Development of automatic and controlled processes in relation to addictive behaviours in youth

Addictive behaviors are a prime cause of death and economic loss in the Western
world. Recently, scientist from the W.H.O. estimated that addictive behaviors
account for 12.4% of deaths worldwide (Goldman, Oroszi & Ducci, 2005). Alcohol
is the primary drug of choice among adolescents, both in the U.S. and in The
Netherlands. In fact, Dutch 15-16 year-olds are the heaviest-drinking of
European youths (Hibell et al., 2004). The second major addictive behavior
among adolescents concerns smoking. Smoking is a major preventable cause of
premature death and disability throughout the world. In the Netherlands around
28% of the adult population are smokers; 38% are considered heavy smokers
(defined as smoking more then 18 tobacco units a day). Although prevalence of
smoking decreases in adults, this is not the case in adolescents (Stivoro,
2006). During the past decade, neurobiological research has made great progress
in delineating the neural adaptations that play a role in different phases of
addictive behaviors. Although there is discussion about the relative importance
of different processes in different phases of addictions (Everitt & Robbins,
2005; Robinson & Berridge, 1993; 2003), there is broad consensus about
different neuro-adaptations:
- sensitization (i.e. a progressively stronger psychomotor stimulant reaction
with repeated use of a drug, mediated by the mesolimbic dopaminergic system,
Robinson & Berridge, 1993; 2003);
- increasing automatization of addictive behaviors (Everitt & Robbins, 2005);
- decreasing controlled processing ability as a result of prolonged alcohol or
drug use (e.g., Giancola, 2000),
particularly inhibitory control (e.g., Fillmore & Vogel-Sprott, 1999, 2000).
- decreases in motivation to perform alternative behaviors than alcohol-use as
a result of prolonged alcohol-use (e.g., Kalivas, & Volkow, 2005).
In addition, recent animal studies consistently reveal that alcohol exposure
(especially at intoxicating doses) impairs cognitive abilities more in
adolescents than in adults (e.g., White et al., 2000; White & Swartzwelder,
2004) and that this impairment is long-lasting. Moreover, chronic alcohol
involvement in adolescence can have negative consequences for the maturation of
a number of brain regions, including structures with extensive
projections to prefrontal cortical areas that support executive control
functions (e.g., DeBellis et al., 2000; White & Swartzwelder, 2004).
Note that most of the neurobiological research in addiction is primarily based
on animal models and that it is unclear to what extent this translates to
humans. For example, there are indications that alcohol abuse during
adolescence in humans also results in deficits in executive abilities (e.g.,
Brown et al., 2000) and/or abnormal patterns of brain activity during the
performance of executive cognitive tasks (Tapert et al., 2004), but the problem
of the interpretation of these studies is that there is also evidence that
deficits in executive functions can be a premorbid characteristic of risk for
alcoholism or other addictions (Peterson et al., 1992; Wiers et al., 1998).
Note further that in addition to deteriorative effects of prolonged alcohol
abuse on executive functions, there is also evidence that the acute effects of
alcohol and many other drugs are a deterioration of executive control
functions, while implicit associative processes are left unaffected and can
even be primed by alcohol use (Fillmore & Vogel-Sprott, 2006). In particular
prospective research can provide insight into these mechanisms as this allows
the investigation of transitions in alcohol and cigarette use, controlling for
previous use and all kinds of confounders and third variables (for instance
family history of drinking).
Recently, Wiers and colleagues (Wiers, Bartholow, et. al., 2006) reviewed the
literature on human development and proposed an integrative model for the
development of alcohol and drug use during adolescence (see also Wiers & Stacy,
in press). Based on a number of recent studies involving adolescents and young
adults, it is proposed that sensitization in humans results in a strengthening
of largely automatic appetitive responses, which results in an attentional bias
for the substance involved (e.g. Franken, 2003), automatic associations between
the substance and positive-arousal (e.g. Houben & Wiers, 2006; Wiers et al.,
2002; 2005) and in automatically triggered approach-action tendencies (Palfai &
Ostafin, 2003). Meanwhile, the executive functions are expected
to be negatively affected by prolonged alcohol/drug use.
Almost all research involving implicit cognitive processes in addictive
behaviors so far has been cross-sectional or used short-term follow-ups
(typically one month, e.g., Stacy, 1997; Wiers et al., 2002). In these studies
it was found that both implicit and explicit cognitions uniquely contributed to
the prediction of later alcohol/drug use.
Some recently performed studies also found evidence for the proposed moderating
role of executive functions on the effect of automatic associations on
(concurrent) alcohol and cigarette use (Grenard et al., 2006): automatic
associations positively predicted alcohol use and smoking in adolescents
scoring low on a working memory task, but not in adolescents scoring high on a
working memory task. We recently found a similar pattern of results using a
different measure of automatic associations (unpublished data). In another
recent study in young adolescents, we found that automatic alcohol-arousal
associations and explicit negative expectancies predicted alcohol involvement
in boys a year later (Thush & Wiers, resubmitted). Both findings are in line
with the model outlined above: we hypothesized that moderation of alcohol use
is a function of motivation to moderate drinking, which is predicted by
negative drinking experiences that translate into explicit negative
expectancies (cf. Jones & McMahon, 1998).
The model is primarily focused on the development of alcohol misuse in
adolescents, but given the partially overlapping risk-factors and interactions
between smoking and alcohol misuse, we also assess smoking associations and
smoking behavior. In the animal literature, there is much evidence for
cross-sensitization, i.e. once the mesolimibic dopaminergic system gets
sensitized by one substance (alcohol, nicotine, other
substances), sensitization develops more rapidly for the second substance
(alcohol, nicotine, other substances, Robinson & Berridge, 2003; Schoffelmeer
et al., 2002). Hence, we expect that the sensitized response to alcohol and
smoking develops more rapidly for individuals who smoke and (binge-) drink. The
negative effects on the
ability to self-regulate the sensitized action-tendencies are thought to be
primarily related to alcohol misuse (there is evidence for negative effects of
alcohol misuse on the development of executive functions, we are not aware of
such data for nicotine).

In the research project proposed here, we assess all important elements of the
model in a longitudinal set-up, with four assessments, separated by 6 months.
This approach will allow us to study the different hypothesized mechanisms in
the development of addictive behaviors in human adolescents, with a focus on
alcohol use and smoking. Specifically, the research will generate answers to
the following main hypotheses:
H.1 - premorbid differences in executive functioning predict later alcohol
involvement and smoking
H.2 - after initiation of binge-drinking/smoking automatic appetitive processes
increase in strength
H.3 - we expect this to happen especially fast in individuals who develop a
pattern of smoking and bingedrinking
due to cross-sensitization
H.4 - after initiation of binge-drinking executive functions development is
impaired (that is, people who do not
start to binge-drink are expected to show a stronger increase in scores on
EFs). Given the absence of research
findings, we do not have specific hypotheses concerning the effects of nicotine
on EFs, after controlling for the
effects of alcohol involvement.
H. 5 - We expect that in individuals scoring low on Executive Functions, the
associations between alcohol
and/or smoking with arousal are stronger predictors of prospective
drinking/smoking than in individuals scoring
high on Executive Functions (Moderation, cf. Grenard et al., 2006)
H.6 - negative experiences with alcohol are hypothesized to predict negative
explicit expectancies and
motivation to change drinking.
H.7 - heavy alcohol-involvement is hypothesized to decrease motivation for
alternative behaviors (e.g., sports,
hobbies) and on social relationships.

The general approach taken is a cohort-sequential design. Based on a
power-analysis, we include 120 12-yearolds, 120 13-year-olds, 120 14-year-olds,
and 120 15-year-olds at T0. We aim for an approximately equal representation of
boys and girls. The two youngest cohorts will be most informative regarding
initiation of addictive behaviors, the two oldest samples for investigating the
escalation of addictive behaviors. Participants are requested to take part in
the study which will take four times approximately one hour of testing (one
hour extra of background variables at T0), separated by 6 months intervals.

Like previously described, there is no risk associated and only a minor burden.
Participants only fill in questionnaires and participate in computertasks every
session and donate some saliva once. This is done by spitting in oragene boxes.