Conditioning Models of Addiction, Part 1

There is a substantial body of research that shows that the ingestion of psychoactive substances and the development of problematic substance use or addiction involve psychological processes similar to those involved in normal appetitive behaviours such as eating, drinking and sex.

Research in laboratory animals has provided many insights into the role of reinforcement, learning and conditioning in normal appetitive behaviours, as well as in the misuse of psychoactive substances. In this regard, it is important to note that when given the opportunity, laboratory animals, such as the rat, learn to self-administer psychoactive drugs (except LSD).

Over millions of years, the brains of animals have evolved a motivational system that helps animals’ survive and reproduce. Behavioural responses that lead to positive consequences, such as the reduction of hunger, are likely to be repeated. Moreover, animals learn to escape from or avoid painful or noxious stimuli.

Operant conditioning, or instrumental learning, refers to the way in which the consequences of behaviour influence the likelihood of that behaviour being repeated. One class of consequence which can affect behaviour, positive reinforcement, is illustrated by a laboratory rat learning to press a lever to obtain food, or a dog sitting up to beg for a biscuit.

Drugs of dependence tap into the motivational system underlying this behavioural change. The drug acts as a reward, or positive reinforcer, and with repetition the association between cue, response and reward becomes stronger and stronger.

Another important principle here for understanding problematic substance use is the immediacy of reinforcement. It is well-established that the sooner a reinforcer follows a behaviour, the more powerful its effect will be on that behaviour and the more likely the behaviour is to be repeated.

A second class of consequence that can affect behaviour (negative reinforcement) can be demonstrated in the laboratory by training a rat to press a lever to avoid being punished by, for example, a small electric shock to the feet. Each time the animal receives the cue (e.g. a light predicting impending shock), it will perform an operant response to avoid the shock being delivered.

Similarly, the dependent heroin user may take the drug (perform an operant response) to avoid impending withdrawal symptoms and the associated physical and psychological discomfort.

It is important to emphasise that these instrumental learning mechanisms can operate outside conscious awareness and not involve a decision-making process.

West points out that in this model, addiction can be viewed as involving the “development of a habitual behaviour pattern that is independent of any conscious evaluation that might be taking place about the costs and benefits of the behaviour. The impulses to engage in addictive behaviour that are generated by this mechanism can be so strong that they overwhelm the desire of the addicts to restrain themselves.”

Classical (or Pavlovian) conditioning is a process that involves a neutral stimulus (such as a red light) become rewarding and influencing behaviour because it has reliably preceded a natural reward such as food.

In Ivan Pavlov’s seminal experiments at the turn of the 20th century, salivation was demonstrated in dogs presented with food. After a neutral stimulus (bell) was presented in combination with the food on a number of occasions, the bell became capable of eliciting salivation in the absence of the food. Thus, the bell had become a conditioned stimulus capable of influencing behaviour, i.e. producing a conditioned response.

Conditioned stimuli play an important part in our daily life, and they have played a significant role in evolutionary terms, in respect of the survival of the species. They allow us to react to threatening situations and alert us to such necessities as food and sexual partners; they shape behaviour.

As discussed earlier for operant conditioning, classical conditioning processes can become automatic. Behaviour can be influenced without conscious, decision-making processes.

I know this well from lighting the gas ring above an oven that had been left on for many hours: I was blown across the room, fortunately with only hairs singed. But I was left with a strong conditioned response, such that every time I heard a sound near a gas stove, I literally jumped out of my skin. The response took years to extinguish.

These stimuli, such as Pavlov’s bell, are known as secondary reinforcers because they derive their ability to influence behaviour by association. Secondary reinforcers can generalise in the sense that stimuli with similar characteristics (e.g. similar colour light) will produce a similar, but not necessarily identical, impact on behaviour.

The impact of the conditioned response can also extinguish, in that if presentation of the bell is not followed by food on a number of occasions, salivation in the dog will disappear.

In the next Briefing, we will look at the role of classical conditioning in substance use and addiction, considering conditioned withdrawal, conditioned drug-opposite responses and conditioned tolerance, and conditioned drug-like responses.

Recommended reading:

Robert West (2006) Theory of Addiction. Blackwell Publishing.

Nick Heather and Ian Robertson (2001) Problem Drinking. Oxford Medical Publications.

<. pdf document

> Conditioning Models of Addiction, Part 2

Conditioning Models of Addiction, Part 2

In our last Briefing, we described classical conditioning as a process that involves a neutral unconditioned stimulus (UCS), such as a coloured light, becoming rewarding and influencing behaviour because it has reliably preceded a reward such as food.

