APS
31st APS Annual Convention · 2019
Predicting Irritability in Youth: A Machine Learning Approach with fMRI Data
- Olga Revzina
National Institute of Mental Health - Christen Deveney
Wellesley College - Joel Stoddard
University of Colorado School of Medicine - Katharina Kircanski
National Institute of Mental Health - Jennifer Yi
The University of North Carolina at Chapel Hill - Derek Hsu
Emory University School of Medicine - Elizabeth Moroney
University of California, Los Angeles - Laura Machlin
The University of North Carolina at Chapel Hill - Laura Donahue
University of Michigan Medical School - Alexandra Roule
The Pennsylvania State University - Gretchen Perhamus
National Institute of Mental Health - Daniel Pine
National Institute of Mental Health - Melissa Brotman
National Institute of Mental Health - Ellen Leibenluft
National Institute of Mental Health - Wan-Ling Tseng
National Institute of Mental Health
Abstract
Using linear Support Vector Regression with activations from 268 nodes across the brain, we predicted childhood irritability (measured by parent- and child-reports) with high performance (r =0.47, root mean square error [RMSE]=2.40). Activations in the prefrontal cortex, caudate, thalamus, parietal lobe, and cerebellum were the most predictive features of irritability.
Neuroscience
