Alcohol misuse during adolescence (AAM) has been associated with disruptive development of adolescent brains. In this longitudinal machine learning (ML) study, we could predict AAM significantly from brain structure (T1-weighted imaging and DTI) with accuracies of 73 - 78% in the IMAGEN dataset (n ~1182). Our results not only show that structural differences in brain can predict AAM, but also suggests that such differences might precede AAM behavior in the data. We predicted ten phenotypes of AAM at age 22 using brain MRI features at ages 14, 19, and 22. Binge drinking was found to be the most predictable phenotype. The most informative brain features were located in the ventricular CSF, and in white matter tracts of the corpus callosum, internal capsule, and brain stem. In the cortex, they were spread across the occipital, frontal, and temporal lobes and in the cingulate cortex. We also experimented with four different ML models and several confound control techniques. Support Vector Machine (SVM) with rbf kernel and Gradient Boosting consistently performed better than the linear models, linear SVM and Logistic Regression. Our study also demonstrates how the choice of the predicted phenotype, ML model, and confound correction technique are all crucial decisions in an explorative ML study analyzing psychiatric disorders with small effect sizes such as AAM.

Research on argumentation mining from text has frequently discussed relationships to discourse parsing, but few empirical results are available so far. One corpus that has been annotated in parallel for argumentation structure and for discourse structure (RST, SDRT) are the 'argumentative microtexts' (Peldszus and Stede, 2016a). While results on perusing the gold RST annotations for predicting argumentation have been published (Peldszus and Stede, 2016b), the step to automatic discourse parsing has not yet been taken. In this paper, we run various discourse parsers (RST, PDTB) on the corpus, compare their results to the gold annotations (for RST) and then assess the contribution of automatically-derived discourse features for argumentation parsing. After reproducing the state-of-the-art Evidence Graph model from Afantenos et al. (2018) for the microtexts, we find that PDTB features can indeed improve its performance.

Alcohol misuse during adolescence (AAM) has been linked with disruptive structural development of the brain and alcohol use disorder. Using machine learning (ML), we analyze the link between AAM phenotypes and adolescent brain structure (T1-weighted imaging and DTI) at ages 14, 19, and 22 in the IMAGEN dataset (n~1182). ML predicted AAM at age 22 from brain structure with a balanced accuracy of 78% on independent test data. Therefore, structural differences in adolescent brains could significantly predict AAM. Using brain structure at age 14 and 19, ML predicted AAM at age 22 with a balanced accuracy of 73% and 75%, respectively. These results showed that structural differences preceded alcohol misuse behavior in the dataset. The most informative features were located in the white matter tracts of the corpus callosum and internal capsule, brain stem, and ventricular CSF. In the cortex, they were spread across the occipital, frontal, and temporal lobes and in the cingulate cortex. Our study also demonstrates how the choice of the phenotype for AAM, the ML method, and the confound correction technique are all crucial decisions in an exploratory ML study analyzing psychiatric disorders with weak effect sizes such as AAM.

No abstract