During a history of drug use, certain stimuli, such as environmental contexts or drug paraphernalia, reliably accompany drug administration. These stimuli, by virtue of their pairing with the drug effects, become conditioned stimuli (CS) capable of eliciting conditioned responses (CRs), e.g. drug-seeking behaviour.

There are three ways that classical conditioning may be involved in problematic substance use or addiction.

In the first proposed model of conditioning, the conditioned withdrawal model, Wikler (1948) proposed that environmental stimuli paired with drug withdrawal became conditioned stimuli (CS) capable of eliciting conditioned withdrawal reactions (CRs).

For example, in people dependent on heroin, withdrawal symptoms can occur and be paired repeatedly with environmental stimuli. At a later time, when the individual is no longer dependent, the environmental cues alone can be enough to elicit the symptoms of withdrawal.

The cues that trigger conditioned withdrawal can be both external (places or situations) or internal (moods). Conditioned withdrawal can play a prominent role in relapse.

In fact, the conditioned withdrawal model of addiction involves both classical and operant (or instrumental) conditioning. Repeated pairing of environmental stimuli with withdrawal results in these stimuli being capable of inducing conditioned withdrawal (classical conditioning).

The instrumental conditioning component involves the person taking the drug to alleviate an aversive state, the withdrawal symptoms, which can be regarded as a negative reinforcer.

The second classical conditioning involves the concepts of conditioned drug-opposite responses and conditioned tolerance.

Whenever a disturbance occurs in the body, such as produced by a drug, a physiological process known as homeostasis, in which the body tries to counteract the disturbance, comes into play.

For example, amphetamine enhances release of the neurotransmitter dopamine in the brain, but at the same time regulatory mechanisms reduce dopaminergic function in order to try and maintain the status quo – although the amphetamine still increases dopamine function overall.

Researchers believe that these compensatory mechanisms can eventually be triggered by stimuli and cues previously associated with drug administration, and this can happen even before the drug is taken.

In situations where the predictive stimuli appear but no drug is taken, the body’s compensatory mechanisms come into play and go unopposed because there is no drug effect. This can be expressed as overt physiological reactions and/or form the basis for the subjective experience of withdrawal sickness and craving.

Take for example a person who is drinking alcohol every evening to reduce the anxiety they have experienced from working in a stressful job. The clock at work approaching 17.00, and the sights and sounds of the pub, act as conditioned stimuli to the anxiety-alleviating effects of alcohol.

If the person were to attend a school play one evening, without going to the pub, their body’s compensatory mechanisms would come into play but not be diminished by the physiological effects of alcohol. The person would experience the opposite subjective effects to those produced by alcohol, i.e. anxiety.

According to this model, tolerance and withdrawal symptoms are intimately linked.

Tolerance – the gradual diminution of effect following repeated administration of the same dose of drug – is thought to occur because of the homeostatic processes that occur in the body to counteract the action of a drug. The homeostatic (or opponent) responses are thought to be strengthened by repeated drug administration, and the net effect of the drug (original effect minus the opposing effect) is therefore reduced.

These processes are explained in more detail by the Opponent Process Theory of Solomon and Corbit (1973), summarised in Robert West’s book “Theory of Addiction”.

Shepard Siegel (1975) first proposed that a complete account of tolerance requires an appreciation of the role of environmental influences or cues.

There is now an abundant evidence showing that animals pre-administered a drug repeatedly in one environment and tested behaviourally in another environment, will not show as much tolerance as those animals given chronic drug and behavioural testing in the same environment.

An important consequence of this idea in relation to heroin overdose was illustrated by Shepard Siegel in the early 1980s. Tolerance develops to the effects of heroin, so that users face the possibility of overdose (and death) if they take much larger amounts of drug than normal.

Siegel reasoned that if tolerance to heroin was partially conditioned to the environment where the drug was usually administered, if the drug was administered in a new setting, much of the conditioned tolerance would disappear, and the person would be more likely to overdose.

In his study, many heroin users admitted to hospital suffering from a heroin overdose reported that they had taken this near-fatal overdose in an unusual environment, or that their normal pattern of use was different on that day.

Recommended reading:

Robert West (2006) Theory of Addiction. Blackwell Publishing.

Nick Heather and Ian Robertson (2001) Problem Drinking. Oxford Medical Publications.

