Stimulating brain cells can make false memories

Any crime scene investigator can tell you that memories are unreliable; the way people remember a place or event changes over time and varies between individuals. But for the mice in one lab at MIT, the accuracy of memories is even more suspect. Howard Hughes Medical Institute researchers in that lab have discovered how to alter the animals' memories by turning on neurons in the brain that are associated with the memories and updating them with new information. The new findings, which appear in the journal Science, illustrate that a mouse can be made to fear a cage by giving it a foot shock while at the same time reactivating a memory of the cage to associate the two. "The kinds of things that once existed only in the realm of science fiction movies like Inception and Eternal Sunshine of the Spotless Mind are now experimentally possible," says Steve Ramirez, a graduate student in the lab of HHMI investigator Susumu Tonegawa and first author of the new work. Researchers knew that memories are stored by the brain in a small set of neurons. Understanding how this information is encoded could be key to understanding how human memory works as well as memory disorders. But identifying exactly which neurons are linked to specific memories has been technically challenging. Ramirez and Tonegawa, along with Xu Liu, a postdoctoral fellow in Tonegawa's lab at the Massachusetts Institute of Technology, had previously developed a way to pinpoint the specific handful of neurons that are activated in the brains of mice in any particular situation. The technique relies on optogenetics, a method of controlling brain cells through bursts of light developed by HHMI early career scientist Karl Deisseroth at Stanford University. The researchers engineered brain cells to produce a light-sensitive protein whenever the neurons were activated in a new setting or situation. Then, by shining a light onto the brain through a fiber optic cable connected to the mouse's skull, they could reactivate only that subset of neurons. Even without reactivating the cells, they could determine which cells had been activated by measuring which contained the light-sensitive protein. The approach was described in a 2012 Nature paper. More recently, the scientists wondered if they could alter the way a mouse remembered a setting by activating neurons associated with it. They chose to test this idea by attempting to change whether or not a mouse was afraid of a particular cage. "In mice, fear can be seen as a binary behavioral output," says Ramirez. "Either the animal is exploring a box that it's interested in, and it's curious and sniffing around. Or, if it's displaying fear behavior, it's huddled in a corner not moving. So it's a very easy, very powerful readout of memory." To see whether they could make an animal associate fear with a previously neutral setting, Tonegawa's lab group first exposed mice to one of four unique cages. Each cage had distinct flooring materials, artificial smells, and different lighting. As the mice scouted out the new room, whichever neurons were activated produced the special light-sensitive protein. Next, the mice were moved to a second cage. This time, as the mice explored, the scientists used light to turn on the neurons that had been activated in the first cage and simultaneously shocked the feet of the mice. Then the mice were put back in the first area -- where they'd never received a shock. The mice were clearly fearful of the setting, Ramirez says, spending more than a quarter of their time frozen in place. "We were astonished that this worked on the very first mouse we ever tried," he says. "We got the animal to be scared of an environment where technically, nothing bad had ever happened to it." By contrast, when the mice were put in a third cage that they'd never been in before, they exhibited no fear. And in a control group of mice that had received shocks in the second cage but no neuron reactivation, the first cage never induced fear. After the successful experiment, Tonegawa, Ramirez, and Liu looked at the details of which neurons in the brain had been responsible for inducing the memory of the first cage. The neurons, they found, were located in the dentate gyrus, part of the hippocampus. The dentate gyrus has previously been implicated in the formation of memories, and is one of the areas of the brain with the most new neuron generation during adulthood. But most evidence about its importance came from instances in which the area had been damaged and memories lost. "This study gives us information on the basic mechanism that could be happening in the brain when memories or false memories are formed," says Ramirez. "Next, we want to see if we can do the same with not only fear memories but pleasure memories or memories of objects or memories of other mice."