Cognitive Models, Epidemiology and Neurobiology of Addiction

What are cognitive models of addiction?

Models that give a cognitive or information-processing account of drug use.

What can cognitive models include?

• Cue-related accounts (learning and memory)

• Expectancy theories

• Impulse control

• Some behavioural economic (choice) approaches.

What are the two broad approaches within cognitive models?

Motivational drive vs automaticity (the transition between them)

How does Hyman (2005) view addiction in terms of learning?

As a disorder of learning and memory. Associative learning is a natural process that can become pathological.

Why does the reward system exist?

To promote life-sustaining behaviours (e.g. seeking food) by learning which behaviours give rewards and which cues signal them.

What two types of learning are involved in reward?

• Instrumental learning (behaviour → reward)

• Conditioned learning (cue → reward association).

What is Thorndike's (1898) law of effect?

Animals explore new environments, and actions followed by reward are repeated. Reliable reward cues become important.

How do behaviours become automatic?

With repetition they become overlearned and can be triggered by environmental cues (e.g. availability, predictability).

What are conditioned compensatory responses (Wikler 1948; Siegel 1975)?

• Classical conditioning where the environment becomes a cue for the drug.

• The body prepares by making compensatory physiological responses, which produce cue-related cravings.

What is situational tolerance?

Tolerance built up in the usual drug-taking environment, because the body prepares there.

Why can taking the same dose in a new environment cause overdose?

The cues are absent, so the body doesn't make the compensatory response and the dose has a bigger effect.

What are the treatment implications of conditioning research?

• Cue-exposure therapy

• Cognitive bias modification

• Pharmacotherapies that block the reinforcing effects of drugs.

Who proposed Incentive Sensitisation Theory (IST)?

Robinson & Berridge (1993, 2000).

What are the three components of reward in IST?

• Liking (hedonic reaction)

• Wanting (incentive salience/motivation)

• Associative learning.

• They can be dissociated biologically.

Which brain region controls wanting?

The nucleus accumbens (NAcc).

What is incentive salience?

The motivational 'attractiveness' of a stimulus.

What does repeated drug use do, according to IST?

It sensitises neural pathways, increasing incentive salience for the drug and its cues, which produces intense wanting triggered by cues.

What is the key shift in IST?

From drug liking to drug wanting. Behaviour is driven less by pleasure and more by incentive salience.

How does IST differ from stimulus-response (S-R) positive reinforcement theory?

• S-R says the cue directly drives drug seeking.

• IST says the cue acts through strengthened reward-pathway relationships.

What does incentive salience underlie in drug users?

Attentional biases to drug cues (a phenotypic element).

What does the transition to addiction involve (IST)?

A shift from responding to the drug to responding to the cues.

What did the rat Skinner box study show about cues?

• Dopamine release shifted from the reward itself to the conditioned cue (light/tone).

• Rats only sought reward when the cue was on, and the cue response was larger than the original reward response.

How is dopamine measured in humans with PET?

• Through competitive binding.

• A radiotracer competes with dopamine, so less tracer binding means more dopamine.

What happens to dopamine response to methylphenidate in cocaine users?

• It is greatly blunted compared to controls (barely any change).

• The reward system doesn't respond to the drug itself.

What happens when cocaine users watch cocaine-related cues?

They show a dopamine response, which correlates with self-reported craving.

What is the key dopamine pattern in addiction?

Blunted response to the drug but exaggerated response to drug cues.

Why do withdrawal models fail to explain relapse?

Relapse often occurs after withdrawal has passed, in the presence of cues (people, places, paraphernalia, sensations).

Why is cue-induced relapse clinically important?

Cue-triggered activation of habitual drug seeking is a key relapse mechanism, so it gives a new treatment target.

What is attentional bias (AB)?

Preferential allocation of attention to salient cues (threat, or reward-related cues like drugs). It guides and is guided by motivated behaviour.

How does the dot-probe task measure AB?

• Two images are shown (drug cue and neutral), then a probe appears.

• Faster responses when the probe replaces the drug cue indicate attention was drawn to it.

How is the dot-probe adapted for treatment (AB modification)?

