03-04-2026, 12:50 background module 2: Addiction and Compulsive Disorders

background module 2
Goal-directed drug seeking
In Module 1, we discussed the central role of craving in addiction. The idea that craving fuels
instrumental drug-seeking behavior aligns with the perspective that drug use is goal-directed and
that individuals with addiction have heightened sensitivity to the rewarding effects of substances,
relative to other rewards such as social interaction or food. Drug-seeking behaviour is considered
goal-directed when it is guided by expectations (cognitive criterion) and a positive (hyper)valuation
of the drug (motivational criterion). This account is supported by empirical evidence. Drug use is
commonly linked to positive subjective experiences, favorable expectations about the substance,
and a willingness to invest resources to obtain it. Common goals subserved by substance use
are: feeling great (i.e., recreational use), feeling well (i.e., relief from a negative state; affect
regulation), fitting in (i.e., using the substance facilitates social interactions), and doing better (i.e.,
improve cognitive or physical performance)(NIDA, 2011). In line with a goal-directed account,
experimental studies using instrumental tasks show that individuals with addiction exert more
effort to obtain drug rewards than non-addicted individuals (for a review, see Hogarth & Chase,
2012).

Why then does problematic substance use continue even when people acknowledge that it
contradicts their long-term goals? Within this framework, problematic substance use should be
considered in the light of goal conflict, with the highly valued short-term effects of drug use (i.e.,
feeling great/feeling well/fitting in/doing better) outweighing longer-term goals like maintaining
health (e.g., Buabang et al., 2025).

Habitual drug seeking

However, drug seeking may not always be a goal-directed behaviour. The habit account of
addiction proposes that although drug seeking initially involves goal-directed processes, it can
gradually become governed by abnormally strong habits. According to Thorndike’s Law of Effect,
habits develop when behaviours are reinforced either by rewarding outcomes (positive
reinforcement) or by the removal of an anticipated negative state (negative reinforcement). Such
reinforcement strengthens stimulus–response (S–R) associations between environmental stimuli
(or "cues") and behavior. As a result, encountering these cues can automatically elicit the learned
behavior, even in the absence of strong motivation for the expected outcome. Dual-process
theories suggest that with repeated practice, these S–R associations can come to dominate
behavior, leading to a shift from flexible, goal-directed control to habitual responding. , drug
seeking initially starts out as goal-directed but becomes increasingly driven by aberrantly strong
habits. As a result, encountering these cues can automatically elicit the learned behavior, even in
the absence of strong motivation for the expected outcome. Dual-process theories suggest that
with repeated practice, these S–R associations can come to dominate behavior, leading to a shift
from flexible, goal-directed control to habitual responding.

[Link] 1/5
03-04-2026, 12:50 background module 2: Addiction and Compulsive Disorders

Measuring habits
The outcome-devaluation test is the experimental paradigm that is used to determine whether a
certain behavior is goal-directed or habitual. it will be extensively covered in the lecture, but the
basic set-up of the animal version is briefly described here. This paradigm consists of 3 phases:

Phase 1: this is the instrumental learning phase. For example, a rat is trained to perform a
response (e.g., press a lever) to obtain alcohol from a fluid dispenser in a Skinnerbox (i.e.,
operant chamber). This training usually takes place during several consecutive days.

Phase 2: this is the outcome-devaluation phase. The rat is removed from the Skinnerbox, and
receives unlimited access to alcohol in a separate cage. This way, the alcohol outcome is
"devalued" through satiation. An alternative devaluation method is to pair alcohol with Lithium-
Chloride induced nausea to condition an aversion to alcohol.

Phase 3: this is the critical test phase. The rat is returned to the Skinnerbox and has the
opportunity to perform the response (e.g., press the lever) again (e.g., during a 5-min test, and the
number of lever presses are counted). If pressing the pedal is a goal-directed action, the rat
should adjust it's behavior immediately after devaluation, as reflected in reduced lever pressing.
The rat anticipates that the response will be followed by alcohol, which no longer serves as a goal.
If, on the other hand, the behavior has become a habit (e.g., after extensive instrumental training
in phase 1), the contextual stimuli in the Skinnerbox would simply reactivate the response through
a stimulus-response (S-R) association, even when the alcohol outcome is no longer valuable.
Consequently, the rat should be unable to immediately flexibly adapt its behaviour. It should
continue to press the lever for the devalued alcohol.

Importantly, this (short) test phase is carried out in extinction, which means that the falcohol is no
longer offered when the rat presses the lever. The reason for this is that otherwise we cannot
determine whether lever pressing is goal-directed behavior or a habit. Indeed, according to the
Law of Effect, habits are strengthened when a response is followed by reward, but also weakened
when it is nót followed by a rewarding outcome (or even by an aversive outcome). Offering a
devalued (alcohol) outcome during the testing phase would therefore be predicted to gradually
lead to weaker S-R associations (compared to when a valuable outcome is offered) and
consequently reduced lever pressing. Therefore, if the test is not performed in extinction, it is
impossible to tell what it means when a rat presses less for the devalued reward. That could be
because the behavior is goal-directed, but it could also be a consequence of habit learning. For
this reason, the outcome-devaluation test is conducted in extinction: if the rat then immediately
(without further opportunity for learning) adjusts it's behaviour, we know that it acts in a goal-
directed manner (based on anticipation and evaluation of the outcome). (PS - a drawback of the
extinction procedure is that the behavior will extinguish. Extinction tests are therefore usually kept
short).

