constraints of the laboratory. However, various procedures have been proposed as possible rodent analogues of addictions major elements including compulsive drug seeking, relapse, loss of control/impulsivity, and continued drug consumption despite negative consequences. This review provides an extensive overview and a critical evaluation of the methods currently used for studying druginduced reinforcement as well as specic features of addictive behaviour. In addition, comic strips that illustrate behavioural methods used in the drug abuse eld are provided given for free download under http://www.zi-mannheim/ psychopharmacology.de Keywords Addictive behaviour, drug seeking, drug taking, loss of control, methods, reinforcement, relapse.
Correspondence to: Carles Sanchis-Segura, Department of Psychopharmacology, Central Institute of Mental Health, J5, 68159 Mannheim, Germany. E-mail: [e-mail]
REVIEW
1 1.1 1.1.1 1.1.2 1.2 1.2.1 1.2.1.1 1.2.1.2 1.2.2 1.2.3 1.2.4 2 2.1 2.1.1 2.1.1.1 2.1.1.2 2.1.2 2.1.3 METHODS USED TO ASSESS DRUG-INDUCED REINFORCEMENT Self-administration models Non-operant methods Operant drug self-administration procedures Tests used to measure the reinforcing properties of drugs of abuse Tests based on conditioned preference Place conditioning procedures Other measures of conditioned preference Other procedures based on Pavlovian conditioning: autoshaping (sign tracking) Facilitation of intracranial electric self-stimulation Drug-induced memory enhancement as an alternative assessing the reinforcing effects of drugs BEYOND REINFORCEMENT: ADDICTIVE BEHAVIOUR Modelling features of addictive behaviour Modelling drug seeking and relapse Assessing drug seeking by the reinstatement model Assessing relapse by drug deprivation Modelling loss of control/impulsivity Modelling drug consumption despite negative consequences
Addiction Biology, 11, 238
2006 The Authors. Journal compilation 2006 Society for the Study of Addiction
Behavioural assessment of drug reinforcement and addictive features
3
2.2 2.2.1 2.2.2 2.2.3 3
Tests currently used in the study of addictive behaviour Behavioural sensitization Second order schedules in drug-seeking behaviour assessment The use of conditioned place preference to measure the reinstatement of drug-seeking behaviour CONCLUDING REMARKS
FOREWORD Addiction is dened as a syndrome in which drug use pervades all facets of the users life, even precipitating in the loss of social compatibility (e.g. loss of partner and friends, loss of job, crime . . .) [1]. It is obvious that addiction is a genuinely human phenomenon; and one that is therefore no reproducible within the unavoidable constraints imposed by the laboratory setting. However, some of the behavioural characteristics of this syndrome, such as resumption of drug seeking/drug consumption after a protracted abstinence (relapse), can be satisfactorily modelled in laboratory animals. For example, experimental procedures can be designed to be as simple as possible, thereby maximizing internal validity and thus reproducibility. Conversely, procedures can aim to be as holistic as possible, thereby favouring the possible relevance to human situations. Neither one of these two approaches is perfect; each has its respective drawbacks. Indeed, methods designed attending only to internal validity may nally exclude variables of relevance to understanding, explaining or predicting the phenomenon of interest. On the other hand, very complex procedures enhance the difculty of arriving at conclusions and reduce the capability of both inferring causal relationships between variables and of establishing predictions. In the following pages, we summarize various methods currently used in the drug addiction eld. Methods that evaluate features of addictive behaviour and those measuring the reinforcing properties of drugs are discussed separately. It is assumed here that drug intake (promoted and maintained by the reinforcing properties of those substances) is a requirement for the development of addiction; however, addiction is neither a necessary nor a universal consequence of drug consumption. This review also considers that physical dependence and other consequences derived from longterm drug consumption can be concurrent to the development of addiction, but that they are not aetiologically related and can be dissociated for their study (Li & Volkow 2005). Therefore, procedures measuring phenomena such as tolerance or dependence are not included in the present review. Other methods (i.e. drug discrimination procedures) that provide additional kinds of information relevant to understanding drug consumption and addictive behaviour in all its complexity, but that do not fall in any of these two main categories (reinforcement versus addictive features), are not discussed as well. It is important to note that in this review the term model will be restricted to those methods that display clear face validity towards human behaviour and phenomena. For example, the rodent drug self-administration procedures currently used in this eld are aimed at addressing the main features of the same behaviour in humans, i.e. they show clear face validity. On the other hand, the term test will be applied to describe the experimental methods that do not have a direct resemblance to the human condition (i.e. place conditioning, intracranial self-stimulation-based procedures, etc.).
