Meeting Banner
Abstract #1645

Disease State Prediction from Resting State Functional Connectivity

R. Cameron Craddock1,2, Paul Holtzheimer2, Xiaoping P. Hu3, Helen S. Mayberg2

1School of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA, USA; 2Dept. of Psychiatry and Behavioral Sciences, Emory University, Atlanta, GA, USA; 3Dept. of Biomedical Engineering, Georgia Institute of Technology and Emory University, Atlanta, GA, USA


Support vector classification is applied to predict disease state (MDD) from resting state functional connectivity. Additionally two feature selection methods are proposed that score features based on reliability. The resulting classifier was able to distinguish MDD from controls 100% of the time. The two reliability based feature selection algorithms outperform t-test filter and recursive feature elimination methods.