In many functional MRI experiments, structural MRI scans of the brain are customarily collected for co-registration but are otherwise left unanalyzed. Given the intricate relationship between structure and function, this standard practice represents a missed opportunity to deepen the understanding about the brain. In this seminar, Dr. Fong will discuss his ongoing work in leveraging convolutional neural networks (CNNs) for predicting the cognitive health, based on the structural MRI images made available by the Human Connectome Project. He will focus on two aspects: (1) comparisons between single- and multi-channel approaches of CNN model construction in estimating the brain age; and (2) the prediction of cognitive functions of individuals across the adult lifespan. In addition, he will provide preliminary indications of how well these models, trained on foreign data, generalize to the data collected here at the UBSN MRI Facility.