The probe appears more often behind the neutral cue, training attention away from the drug cue.

What are the practical advantages and disadvantages of AB modification?

• It can be done remotely without a clinician

• It is tedious

• Only generalises if repeated

• Has low motivation to engage.

What are the issues with AB training?

• Task reliability/validity

• Multiple cues not modelled in animal extinction

• User motivation

• Unclear mechanism (e.g. perceived availability)

• AB being transient.

What did the real-time handheld AB study show?

AB and craving were measured via randomly sent links vs links sent when people were struggling. AB was strongly predictive of relapse, but it fluctuates within individuals.

What are the limitations of learning models?

They ignore conscious desires and motivations not acquired through experience, and cue-based treatments haven't been very successful (though this doesn't invalidate the theory).

Why is cue-based treatment difficult?

There are many possible cues, learning resists change, and cognitive tasks are effortful and repetitive.

What is the anatomical link between reward and cognitive control?

The reward pathway projects to the nucleus accumbens, with a parallel pathway to the frontal cortex, where cognitive tasks occur.

What does the PFC do normally (Hyman, 2005)?

Goal selection: representing goals, assigning value to them, and selecting actions. It actively maintains goal representations.

What happens to goal selection in addiction (Hyman, 2005)?

A pathological narrowing of goal selection to drug-related goals.

Why don't choice models fully explain addiction?

They don't explain the intense craving and compulsion that develop. These, plus poor self-control, are key arguments for the disease model.

Is impulsivity a cause or consequence of addiction?

Both. It raises risk of use and dependence, and drug use impairs frontal function and worsens impulsivity.

What are the dimensions of impulsivity?

• Inability to delay

• Risky behaviour

• Insensitivity to negative consequences

• Difficulty withholding responses.

What three lines of evidence link impulsivity and drug use?

• It's a risk factor (longitudinal studies)

• It's higher in users than controls

• Some drugs increase state impulsivity (e.g. delay discounting, stop-signal, go/no-go).

Why are longitudinal studies useful for impulsivity research?

They show what comes first (impulsivity predicts drug use), which cross-sectional studies can't.

Why are adolescents vulnerable to drug use?

• Exploration

• Rebelliousness

• Peer influence

• Limited responsibility

• An immature brain (frontal lobes, synaptic pruning).

What did Braet et al. (2009) find?

Adolescents show poorer response inhibition, linked to altered PFC function.

What did the opiate user delay-discounting study find?

Current users and abstinent ex-users both chose smaller immediate rewards more than controls, so impulsivity persists in abstinence.

What is the balance between impulse drive and impulse control?

Impulse drive is essential for survival, but frontal control is needed to weigh up the best choice.

What is the temporal transition in addiction?

Impulsivity (initiation) → compulsivity (addiction).

What is compulsivity?

Rigid, habitual (not goal-directed) drug use with loss of control, continuing despite negative consequences.

What do Robbins et al. hypothesise about the brain systems?

• Impulsive and compulsive actions are governed by different cortico-striatal loops under different chemical modulation.

• Impulsivity involves the ventral striatum, and compulsivity involves structures like the putamen.

• This is a theory, not proven.

What is the ventral-to-dorsal striatum shift?

Habit formation shifts control from ventral to dorsal striatum, which is involved in motor control.

What are the two types of instrumental control?

• Goal-driven (aware of action-outcome relationship)

• Stimulus-driven (reinforcer strengthens response-context/cue association).

How does impaired top-down control contribute to addiction?

• Reduced PFC function exacerbates the shift to S-R habits

• Addicts show loss of frontal white matter and altered D2 receptors.

What does dysregulated dopamine in addiction look like?

• Reduced activity in the OFC, cingulate gyrus and DLPFC

• Less top-down control

• A more active dorsal striatum

• An unbalanced system.

What is the summary of the transition to addiction?

Impulsive, goal-directed use becomes compulsive, habitual use, reflecting a ventral-to-dorsal striatal shift and weakened prefrontal control.

What is epidemiology?

The study of health at the population level: prevalence, influencing factors, and outcomes.

What is prevalence?

