In the Penn-led project, researchers will measure brain activity as the patients engage in memory games and other tasks, to determine what electrical patterns are associated with memory when it is operating at peak performance.
The resulting data will then be used to guide delivery of mild electric stimulation to nudge the brain into that optimal learning state as needed, said Michael J. Kahana, a Penn psychology professor who is overseeing the effort.
Preliminary research, including a 2012 UCLA study in the New England Journal of Medicine, suggests that the concept is not far-fetched.
"You may not be able to fix the entire system after it's been damaged," Kahana said. "But at least you could potentially give people . . . their best function much of the time."
The project is unusual in both its speed and the involvement of multiple disciplines from academia and the private sector. The Defense Advanced Research Projects Agency (DARPA) - known for such exotic projects as competitions to develop a driverless car - wants the two teams to come up with implantable memory-boosting devices within four years.
"This is not like, 'Here is some money, go do great science,' " Kahana said. "This is DARPA saying, 'We've got veterans, they have a problem, and we've got to fix it.' "
More than 270,000 members of the military have suffered traumatic brain injury since 2000, DARPA said, in many cases with persistent memory problems that medicine can do little to treat. Even more civilians have suffered brain injury during that time.
In theory, the fruits of the research also could be used to treat patients with Alzheimer's, Kahana said.
Patients for the Penn-led study will be recruited at seven medical centers across the country, including the Hospital of the University of Pennsylvania, Thomas Jefferson University Hospital, and the Mayo Clinic. In addition, neuroscientists, engineers, and mathematicians from Drexel University, Boston University, and the University of Washington will help by developing computer models of memory formation.
Medtronic Inc., a prominent maker of the electrode systems used to treat Parkinson's, would then use the results to develop an implantable device with 256 electrodes - 32 times as many as in the devices now used for Parkinson's.
NeuroPace Inc., a California company that has developed brain-stimulation technology to treat epilepsy, also would contribute.
The project partners from Jefferson include neurosurgeon Ashwini D. Sharan and neurologist Michael R. Sperling, who treat epilepsy patients.
In previous research, the Jefferson team discerned electrical brain patterns associated with learning, said Sharan, a professor of neurosurgery at Jefferson's Kimmel Medical College.
"We're able to tell when the patient is concentrating, when they're encoding memory, when we're confusing them and when we're asking them to retrieve memory," he said.
Still, team members acknowledge there is no guarantee of success. Gordon Baltuch, a professor of neurosurgery at Penn's Perelman School of Medicine, characterized the project as "extremely ambitious."
"This is way out there," Baltuch said. "But if there's a group that's going to do it, it's DARPA."
Kahana, the project leader, said that all people tend to vary widely in their ability to form and retrieve memories. On memory tests, people perform up to twice as well on good days as on bad days.
Kahana said the goal is not to help people exceed their highest performance level, if that were even possible.
"I'm not trying to make you do anything better than you can do," he said. "I'm trying to get you to look on those bad days just like you do on your good days."