Anxiety and Mood Risk Algorithms

Anxiety and Mood Risk Algorithms (K01 MH106710)

This project used publicly available psychiatric epidemiological datasets (e.g., NCS-R, NESARC) and machine learning methods (e.g., super learning) to develop risk algorithms that predict the onset and chronic course of DSM-defined posttraumatic stress disorder, generalized anxiety disorder, major depressive disorder, and bipolar disorder. This project also involved collecting data to develop additional machine learning risk algorithms using other measures and units of analysis. Participants were recruited through Amazon Mechanical Turk (MTurk) to complete detailed surveys assessing current and past anxiety and mood symptoms and associated risk factors (e.g., personality traits, traumatic events, childhood experiences and adversities, recent stress). Several follow-up surveys occurred over a one-year period. In addition to the surveys, participants completed neuro-cognitive tasks (e.g., assessing emotion recognition, decision-making) through TestMyBrain.

Publications

Jiang, T., Gradus, J. L., & Rosellini, A. J. (2020). Supervised machine learning: A brief primer. Behavior Therapy, 51, 675-687. PubMed

Rosellini, A. J., Liu, S., Anderson, G. N., Sbi, S., Tung, E., Knyazhanskaya, E. (2020). Developing algorithms to predict adult onset internalizing disorders: An ensemble learning approach. Journal of Psychiatric Research, 121, 189-196. PubMed

Rosellini, A. J., Dussaillant, F., Zubizarreta, J. R., Kessler, R. C., & Rose, S. (2018). Predicting posttraumatic stress disorder following a natural disaster. Journal of Psychiatric Research, 96, 15-22. PubMed; Media Coverage

Rosellini, A. J., Liu, H., Petukhova, M. V., … & Kessler, R.C. (2018). Recovery from DSM-IV posttraumatic stress disorder in the WHO World Mental Health surveys. Psychological Medicine, 48, 437-450. PubMed