Reading the brain, moving the muscles
About 450,000 Americans are living with damaged spinal cords, and controlling the limbs by thought alone remains a Holy Grail for those with severe paralysis. To date, scientists have tried to detect the brain's intentions by monitoring groups of nerve cells that normally control movement.
It's complicated, but the idea has worked in preliminary tests.
Now, a University of Washington research group has taught monkeys to control one muscle using one neuron -- through an external circuit that bypasses the spinal cord.
taught two monkeys to move their wrists to play a simple video game
implanted several electrodes in the monkey's brain and others in muscles in its arm, then connected both electrodes to a control device
paralyzed the arm for several hours with a reversible anesthetic and
allowed the monkeys to play the video game by activating individual neurons in the brain
Bypassing the spinal cord
Through a feedback process, the monkeys learned that firing a particular neuron could stimulate the paralyzed muscle, says co-author Chet Moritz, a postdoctoral fellow at the University of Washington. "The only way they could move the wrist was to change the neural activity in their brains," Moritz explains.
It's not clear how the monkeys figured out that firing the specific neuron would move the wrist, but doing so took them only two to 60 minutes, Moritz says. "It was very rapidly adopted," he says. "We don't know what goes on in the brain, but the technique we use is based on operant conditioning. They get applesauce or fruit juice every time they do something close to what we want them to do."
You don't have to force the monkeys to play the games, Moritz emphasizes. "These are teenage male monkeys playing video games; it's the favorite part of their day, they enjoy the challenge, they enjoy going to the next level," Moritz says. Temporarily paralyzing the arm, he says, makes the game harder, and "that increases the motivation to learn."
Going it alone
The new technique has yet to be replicated, but its emphasis on watching one neuron confers some theoretical advantages. The technique was "distinctly different from the conventional approach, which is typically based on recording the activity of populations of cells," says co-author Eberhard Fetz, a professor of physiology and biophysics at Washington.
Reading individual neurons reduces the need for signal processing, says Moritz. "We use a very simple conversion that does not require a complicated computer algorithm or a large amount of processing power," he says, "so we may be one step closer to a low-power device than previous studies that required several desktop computers."
Small is big, Moritz adds. The control box is only as big as a cell phone, and further miniaturization is possible. Smaller devices, in turn, allow smaller batteries, a key constraint for prosthetics -- especially for brain implants.
Got those "how long?" blues
Now we reach the bummer paragraph. Although promising, nerve damage may put patients with multiple sclerosis or Lou Gehrig's disease, also known as ALS, out of reach of the new approach.
And even those with an injured spinal cord must be patient, Moritz says. "This was an initial demonstration that this type of technology is possible," he says. "It will take several years -- even several decades -- before it's ready for a clinical application." Among the important obstacles: Creating a system that avoids the infection-prone step of passing wires through the skin, and making an electrode that can record for years on end.
After a few weeks, many existing brain electrodes begin to lose their sensitivity, Moritz adds.
Nonetheless, the new approach may be simpler to learn and to implement, and it takes advantage of the brain's ability to alter itself to satisfy a changing environment.
Perhaps the most remarkable result of the new experiment was the finding that neurons outside the motor cortex were also able to play video games. "Even neurons that were unrelated to the movement of the wrist could be used to control the wrist muscles," says Moritz, "so it dramatically expands the potential pool of control signals available for brain-computer interfaces."
- David J. Tenenbaum
• Direct control of paralysed muscles by cortical neurons, Chet T. Moritz et al, Nature, 16-Oct-2008 (Vol. 455, No. 7215).