Bing W. Brunton, Ph.D., Associate Professor: Natural Decoding in the Wild
SEMINAR (In-Person + Zoom)
This CBE Seminar will be hosted in person as well as online via Zoom. RSVP to receive the zoom link by emailing info@bioengineering.ucsb.edu.
Zoom will open after the host has joined at the start of each seminar. You can ask questions through the chat forum and by raising your "hand" and the speaker will call on you.
Speaker
Bing W. Brunton, Ph.D.
Associate Professor
Department of Biology
H. Stewart Parker Endowed Faculty Fellow
Adjunct, Computer Science & Eng. and Applied Math
Data Science Fellow of the eScience Institute
Weill Neurohub Investigator
Graduate Program in Neuroscience
University of Washington, Seattle WA
Faculty Host
Ryan Stowers
9:00 am in Bldg. ESB Room #1001
Natural Decoding in the Wild
Abstract
Scientists, engineers, and speculative fiction authors have long imagined possible futures when people interact meaningfully with machines directly using thought. Technologies that interpret brain signals to control robotic and virtual devices have tremendous potential to assist individuals with physical and neurological disabilities, to augment engineered systems integrating humans in the loop, and to enhance one’s daily life in an increasingly information-rich world. One key challenge is how neural decoding may be approached 'in the wild,' where sources of behavioral and recording variability are significantly larger than what is found in the lab. In this talk, I will highlight a set of studies where we leveraged recent advances in machine learning and computer vision to study intracranial electrophysiological recordings from human participants recorded during thousands of spontaneous, unstructured arm movements, observed over up to a week for each participant. We found considerable individual variability in both movement behaviors and their neural correlations. Further, we make use of this multi-modal, large-scale dataset (now published in NWB format) to develop several novel approaches to robust neural decoding, including learning self-supervised decoders without labels.
BIO
Bing W. Brunton is an H. Stewart Parker Faculty Fellow at the Department of Biology at the University of Washington. She currently holds affiliations with the eScience Institute for Data Science, the Paul G. Allen School of Computer Science & Engineering, and the Department of Applied Mathematics. After studying biophysics at Caltech and then computational and systems neuroscience at Princeton, Brunton expanded her expertise into applied mathematics, dynamical systems, and neuroengineering at the University of Washington.
Now, the Brunton lab works at the intersection of data analysis and neuroscience, and helps develop the data-driven analytic methods that are applied to neuroscience. The Brunton Lab is particularly interested in uncovering spatiotemporal patterns in high-dimensional, time-series data, especially exploring the neural basis of naturalistic movements in diverse animals. Brunton took the time to develop a strong record working collaboratively as a part of interdisciplinary teams, as well as mentoring early-career researchers interested in advancing both methods development and neuroscience. Brunton's work has been recognized by the Alfred P. Sloan Foundation Fellowship in Neuroscience (2016), a UW Innovation Award (2017), a Young Investigator Program award from the Air Force Office of Scientific Research (2018), and as a Weill Neurohub Investigator (2020) and a Moore Distinguished Scholar for visiting faculty at Caltech (2021).