1 METHODS USED TO ASSESS DRUG-INDUCED REINFORCEMENT Nowadays, it is assumed that drugs are voluntarily taken and potentially abused because of their reinforcing properties, that is because they act as reinforcers of drug-seeking and drug-taking behaviours. While this is probably true, it is also important to consider what the term reinforcer actually means. In the context of drugs of abuse, reinforcers are
[1] The denition of addiction has changed across time and different denitions have been related to the specic characteristics of specic drug classes. Thus, the current denition of addiction ts much better to (and it is mainly studied in) psychostimulants such as cocaine or amphetamine, whereas 10 years ago physical dependence and withdrawal were considered as the landmarks of an alternative view of addictive behaviour more suitable for opiate drugs. Interestingly these changes are not independent of the trends in drug consumption in our society and therefore the consideration of specic drugs as health/social problems. In the clinical language dependence is the most general term to refer to the syndrome that most preclinical researchers dene as addiction. This produces some confusion because the term dependence (or less preferably physical dependence) is used in preclinical research to dene a latent withdrawal state that would be potentially triggered in the absence of the drug. According to the current evidence obtained in animal models, withdrawal symptoms do not seem to be necessary or sufcient for the development of addictive behaviour, but this observation is often ignored in the medical environments. In fact, the DSM-IV (one of the most accepted diagnostic manuals for behavioural disorders) still includes tolerance and dependence as important symptoms of human addiction.
2006 The Authors. Journal compilation 2006 Society for the Study of Addiction Addiction Biology, 11, 238
4
Carles Sanchis-Segura & Rainer Spanagel
mainly considered in a Skinnerian view and therefore dened as events that follow a response and change the probability of future occurrences of that response. However, this perspective often ignores that the reinforcement process also accounts for changes in behaviour when the reinforced response appears in the absence of the reinforcer or when the reinforcer is non-contingently administered. In addition, any comprehensive denition of reinforcement should account for the generalization and pre-eminence of previously reinforced behavioural outcomes in novel situations. Therefore, the actions of reinforcers should be understood in a broader context intimately related to learning and memory processes (White & Milner 1992; White 1996; Hyman 2005). At this point it is necessary to highlight that very often the term reward is misused and confused with the term reinforcer (or the reinforcement process). Reward, as a scientic concept, was coined within the eld of experimental psychology and has three possible meanings. First, reward can be used as it would be used in a non-scientic context to describe stimuli with appetitive (desirable) consequences. Second, reward can be used (as opposed to punishment) to refer the learning contingency in which the emission of a response brings such an appetitive stimulus. This kind of contingency is often referred as positive reinforcement. Third, reward is also used to refer to a hypothetical pleasurable internal state (hedonia), which derives from the acquisition, use or consumption of appetitive stimuli. In this regard, as summarized by Everitt & Robbins (2005), reward (or related concepts, such as liking) refers to the subjective responses associated with the post-presentational consequences of reinforcers, becoming later on important characteristics of the internal representation of these stimuli (and those others surrounding their occurrence). It is also important for the purpose of this review to clarify that so far these hedonic attributive processes are simply hypothetical and that their measurability in rodents is an open debate. Regardless of which one of these three meanings is ascribed to reward, this concept cannot be equated to reinforcement. Reinforcement is a broader concept that refers to the ability of some stimuli (reinforcers) to change the probability of specic behavioural repertoires in different learning contingencies (positive versus negative reinforcement [2]). This concept includes reward-related processes but also reward-independent mechanisms which lead to an increase of the emission probability of a particular response. Acting as reinforcers, drugs can promote changes in the probability of emission of specic responses in three ways: rst, reinforcers can reduce specic needs or drives (negative reinforcement). This perspective is perhaps not essential to understand the reinforcing properties of drugs of abuse in initial states of its consumption but the importance of this phenomenon is