< pdf document

> Conditioning Models of Addiction, Part 3

Conditioning Models of Addiction, Part 3

In our last two Briefings, we looked at two ways that classical conditioning may be involved in problematic substance use or addiction. We described the conditioned withdrawal model, as well the concepts of conditioned drug-opposite responses and conditioned tolerance.

In the conditioned incentive model of addiction, proposed by Jane Stewart and colleagues in the mid-1980s, environmental stimuli previously associated with the pleasurable effects of drugs become conditioned stimuli (CS) via classical conditioning processes.

These CS are considered to activate the same neuronal pathways in the brain that mediate the direct pleasurable effects of drugs, albeit weakly, and they thereby elicit a motivational state that directly primes drug-taking behaviour. The CS are positive incentives that drive drug use.

Thus, when a heroin user sees the paraphernalia that they usually use for administering the drug, the paraphernalia act as a CS that elicits feelings somewhat similar to that triggered by the drug itself, which result in the person wanting to use the drug.

The present model is derived primarily from incentive motivation theory, which was developed on the basis of work with laboratory rats using natural reinforcers such as food. This theory asserts that organisms are motivated by incentives, stimuli that predict a primary reward. The motivation is the expectancy of the primary reward, be it food or drug.

Thus, one person may want to eat a doughnut when they see the bakery assistant who regularly sells them their favourite vice, whilst another person may want to inject heroin when they see their regular dealer.

There is considerable evidence from animal research that positive incentive effects of drugs motivate drug-seeking behaviour.

In the place conditioning paradigm, rats are introduced to a three-compartment box, containing two end compartments with distinctly different environments (light walls, grid floor vs. dark walls, smooth floor), and a smaller “neutral” central area. The time spent in each of the end compartments is measured over a 15-minute period, and one side is assigned as the original least-preferred side.

In subsequent sessions (days 1, 3 and 5), animals are administered a drug of abuse such as amphetamine and restricted to their original least-preferred side for 30 minutes. On days 2, 4 and 6, they are administered an inert substance (saline) and restricted to the original preferred side.

On the following day, the rats are given free access to all parts of the box, with the time spent in each end compartment measured. When given this free choice, rats show a shift in preference towards the side in which they had received the drug – even though no drug was administered in this test session.

These studies demonstrate that a wide variety of drugs of abuse (e.g. amphetamine, cocaine, heroin), as well as natural reinforcers such as food, can induce place conditioning. Thus, environments associated with the pleasurable effects of drugs, or natural reinforcers, become positive incentives that motivate approach behaviours.

We can safely assume that animals find the effects of drugs of abuse to be pleasurable in that they will learn to perform specific tasks (e.g. pressing a lever in a Skinner box) to obtain intravenous injections of drugs of abuse such as amphetamine, cocaine and heroin. They also learn to respond to a stimulus (e.g. a light) that was previously associated with their lever presses for drug.

Brain dopamine neurons, in particular those projecting from a midbrain region known as the ventral tegmental area to forebrain regions such as the nucleus accumbens (mesolimbic dopamine neurons), are thought to play a major role in mediating drug self-administration.

Terry Robinson and Kent Berridge, two leading researchers from the States, propose that the primary role of mesolimbic dopamine neurons is to mediate what is called incentive salience.

Incentive salience is a characteristic of the mental representation of a stimulus that allows it to become attractive and wanted, thereby eliciting approach behaviours towards a specific goal. [A juicy piece of apple pie possesses a high degree of incentive salience – at least to me!]

In their incentive sensitisation model, Robinson and Berridge propose that drugs of abuse produce a long-lasting sensitisation of the neural system mediating incentive salience (mesolimbic dopamine system), so that the incentive salience attributed to drug-taking and to drug-associated stimuli become pathologically amplified, leading to compulsive drug-seeking and drug-taking.

The sensitisation of incentive salience can occur at the same that the pleasurable effects of the drugs are diminished, due to the repeated drug administration producing tolerance to this effect.

In fact, these researchers emphasise that the neuronal systems responsible for excessive incentive salience are dissociable from the systems mediating the pleasurable effects of drugs. ‘Wanting’ is not ‘liking’ – a person may strongly want a drug without actually liking the experiences that it produces.

Moreover, it is also proposed that the wanting system can be activated and influence behaviour without a person having conscious awareness of ongoing processes.

A considerable degree of animal research has been focused on drug-induced sensitisation, and the incentive salience model is very popular amongst neuroscientists. Whilst it has been argued that there is little evidence in humans supporting the model, this is in part due to a difficulty in testing the ideas.

Recommended reading:

Robert West (2006) Theory of Addiction. Blackwell Publishing.

> pdf document