The rapid growth of neuroscience data offers an unprecedented opportunity to build models rich enough to capture the complexity of brain function. However, much of our modeling has focused on single tasks and understanding the brain in relatively simple contexts. In this talk, I will present our efforts to develop large-scale models for neural data that integrate data across diverse tasks, brain regions, and species. In both motor and visual decoding tasks, we show that scaling to many sessions of data unlocks entirely new capabilities, enabling significant advances in brain decoding and cell type prediction that are unattainable with small or single-animal datasets. This work lays the foundation for a new era of integrative neural data analysis and next-generation brain-machine interface technologies.
Biosketch:
Eva Dyer (she/they) is an Associate Professor in the Department of Biomedical Engineering at the Georgia Institute of Technology. Dr. Dyer leads the NerDS Lab which focuses on data-centric AI representation learning and AI for science. A key area of the lab’s research focuses on AI for neuroscience where they aim to develop tools to better understand the brain and neural computation and to uncover abstractions of natural intelligence for creating new brain-inspired AI. Eva earned all of their degrees in Electrical & Computer Engineering including a Ph.D. and M.S. from Rice University and a B.S. from the University of Miami. Eva has received numerous honors including a Sloan Fellowship NSF CAREER Award McKnight Foundation Technological Innovations in Neuroscience Award and CIFAR Azrieli Global Scholar Award.