Home | People | Research | Blog | Software

Brain Networks Laboratory

Our lab is part of the Cognitive Neuroscience Division in the Department of Neurology, Columbia Medical Center. Our research consists of both methodological and applied neuroimaging analysis using multivariate and multimodal machine-learning techniques. A major focus of the applied and developed tools is the practical utility for clinical and basic cognitive neuroscience. Multivariate decompositions (like Principal/Independent Component Analysis) are used in conjunction with simple Machine-Learning tools (Naïve Bayes, k-nearest-neighbor classifier, Linear and Quadratic Discriminant classifier, Support Vector Machines) and resampling techniques (permutation test, bootstrap test) for statistical inference. The data modalities include structural neuroimaging (white-matter integrity, cortical thickness, etc.) and functional activation and connectivity during rest and task-performance. In our community-based cohort sampled from the whole adult life span (age 20-80) we can ask detailed questions about age-related changes in cognition and their attendant neural mechanisms.