[Link] 2/5
03-04-2026, 12:50 background module 2: Addiction and Compulsive Disorders

Figure: illustration of outcome-devaluation paradigm when applied to study the goal-directed/habitual status of
alcohol-seeking in rats (from Corbit & Janak, 2016)

The habit theory of addiction

According to the habit theory of addiction, drug-seeking behavior is initially goal-directed, but
quickly turns into a habit. In the article by Berridge & Robinson (2011), you have read that they do
not view habits as an important factor in compulsivity, but other researchers have a different take
on this. Two prominent researchers in the field of addiction, Everitt & Robbins, see the transition
from goal-directed control to drug habits as crucial to the development of compulsive drug-seeking
behavior. Central to this account are three ideas: (1) there are individual differences in the
tendency to form dominant habits, and with a strong tendency, the person in question is
vulnerable to developing an addiction (or other compulsive behaviours); (2) drugs are extremely
strong reinforcers of S-R habits (through their effect on the nigrostriatal dopamine system); and
(3) drugs lead to structural changes in the brain, that lead to a generally stronger tendency to rely
on habitual control (including detrimental effects on prefrontal areas involved in cognitive and
goal-directed control functions). These different ideas are illustrated in the Figure below, and will
be discussed in the lecture of Module 2.

[Link] 3/5
03-04-2026, 12:50 background module 2: Addiction and Compulsive Disorders

Dual-process models
Drug habits are best explained within dual-process models, which propose that addiction arises
from disruptions in two interacting systems. On the one hand, automatic, reflexive, and impulsive
bottom-up processes become excessively reinforced; on the other hand, top-down cognitive,
reflective, and executive control processes are weakened. As a result, reduced cognitive control
increases reliance on habitual behaviour. Support for dual-process models comes from
neuroscientific evidence demonstrating that goal-directed actions and habits are mediated by
distinct neural systems. These mechanisms will be addressed in both the lecture of Module 2 and
the assigned readings (de Wit, 2017). The role of impaired executive functions - specifically
cognitive control processes - is further discussed in the knowledge clip in the Study Material.

Note: The precise definition of the two processes varies across dual-process models. However,
Pavlovian conditioning and the development of instrumental habits are typically considered
bottom-up processes, whereas goal-directed control and executive functions—such as error
monitoring, decision making, and impulse control—are classified as top-down processes.

Stigma
Finally, a note on stigma in addiction. In tutorial 1, you considered stigmatization in addiction.
Stigma refers to a set of negative and unfair beliefs that a society or group holds about a
particular issue or population, and it plays a central role in shaping how substance use disorders
are perceived and treated.

[Link] 4/5
03-04-2026, 12:50 background module 2: Addiction and Compulsive Disorders

In psychology and clinical practice, reducing stigma begins with language. The American
Psychological Association (APA) recommends person-first expressions (e.g., “person with a
substance use disorder”), avoiding moral judgments such as “abuser” or “they chose this,” and
using neutral, clinical terms like “in recovery” rather than “ex-addict.” Such language should
emphasize individuals’ agency and potential for recovery, highlight treatment options, and avoid
deterministic statements like “once an addict, always an addict.” These guidelines are not merely
semantic, they actively contribute to more respectful care, greater help-seeking, and improved
outcomes.

Importantly, stigma is not uniform but shaped by multiple interacting factors. The type of
substance matters: stimulant use (e.g., cocaine) is often viewed as dangerous and reckless
(aligned with a moral model), whereas alcohol use tends to be more culturally accepted and
framed as problematic yet understandable (closer to a brain disease model). Similarly, addiction to
prescription drugs is often perceived as “not their fault.” Visibility also intensifies stigma—
behaviors such as injection use, intoxication, or homelessness increase negative judgments.
Social context further complicates this picture: class, race, and marginalization can lead to
“double stigma,” where individuals are judged for both for their substance use and for another
stigmatized identity, with these stigmas reinforcing one another (e.g., “a woman who uses drugs is
irresponsible and unfit to be a mother” or “he is an addict from the streets, so what else would you
expect?”). Cultural norms also shape meaning—for example, psychoactive substance use by
shamans may be viewed as sacred within certain communities.

Models of addiction that reduce moral responsibility can decrease stigma by others, thereby
potentially increasing treatment seeking because one is less afraid to face judgement. On the
other hand, these same models may also inadvertently lower perceived agency/self-efficacy,
potentially discouraging treatment seeking. At the same time, these models guide intervention
strategies; for instance, recognizing social and structural drivers of addiction (i.e., environment
and policies regarding inequality, poverty, racism, and access) underpins approaches like
Housing First, which provide immediate housing without requiring abstinence. Overall, stigma is
a complex, multidimensional phenomenon: different substances, models, and social contexts
produce distinct “stigma profiles,” rather than simply more or less stigma.

[Link] 5/5