The researchers are also working on restoring bladder control, a problem for a large percentage of paralyzed people. To empty the bladder, a signal from the brain travels down the spinal cord, signaling the sphincter to relax and release urine. That connection is severed in spinal cord injury, preventing normal urination. "A very powerful local circuit keeps the sphincter closed," says Silver, "which is why patients have to catheterize themselves" -- a process that can lead to infections and other complications. "This is high on the list of things people with paralysis would like to see fixed."

The researchers are now testing a slightly different molecular light switch called vertebrate rhodopsin 4, which turns cells off in response to light rather than activating them. Injecting the gene into the neurons that control the sphincter would allow those cells to be turned off in response to light, allowing the sphincter muscle to relax.

Ultimately, the researchers would like to combine this off-switch with a mechanism to squeeze the bladder, another part of normal urination. "We could put an on-switch into the nucleus that squeezes the bladder, or we could work with people who squeeze the bladder using functional electrical stimulation," says Silver. In functional electrical stimulation, implanted electrodes are used to control paralyzed muscles.

Researchers still must surmount several obstacles to develop the technology into a practical treatment. They will need to figure out how to safely deliver both the DNA and the light to the appropriate nerve cells. LucCell is developing a version of the light switches that are delivered using viruses already common in human gene therapies. The researchers are collaborating with Boyden to modify an implantable light source made from a miniature laser or LED attached to an optical fiber.

"The biggest challenge will be safety," says Karl Deisseroth, a neuroscientist at Stanford who was not involved in the research. "You have to worry about things like possibility of rare but serious immune reactions to the proteins and devices." Because the protein is derived from algae, there is some concern it could trigger an immune response or prove toxic to the cell in the long term. Earlier this year, Boyden and his collaborators published the first paper testing the channelrhodopsin technology in primates, expressing the protein in the frontal cortex of macaque monkeys. The animals showed no unusual signs of damage nine months later. But given the novelty of the technology, extensive safety testing will likely be required.