Caroline Matthis: Neurobehavioural patterns of alcohol abuse in adolesence

Excessive alcohol consumption has a detrimental effect on public health. Alcohol abuse is a top-ranked disorder of the brain with respect to total costs to economy and is linked to an estimated 3.8 % of global deaths. Often, first experiences with alcohol are made during adolescence, the time of transition between childhood and adulthood. Adolescence marks a period of complex social, biological and psychological development; the interplay between alcohol consumption and these manifold developments are not yet fully understood. In my thesis, I use the IMAGEN database, a large-scale, longitudinal study of around 1000 healthy adolescents, to shed light on the neurobehavioural patterns of heavy drinking in adolescence.

In the first part of my thesis, I introduce a framework for the systematic analysis of predictive and explanatory quality of neurobehavioural features. I adapt a common cross-validation scheme to assess the predictive power in the presence of confounders. Moreover, I develop a novel method to estimate the influence of a single (neurobehavioural) feature on a binary outcome.

In the second part of my thesis, I build a comprehensive phenotype describing alcohol drinking behaviour for the subjects from the IMAGEN database. I show that the constructed phenotype outperforms all original drinking behaviour variables in terms of consistency.

In the third part, I apply the established methods on neurobehavioural features from the IMAGEN study to differentiate between drinking behaviour phenotypes. First, I focus on grey matter volume and psychosocial features. I show that heavy drinking in adolescence is associated with reduced grey matter volume across various cortical and subcortical structures, especially in females. Moreover, I observe that impulsivity and facets of novelty seeking are associated (also longitudinally) to heavy drinking. Then, I focus on functional imaging and cognitive features. I show that adolescent drinking is associated with various markers of impaired response inhibition, both neural and behavioural.

Overall, I shed light on both risk profiles and potential consequences of heavy drinking in adolescence. On the one hand, I contribute to the understanding of the detrimental effects of alcohol on the adolescent brain, calling for targeted intervention programs. On the other hand, I show that risk profiles based on personality traits may offer the potential for prevention procedures before the treacherous spiral of addiction begins.