likely to grow after prolonged drug exposure and/or when introducing deprivation/abstinence phases. Second, drugs of abuse can act as primary motivators in positive reinforcement contingencies (positive reinforcement). In those situations, other stimuli associated with the presence of the drugs can also acquire incentive-motivational properties becoming conditioned reinforcers. Finally, and referring to the initial meaning attributed to this concept by Pavlov, Thorndicke or Hull, reinforcers can enhance the storage of information about situations in which they occur, via a process that does not involve learning about the reinforcer itself. Thus, reinforcers by promoting an increased associative strength of specic stimuli-response contingencies can bias the choice of particular responses and increase their probability of emission. These different dimensions of reinforcers can be used to study drug-induced reinforcement. Thus, as mentioned before, the traditional view of reinforcers in this eld considers them as primary appetitive stimuli, and their motivational properties are often evaluated through consumption/preference-based measures. However, methods that measure the strengthening action of drugs on learning/memory processes are briey introduced together with their possible usefulness in the context of drug abuse research. 1.1 Self-administration models Self-administration-based methods are widely used in basic/preclinical drug abuse research. This is because these procedures have a good construct as well as appealing face validity towards drug consumption in humans. Of course, how the drug is obtained or even why the drug is consumed varies notably between both a human being and his/her social environment and a non-human experimental subject in a constricted laboratory set-up. However, it is assumed that the neural chemistry and anatomical circuitry involved generating, selecting and setting in motion these behavioural patterns is similar in both situations. Consequently, these procedures appear to be adequate models in unravelling
[2] Positive and negative reinforcement result in an identical outcome: the increase of probability of a particular response. The difference between both phenomena refers to the underlying process. Thus, the term positive reinforcement refers to situations in which the emission of a response leads to the presentation of an appetitive consequence. Conversely, in a negative reinforcement contingency the probability of emission of a particular response is increased because of its ability to suppress an undesirable event. Drugs of abuse can promote both kinds of reinforcement, and self-administration behaviours can be elicited and sustained to obtain the drug but also to suppress the effects of its absence (i.e. prevent withdrawal).
2006 The Authors. Journal compilation 2006 Society for the Study of Addiction Addiction Biology, 11, 238
Behavioural assessment of drug reinforcement and addictive features
5
common neural mechanisms and therefore help to identify strategies useful in the intervention regarding human drug consumption. The self-administration procedures can be classied according to different criteria. Thus, from a pharmacological perspective they can be classied according to the route of administration via which the drug is ultimately delivered to the organism. This is not a trivial consideration because by determining the latency between the response and the perceived effects as well as the amount of the drug, the route of administration partially determines several drug effects, including those that allow a substance to act as a reinforcer. From a behavioural perspective self-administration methods can be classied as operant and non-operant procedures. When using operant procedures the dependent variables analysed refer to the response itself (frequency, rate, etc.) whereas the most commonly reported dependent variables in non-operant procedures are centred in the amount consumed. Thus, methods based on operant and non-operant responses differ in procedural characteristics, but also may differ in their sensitivity to the manipulation of specic brain substrates and may require a differential framework in the experimental design and in the results interpretation. 1.1.1 Non-operant methods In rodents, non-operant procedures are restricted to oral self-administration procedures. These kinds of methods are very common in the context of alcohol research but they have been also, although less frequently, used with other drugs of abuse such as nicotine (Slifer 1983), cocaine (Falk & Lau 1997), amphetamine (Meliska et al. 1995) or morphine (Schuster, Smith & Jaffe 1971). The most obvious reason for choosing a non-operant self-administration procedure is that rodents hardly consume alcohol via other routes of administration. Conversely, when drugs such as morphine or amphetamine ar...
Sanchis-Segura & Spanagel 2006 rev addiction models