GT Neuro Seminar Series
“Advancing Brain-machine Interfaces Toward Clinical Viability”
Chethan Pandarinath, Ph.D.
Wallace H. Coulter Department of Biomedical Engineering,
Georgia Tech/Emory University
Brain-machine interfaces (BMIs) aim to restore function for people with disabilities by directly interfacing with the nervous system. A key challenge in advancing these systems is developing frameworks to accurately estimate and perturb the state of the brain in real-time. I will demonstrate the development and application of such frameworks to an intracortical motor prosthesis for people with paralysis. As part of the BrainGate2 pilot clinical trial, we developed advances in neural signal processing, systems design, and algorithms and demonstrated the highest performance 2-dimensional control (Gilja*, Pandarinath* et al., Nature Medicine 2015) and communications rates (Pandarinath*, Nuyujukian* et al., eLife 2017) ever achieved by people with paralysis controlling a BMI. Moving forward, I will highlight ongoing work to precisely understand the dynamics of neural population activity, based around deep learning approaches (Sussillo et al., arXiv 2016), that can dramatically increase our ability to extract information and intention from populations of neurons in the brain. The insights gained from these studies motivate interdisciplinary approaches towards the control of complex end effectors (e.g., dextrous robotic arms) that leverage innovations across neuroengineering and systems neuroscience.
This presentation can be seen via videoconference on the Emory Campus HSRB E160
- Workflow Status: Published
- Created By: Floyd Wood
- Created: 03/27/2017
- Modified By: Fletcher Moore
- Modified: 04/13/2017