Meeting Banner
Abstract #1236

Characterizing Structural MR Brain Changes of Child and Adolescent Bipolar Disorder Patients Using Random Forests Classification

Ping-Hong Yeh1, Hongtu Hongtu Zhu, Mark Nicoletti, Hasan Baloch, John Hatch2, Giovana Zunta-Soares, Jair C. Soares

1Psychiatry, University of North Carolina, Chapel Hill, Chapel Hill, NC, USA; 2Department of Psychiatry, The University of Texas Health Science Center at San Antonio, U.S.A

To investigate important structural brain characteristics in classifying child and adolescent bipolar disorder (BD) patients, we used Random Forests, an improved Classification and Regression Trees machine learning method, for doing disease classifications in pathway-based analysis. Our results indicate that cortical thickness is more accurate than gray mater volume in classifying adolescent BD vs healthy controls. The study provides evidence of brain structural abnormalities within the cortico-striatal-thalamic-cortical and limbic-cortico-striatal-thalamic-cortical circuits in BD.