Functional Measures Improve Separation of Marijuana Smokers with Comorbid Psychiatric Disorders from Healthy Controls and Marijuana Smokers Without Psychiatric Disorders in a Three-Group LDA Classification

Piotr Bogorodzki¹, Melissa Lopez-Larson², ³, Jadwiga Rogowska, Deborah Yurgelun-Todd², ³

The objective of this study was to determine whether morphometric or BOLD fMRI features obtained from structural and functional MRI scans could discriminate between three groups of subjects including healthy controls, marijuana using adolescents without comorbid psychiatric illness, and marijuana using adolescents with comorbid psychiatric diagnoses. We found that the combination of morphological and functional features significantly improves classification accuracy and decreases diagnostic misclassification rate. Our findings indicate that diagnostic classification algorithms may be improved with the addition of both morphology and functional measures obtained in fMRI studies.