The researchers used something called optogenetics. If you're not familiar, it's super science fiction-y. Scientists can pick out specific neurons and introduce a special protein to them by way of an engineered virus. Once that protein is present in the brain cells, the cells are sensitive to blue light. That allows researchers to turn particular neurons on and off at will.
In this case, they created a bad memory by shocking mice repeatedly in a particular enclosure. They picked out which neurons were stimulated when the mice relived the memory (as evidenced by their visible fear of being in the shock chamber) and made those neurons, which together are called a "memory engram," light-sensitive in a new batch of mice before training them with the same shock.
Then they made the memories go away the way they might in someone with retrograde amnesia: Some of the mice were injected with a drug called anisomycin that messes with memory formation. Sure enough, those mice stopped being afraid of the shock chamber -- until researchers used blue light to activate the neurons they knew held the repressed memory. Then their fear response returned.
Based on their findings, the researchers believe that "lost" memories may still leave engrams active in the brain. From The Verge:
Eventually, their search lead them to figure out that the engram cells in one part of the hippocampus were strongly connected to another part of the hippocampus, and it was this pattern of connectivity that survived the drug treatment. When they extended the search further, they realized that other regions of the brain, including the amygdala, where where fear-based memories can be found, were also involved in this network. The researchers were therefore able to retrieve the memories because other connections in the brain — connections that were unaffected by the drug, but inaccessible without the light treatment — were storing information related to the shock treatment as well.
"Our conclusion is that in retrograde amnesia, past memories may not be erased, but could simply be lost and inaccessible for recall," lead author Susumu Tonegawa, director of the RIKEN Brain Science Institute in Saitama, Japan, said in a statement. "These findings provide striking insight into the fleeting nature of memories, and will stimulate future research on the biology of memory and its clinical restoration."
The findings aren't fit to help humans with amnesia anytime soon. "It's very difficult to be doing this in humans, partly for the ethical reasons — the work is invasive — but also because we tag the memories in the brain before they're learned," Tomas Ryan, a neuroscientist at MIT and a co-author of the study, told the Verge.
But the work does suggest that "lost" memories might still lurk in the brain, so it's possible that humans could one day find a more plausible way of retrieving them.
Want more science? Give these a click: