A student team led by researchers at Duke Surgery and Global Health Institute will further develop the computer application - Alcohol Use Behavioral Phenotyping Test (AUBPT) that can help predict alcohol use and alcohol use disorder risks based on personal characteristics and behavioral performance on Research Domain Criteria paradigms/games. Students will build multi-tasking simulated AI agents using computational neuroscience and deep learning methods. These agents can 'mimic' human behaviors. Students will also use machine learning to model substance use risk from previously collected data. Together, these approaches will give a more rigorous understanding of causal relationships across behavioral paradigms and make AUBPT an adaptive application. This project is a part of the Bass Connections project - 'Alcohol Use Behaviors across Countries and Cultures' (https://bassconnections.duke.edu/project-teams/alcohol-use-behaviors-across-countries-and-cultures-2021-2022). The Bass Connections team is highly interdisciplinary where students are working at the interface of digital, mental, and global health to deployment, evaluation, and cultural adaptation of our AUBPT across multiple global samples. The Bass Connections team is highly interdisciplinary where students are working at the interface of digital, mental, and global health for deployment, evaluation, and cultural adaptation of AUBPT across multiple global populations.
Project Lead: Siddesh Zadey, Dr. Catherine Staton and Dr. Joao Vissoci