Title:  Data Driven Approaches to Address Inaccurate Nosology in Mental Health from Neuroimaging Data

Committee: 

Dr. Calhoun, Advisor    

Dr. Anderson, Chair

Dr. Keilholz

Abstract: The objective of the proposed research is to identify biological biomarkers for the current psychosis disorders from structural and functional MRI data using data-driven and data-informed machine learning and deep learning approaches. We plan to investigate the relationship between the current categorization and learned features from different classes. To do this, we propose and implement machine learning and deep learning frameworks to capture homogeneous features from different modalities in multi-class analysis. In addition, we apply data cleansing and label noise robustness methodologies on neuroimaging and biological data across psychosis (e.g., structural and functional brain conditions for different brain and mood disorders).