The proportion of a population with a condition at a given time.

Why is epidemiology useful for drug use and mental health?

It helps untangle whether drug use causes poor mental health, vice versa, or a third factor causes both.

When can we be more confident an effect is causal?

• When the exposure precedes the outcome

• The association appears across contexts

• Manipulating the exposure changes the outcome

• Other explanations (bias, chance, confounding) are ruled out.

What is confounding?

• When an observed effect is due to another factor influencing both exposure and outcome.

• Examples: childhood trauma, socioeconomic status, family history, chronic pain.

How do we deal with confounding?

Adjust for confounders in analyses or use methods robust to confounding.

What are examples of bias?

Collider bias, selection bias and measurement bias.

What is reverse causation?

When the outcome actually influences the exposure, e.g. mental health problems leading to drug use.

What is the main problem with cross-sectional studies here?

Data is collected at one time point, so we can't tell whether exposure came before substance use.

What is the problem with cohort studies for this question?

The outcomes (e.g. injecting drug use) are rare, giving small samples, and it's hard to recruit in early life and retain participants until adulthood.

What is the problem with case-control studies?

They rely on retrospective reporting, which leads to recall bias.

What is recall bias?

When people with the outcome are more likely to recall and report exposures than those without it, potentially creating false associations.

Why are RCTs the gold standard, and why can't we use them here?

Randomisation reduces bias and confounding, but it's unethical to randomise people to take drugs.

What does Mendelian randomisation (MR) do?

It uses genetic variants as instruments to overcome confounding and reverse causation, giving greater confidence in causality.

What are the limitations of MR?

• It needs good genetic instruments (large genetic studies of substance use aren't always available)

• Pleiotropy (variants influencing outcomes via other pathways) is hard to rule out.

What do 95% confidence intervals show?

• The uncertainty around an effect estimate

• Narrower means more precise

• If the CI crosses the null, the effect could be due to chance.

What is a p-value?

A value from 0 to 1 measuring the strength of evidence against the null hypothesis. Smaller means stronger evidence.

How should p < 0.05 be treated?

• It's commonly called 'significant', but avoid treating it as an arbitrary cut-off.

• Interpret it alongside prior evidence, CIs, effect size, bias and confounding.

What is Volkow's (2011) core argument?

• Dopamine is crucial for drug reward but its role in addiction is less clear.

• Addiction also involves circuits for conditioning/habits, motivation and executive function

• The review draws mainly on PET imaging in addicted humans.

What are the main risk factors for addiction (Volkow)?

• Genetics (~50% of risk)

• Development (adolescence)

• Environment (access, stress).

What are the three dopamine pathways, and what do phasic vs tonic firing do?

• Mesolimbic (VTA → NAc, drug reward)

• Mesostriatal (substantia nigra → dorsal striatum)

• Mesocortical (VTA → frontal cortex).

• Phasic firing mediates reward and conditioning; tonic firing relates to executive function changes.

How do different drugs raise dopamine in the NAc?

• Stimulants block DAT (cocaine) or release DA (amphetamines)

• Opioids disinhibit VTA DA neurons via MORs on GABA neurons

• Nicotine activates nicotinic receptors on VTA DA neurons

• Alcohol/inhalants act on GABA/glutamate receptors, and cannabinoids act via CB1.

Why does speed of drug delivery matter?

Faster brain entry (e.g. i.v. or smoked) gives a stronger high and stronger reinforcement, likely because abrupt DA increases mimic phasic firing. Larger DA increases are linked to a more intense high.

What is dopamine's role beyond pleasure?

• Driving conditioned responses and incentive motivation.

• DA firing shifts from the natural reward to the conditioned cue, though whether this occurs with drugs is unclear since they act on DA neurons directly.

What is the paradox of the blunted dopamine response?

• Addicted subjects show a markedly blunted DA response to drugs and an attenuated high, yet still seek them.

• Active (non-detoxified) cocaine addicts had ~80% lower response than controls, so withdrawal isn't the explanation.

What is the caveat for the blunted DA finding?

• Low baseline D2R might limit detection of further decreases.

