Concepts familiar from grade-school algebra have broad ramifications in computer science.
In work that could aid the development of robotic prostheses, neuroscientists at MIT's McGovern Institute for Brain Research have gotten one step closer to understanding how the central nervous system solves a gigantic problem -- the production of voluntary movements.
The simplest movement requires choosing which combination of motor neurons will stimulate which of thousands of muscle fibers with just the right amount of force and at the proper time.
But no existing computer can analyze the superabundance of variables involved in the movements of a multijointed limb, such as an arm picking up a coffee cup. That inability poses a major obstacle to designing neuroprosthetics for amputees or people with motor disabilities. (In neuroprosthetics, a person's brain or spinal cord signals operate a device.)
As a result, engineers designing robots and prosthetics hope to mimic the way that biological systems approach the challenge.
For many years, scientists wondered whether vertebrates tackle this problem from the top down, with the brain micromanaging the process, or by establishing mini command centers in the spinal cord that relieve the brain of this onerous oversight. MIT Institute Professor Emilio Bizzi, a principal investigator in the McGovern Institute, has proposed the latter, that the central nervous system handles the daunting number of variables involved in a single movement by grouping sets of muscles and their innervating neurons into an integrated unit called a muscle synergy.
In recent studies in frogs, Bizzi and his collaborators found solid evidence for muscle synergies. They showed that grouping muscles in a small set of muscle synergies simplifies the central nervous system's control issues.
But do muscle synergies in the spinal cord operate independently of sensory input, or do they receive feedback from that input (and if so, to what degree)?
Apparently it's a little of both, according to another recent study by Bizzi and colleagues in the Journal of Neuroscience.
Vincent Chi-Kwan Cheung, a graduate student in the Harvard-MIT Division of Health Sciences and Technology and first author of the paper, recorded the electrical activity of a frog's hind leg muscles both before and after severing the nerve roots feeding sensory information into the spinal cord from the muscles. He left intact the nerve roots carrying the commands to the muscles.
Cheung found that, for the most part, shutting off sensory input from the muscles did not perturb the synergies involved in natural jumping and swimming movements.
Bizzi explains the value of having both fixed motor synergies and some feedback from the environment. "If you're walking on a mountain trail, you need to be able to make many small adjustments as you walk, and having a little sensory feedback helps you match your movements to specific conditions."
In practical terms, the near autonomy of the muscle synergies makes it possible to control a large number of muscles with just a few signals generated in the areas of the central nervous system involved in programming voluntary movements. According to Cheung, "That simplifies the future design of neuroprosthetics." Importantly, using a rigorous mathematical analysis, the researchers also found that a computer model representing specific combinations of muscle synergies could predict the movements produced by the animal.
This research was supported by the National Institute of Neurological Disorders and Stroke.