By David Cyranoski of Nature magazine

Memories fade, events get conflated, names get attached to the wrong faces, or, in the case of post-traumatic stress disorder, signals in safe environments can mistakenly evoke emotions that rightly belong to a battlefield tragedy. Understanding the proper functioning and malfunctioning of memory retrieval holds clues for how to deal with memory disorders and learning disabilities.

In research published in Nature, a group led by neuroscientist Pico Caroni at the Friedrich Miescher Institute in Basel, Switzerland reveal structural 'traces' that help to keep memories precise by preventing unrelated events from evoking them.

The team looked at changes in the synaptic structures--the structures involved in signal transmission from neuron to neuron--in the hippocampus of the brain as mice went through a fear-conditioning procedure in which their paws received electric shocks in a particular room. If, after a couple of days, the mice entered that room, they froze in fear. In a different but similar room, they sniffed about as normal. But after about two weeks the mice froze in both rooms. The memory had been 'generalized', producing a functional response to a wide range of cues instead of the specific one that had been learned. "The memory is still there, it might be there forever. But it changes," says Caroni.

When the team followed the structural changes in the hippocampus during the conditioning process, they found a large increase in the numbers of synaptic structures at the terminals of neurons called granule cells, which have axons called large mossy fibers after the memories were formed. By the time the memory had been generalized, these synaptic structures had disappeared.

New synapses

To confirm that the formation of the synaptic structures was involved in memory precision, Caroni's group put mice that lacked a protein necessary to form the synaptic connections through the same shock treatments. In these mice, the memory became generalized after only one day. But when researchers introduced that protein into the mossy-fiber neurons in the hippocampus--even if into only 20% of the cells--the mice recovered the ability to maintain precise memories for weeks. "That really nails it down," says Caroni.

New synapses form when new skills are learned--the property known as 'plasticity'--and Caroni's experiments suggest that one of their roles is to inhibit the retrieval of the memory in response to unrelated cues, says Sam Cooke, a postdoctoral fellow at the Massachusetts Institute of Technology in Cambridge who works on the neurobiological basis of memory. "The fact that so many mossy-fiber terminals seem to undergo plasticity as a single association is formed suggests that it is not storing information about the actual context in which the shock was received itself, but rather encoding the much larger number of contexts in which it would be inappropriate to produce a freezing response," says Cooke. "It is very interesting that suppression of spurious associations may be a major form of memory."

According to Mark Hübener, a neuroscientist at the Max Planck Institute of Neurobiology in Martinsried, Germany, the most exciting finding is the high spatial and temporal correlation between the formation of specific memories and the structural changes. "The data go further than any previous paper in linking structural changes in the mammalian brain to the formation of very specific memories," he says.

Bad memories

Caroni says that the phenomenon could improve the understanding of post-traumatic stress disorder (PTSD), in which traumatic memories are evoked by environments very different from those in which they originated. The generalization of the mouse memories was a form of PTSD, but when the mice were reintroduced to the environment in which they were fear-conditioned, the synaptic structures reappeared within a couple hours and the precise memories were reinforced, saving them from anxiety in other environments. In the same vein, one treatment for PTSD is 'exposure therapy' in which patients are reintroduced to the place where the traumatic events happened in order to help them control their fears.

Cooke thinks that a wider range of disorders might be related. "Loss of suppression of spurious associations may also be involved in other neuropsychiatric disorders, including autism spectrum disorders and schizophrenia," he says.

Such links will take time to establish. But the immediate future presents many opportunities. Caroni says there is "big time" room for follow up. He plans to start by analyzing different forms of memory in the hippocampus, comparing, for example, emotional memories versus cold, rational cognitive memories. "It's a way to study how systems learn. We will look at it system by system,' he says.

This article is reproduced with permission from the magazine Nature. The article was first published on May 3, 2011.