• But addicts do show D2R reductions to cues, so the blunting likely reflects decreased DA release.

What did the cocaine cue video study find?

• Cocaine cues (vs neutral video) increased dorsal striatal DA.

• The increase correlated with craving and addiction severity, and was replicated.

• Dorsal striatum is linked to habit learning, suggesting habits strengthen as addiction progresses.

What shift in dopamine release is proposed in addiction?

From the drug (early) to the conditioned cue (later). The act of taking the drug can itself become a cue.

What is the finding about D2R and prefrontal function?

Addicts have reduced striatal D2R availability lasting months after detox, correlated with lower metabolism in the OFC, ACC and DLPFC (salience attribution, inhibitory control, emotion regulation, decision making).

How does low D2R link to impulsivity, compulsivity and vulnerability?

• Low D2R was linked to impulsivity in methamphetamine abusers and to compulsive cocaine use in rodents.

• People with a family history of alcoholism had higher D2R and normal prefrontal metabolism, suggesting intact PFC may protect against abuse.

How is dopamine linked to motivation in addiction?

• DA regulates vigour, persistence and effort.

• Drug seeking displaces other activities.

• Prefrontal activity is decreased at rest but OFC and ACC activate during craving, and methylphenidate increased ventral ACC/medial OFC metabolism only in addicted subjects.

What are the four key circuits in Volkow's systems model?

• Reward/saliency

• Motivation/drive

• Conditioning/habits

• Inhibitory control/executive function.

• Mood regulation (stress) and interoception (insula, ACC) also participate but are less studied.

What distinguishes the addicted brain in the systems model and what is the result?

• The enhanced expected value of the drug in reward, motivation and memory circuits overcomes the control circuit, creating a positive-feedback loop.

• Result: enhanced drug value, reduced value of other reinforcers, impaired inhibition, and compulsive drug taking.

What other neurotransmitter roles are noted (Volkow)?

• DA changes may be compensatory

• Glutamate modulates DA.

• Impaired glutamate-induced plasticity likely contributes to prefrontal deficits, including inability to learn from adverse consequences.

What does Volkow's multipronged treatment approach aim to do?

• Decrease drug reinforcement

• Enhance natural reinforcers

• Inhibit conditioned associations

• Enhance non-drug motivation

• Strengthen inhibitory control.

What is Everitt & Robbins' (2016) core argument?

Addiction progresses from voluntary, goal-directed drug use to habitual and then compulsive drug seeking, via a goal-directed → habit shift, a ventral → dorsal striatal shift, and reduced top-down PFC inhibition.

How is compulsive behaviour defined, and how does it differ from habit?

• Compulsion is maladaptive persistence in responding despite adverse consequences.

• Habit continues without considering outcome value, so not all habit is compulsive.

What can drug-associated cues trigger?

Craving, drug seeking and relapse, even after prolonged abstinence and even when the drug is absent.

What is goal-directed (action-outcome) behaviour?

Behaviour performed because the person expects a particular outcome. It is sensitive to outcome value, so it reduces if the outcome is devalued.

What is habitual (stimulus-response) behaviour?

Behaviour triggered by cues/contexts through repeated reinforcement, regardless of current outcome value. It is automatic, persistent and resistant to extinction.

Are habits pathological, and can habit alone explain compulsion?

Habits aren't inherently pathological; they're an efficient way of learning. Excessive habit becomes maladaptive, but habit formation alone can't explain compulsive drug seeking.

What is the evidence for habitual drug seeking?

• Cocaine-seeking rats became insensitive to outcome devaluation after extended training.

• Alcohol-dependent humans overrely on habitual learning, and drugs can accelerate habit formation for other rewards.

What are the roles of the key brain regions?

• mPFC supports goal-directed behaviour and cognitive control,

• pDMS supports action-outcome learning

• aDLS supports established habits

• Putamen is the human dorsal striatum linked to habit learning.

How does control shift with prolonged drug use?

• From mPFC and dorsomedial striatum (early, goal-directed) to dorsolateral striatum (later, habitual).

• Blocking DA receptors in the DLS reduces established cocaine seeking in animals.