Student: Leo Brueggeman, Graduate Student in Medicine and Genetics, University of Iowa
Faculty Advisor: Jacob Michaelson
Using Deep Learning to Predict Psychiatric Traits
My research focuses on using artificial intelligence to find biomarkers for cognitive and psychiatric traits from medical imaging data. Recently, this imaging data (e.g., retina scans, brain MRIs) has become available in large enough numbers (tens to hundreds of thousands) to be feasible for use by high powered deep learning algorithms, similar to those powering self-driving cars. 
Through applying these algorithms, we are developing methods to reliably predict an individual’s cognitive performance and/or psychiatric disease risk from images of their retina and/or brain. From this research, we are gaining an increased understanding of what structural changes in the retina and brain are indicative of mental states relating to performance and disease. In addition to increasing our understanding of the relationship between the brain and behavior, the long term implications of this research may help guide earlier diagnosis and intervention for psychiatric disease.
Through applying these algorithms, we are developing methods to reliably predict an individual’s cognitive performance and/or psychiatric disease risk from images of their retina and/or brain. From this research, we are gaining an increased understanding of what structural changes in the retina and brain are indicative of mental states relating to performance and disease. In addition to increasing our understanding of the relationship between the brain and behavior, the long term implications of this research may help guide earlier diagnosis and intervention for psychiatric disease.