ABSTRACT
Alcohol impairs a person’s level of consciousness and modifies certain EEG signal patterns in the brain. The analysis of EEG signals has been established as a popular method of distinguishing between people with alcohol use disorder (AUD) and nonalcoholics. Machine learning (ML) models such as support vector machines, random forest, neural networks (NNs), decision trees, and multilevel wavelet packet entropy have been developed for this purpose. In addition to diagnosing AUD, these models can be used to evaluate the likelihood of relapse and the effectiveness of treatment. This chapter aims to provide a comprehensive overview of the recent studies that focus on the applications of ML for detection of alcoholism. Logistic regression models have demonstrated the best performance in patient screening. For diagnostic purposes, combinations of different biomarkers along with decision trees or NNs have shown promising results. These diagnostic models appear to perform better than ML models using unstructured imaging data from magnetic resonance imaging and computed tomography scans. The application of ML models for automatic AUD diagnosis provides a cost-effective solution for larger populations, enabling anonymous diagnosis and patient-specific treatment of AUD to alleviate the social stigma associated with the disorder.
