Microsystems Seminar: Jack Judy, DARPA, "Machine-Brain Interfaces"
Thursday, December 8, 2011
1146 A.V. Williams Building
Microsystems Seminar: Machine-Brain Interfaces
DARPA Microsystems Technology Office
Although the field of brain-machine interfaces (BMIs) is exciting and appears to have great potential, there are two fundamental and well-known obstacles that will prevent BMIs from being deployed widely and with great success until they are addressed head on and solved. Despite decades of focused funding by NIH and NSF, more years of interest, and far more years of demand, high-performance neural interfaces for recording motor-control information are not reliable enough for demanding clinical users, such as wounded warriors.
Building on a long history of innovation in neural-recording interfaces, DARPA has launched a large umbrella progrm to address the key challenges related to transitioning advanced neuroprosthesis technology toward reliable clinical use for amputated servicemembers. The ultimate goal of the Reliable Neural Technology (RE-NET) Program, which is managed by the speaker through the Microsystems Technology Office (MTO), is to develop new technology for extracting and processing information from the nervous system at a scale and rate needed to reliably control modern robotic prostheses over the lifetime of the amputee.
To achieve this goal, DARPA is supporting three broad agency announcements (BAAs): Histology for Interface Stability over Time (HIST), Reliable Peripheral Interfaces (RPI), and Reliable CNS Interfaces (RCI). In this presentation I will describe the motivation and goals of each of the BAAs. These investments by DARPA, outlined below, are directly addressing and solving the fundamental problems that limit the reliability of high-performance neural interfaces, in order to provide amputees with the ability to control state-of-the-art many-degree-of-freedom upper-limb prostheses and achieve their ultimate rehabilitative goals.
DARPA-BAA-10-32: Histology for Interface Stability over Time (HIST), has four technical areas of interest: (1) identify the leading mechanisms of interface degradation and failure; (2) develop new invasive and non-invasive histology methods to gain even greater insight in the assessment of neural-recording interface status and performance; (3) develop accurate predictive models of interface degradation and failure, which should reduce the time required to assess and develop new interfaces; and (4) develop methods to accelerate the degradation and failure of neural interfaces, which again should reduce the time required to assess and develop new robust interfaces. DARPA-BAA-11-08: Reliable Peripheral Interfaces (RPI), has five technical areas of interest: (1) demonstrate peripheral-tissue interfaces that can reliably extract motor-control information; (2) demonstrate peripheral-tissue-interface electronics and packaging to facilitate the reliable extraction of motor-control information; (3) demonstrate algorithms and subsystems that can reliably decode peripheral motor-control intent from recorded signals in real time; (4) combine the breakthroughs derived from efforts in technical areas 1, 2, and 3 to demonstrate the effectiveness of the proposed RPI; and (5) demonstrate technology that can reliably provide direct sensory-feedback signal to the peripheral nervous system.
DARPA-BAA-11-37: Reliable Central-Nervous-System Interfaces (RCI), has five technical areas of interest: (1) demonstrate CNS-tissue interfaces that can reliably extract motor-control information; (2) demonstrate CNS-tissue-interface electronics and packaging to facilitate the reliable extraction of motor-control information; (3) demonstrate algorithms and subsystems that can reliably decode CNS motor-control intent from recorded signals in real time; (4) demonstrate amputee-relevant behavioral-testing methods to accurately evaluate the reliability of CNS-interface systems; and (5) demonstrate technology that can reliably provide direct sensory-feedback signals to the CNS.
Jack W. Judy received the Ph.D. and M.S. degrees from the University of California, Berkeley, CA, in 1996 and 1994 respectively, as well as the B.S.E.E. degree, with summa cum laude honors, from the University of Minnesota, Minneapolis, MN, in 1989. He worked for Silicon Light Machines, Inc., Sunnyvale, CA, an optical-MEMS startup company commercializing a micromachined projection-display technology, from 1996 to 1997. He has been on the faculty of the Electrical Engineering Department at the University of California, Los Angeles, since 1997, where he is currently a Professor. At UCLA he leads research laboratories doing both novel MEMS and neuroengineering research. Since 2009, he has been on leave from UCLA to serve the United States Department of Defense, as a Program Manager at the Defense Advance Research Projects Agency (DARPA) in the Microsystems Technology Office (MTO). At DARPA he created and manages the following three funding opportunities within the Reliable Neural Technology (RE-NET) Program he founded: Histology for Interface Stability over Time (HIST), Reliable Peripheral Interfaces (RPI), and Reliable Central-Nervous-System Interfaces (RIC). The overall goal of the RE-NET program is to understand and overcome the failure mechanisms limiting the truly chronic reliability and ultimate long-term and high-level performance of neural interfaces for amputees.