Abstract:
Neural engineers try to fight disease by interfacing directly with the nervous system to correct or circumvent the pathology, but the complexities of neurophysiology make designing these therapies very difficult. The sophisticated interactions of neural control systems with the body and mind has limited our ability to predict the effects of neural interfaces in advance, which forces the field to rely on painstaking animal experiments to test and design the devices. This work presents an alternate way forward using new functionally predictive hybrid models that let us prototype and test neural interfaces in cheap, low-risk settings to get the most out of the cumbersome and expensive animal and human studies that require implanted electrodes. The key innovation is to combine parts of traditionally separate model systems (animal, human, mathematical, or statistical) into hybrid frameworks that use the strengths of each type to make accurate, use-case predictions about the effects of new neural interfaces.
Details:
DATE: Thursday, December 2, 2021
TIME: 4:00-5:00 pm
LOCATION: Cooley G906
ZOOM LINK: https://umich.zoom.us/j/97723483179