Francisco J Valero-Cuevas
Associate Professor of Biomedical Engineering & Biokinesiology and Physical Therapy
Associate Professor of Computer Science & Aerospace and Mechanical Engineering
The University of Southern California, Los Angeles, CA.
While neuro-prosthetics has received much attention recently, current technological advances can also be applied to reverse engineering an older and equally pressing problem. Namely, versatile sensorimotor function. My work has been devoted to extending our technical and analytical tools to confront this age-old reverse engineering problem. I will describe some of our core technologies including innovative physiological recordings from the hand, muscles and brain; novel experimental paradigms for manipulation; a unique mechatronic system to drive the tendons of actual (unembalmed) cadaveric hands; data-driven model inference; and a mathematical/computational framework to understand motor versatility, muscle redundancy, and robustness to muscle dysfunction.
How does the brain achieve robust function by controlling a nonlinear plant with sluggish actuators and noisy sensors to a level that is unsurpassed by engineered systems? By presenting several examples of successful application of these technologies, I will illustrate our current understanding of the neuro-mechanics of manipulation, future directions and cross-fertilization efforts to other areas of engineering, neuroscience and medicine.
While brain-body coevolution as a mechanisms for achieving this performance has face validity, I present work in my lab aimed at establishing the specific contribution by both passive (i.e., tissues) and active (i.e., nervous system and neural strategies) that make such functional performance possible. We find that, for example, (i) the nervous system uses a time-critical strategy to transition between neural controllers for the control of motion and control of force during grasp acquisition (the alternative would be a strategy that relies on finger impedance), (ii) that muscle redundancy (the so-called central problem of motor control) grants much less latitude to the neural controller than generally thought (in fact the system has barely enough muscles!), and (iii) that tendon-driven systems pose specific control challenges that limit the choice of neural strategies.
Importantly, the number of people whose own hands would benefit from an improved understanding of neural control of manipulation is orders of magnitude greater than the number of people who require prosthetic limbs or assistive robots. Equally critically, understanding this one outstanding biological example of sophisticated manipulation can help guide our engineering work towards truly versatile manipulators.
Funding: NIH R01 AR050520, NIH R01 AR052345, NIH R21 HD048566, NSF EFRI-COPN, NIDRR RERC.
See more of: Body and Machine
See more of: Seminars