Asked by a teacher to define the word "memory," one first-grade pupil obliged: "My memory is what I forget with."
This artless definition might well hold true for grown-ups, who increasingly lose their memories as they approach advancing old age - or neurodegeneration, as in Alzheimer's disease.
A team of trans-Atlantic neurobiologists at the Weizmann Institute in Rehovot, Israel, plus one neurobiologist, Diego Berman, now at Columbia University in New York, authored an article in the current issue of Science, dated Aug. 22, 2003. It's title: "Stability of retrieved memory: Inverse correlation with trace dominance." Post-doc Berman is a co-author.
"The main idea in our Science article," he explained, "is that when a memory trace is retrieved in the brain, when that memory comes to mind, several processes can happen. First of all, this memory becomes feeble, which means it's susceptible to destruction. When a memory is formed the first time in the brain," Berman went on, "it has to be consolidated to a long-term memory - retained for days or months or years. This process is called memory consolidation.
"It can be blocked by specific agents, which inhibit protein synthesis. It's processed by memory consolidation. For example, these blockers inhibit the synthesis of new proteins. It was thought for many years," Berman continued, "that this process of consolidation happens once. Only when the memory is formed the first time - and that was it.
"What we report in Science," he noted, "is our finding that this process of consolidation also takes place each time a memory, any memory, is retrieved. It's novel because it goes against the dogma that memories are formed once and then they are stable forever and ever in the brain. We found this not to be the case. Every time a single memory is retrieved or reactivated, this trace again becomes fragile, which means when the memory is being active it can be erased. It undergoes a process that is similar to the process it underwent the first time the memory was formed. At that point it's called memory consolidation, in which it goes from longer to shorter memory. When retrieved it again goes through consolidation but is now called re-consolidation.'"
Foretelling Future Memories From Past Traces
"Then three things can happen," Berman continued. "Either the memory trace is strengthened, it may be modified by adding new information, or it can be erased. For example, consider a person who has a fear memory of spiders. The next time he meets a spider, he can see that the spider doesn't do anything harmful. So he learns that this spider fear-potential aversion only happens once. Hence, this process of learning new things and new associations is called memory extinction.
"What we also show in the paper is that the trace of the activated memory is going to be strengthened or erased, depending on how this memory was acquired. Which means that the initial experience of the acquisition of the memory already tells the future of this recollection - when it's going to be reactivated. We don't know how this effect is possible but it's what we reported in Science. The trace of the memory once it's activated depends strongly on the way it was formed in the beginning. That's our principal finding. What's novel is that the memory actually controls behavior. In our case we did an in vivo study in rats and in fish. Only the memory that is guiding the behavior is the one that can be blocked by these consolidation blockers. This is new because the consolidation process takes place only once in a lifetime."
Making Rats And Fish Behave In Vivo
"We used two kinds of amnesic agents," Bergman recounted. "The one in the rat studies is called amnesomycin, a drug that prevents protein synthesis and causes amnesia. In rats, it blocks memory consolidation.
"The amnesic agent that we used in the fish studies," he said, "blocks electrical impulses. It has nothing to do with protein synthesis. It's an electrical amnesia agent, which lowers the nervous system communication of the whole medaka fish (Oryzias latipes). These 2-inch fish remember that some milliseconds after a light is shown to them, they all received a mild electric shock. What they do is try to escape from the little aquarium where we keep them. They learn to associate the light that comes first followed by the shock. That's associative learning. The next time the fish see the light by themselves, they will be afraid even though no shock is forthcoming, because they learned that light brings shock.
"We have an aquarium with all the fish, and we separate them into ages like newly born, adult, one month old, and so on. In this very small aquarium where we experiment, we show them the light, the shock, and then record the movement of the fishes with a video camera connected to a computer. It's important to understand how memory mechanisms work in lower vertebrates. One can go to lower evolutionary steps and find the beginning of this kind of cognitive processes. We shall also create neurogenetic mutants on these fishes that are defective on behavior.
"In our ongoing research," Berman said, "our first goal is to understand how memory works. We are not only seeking to identify molecular targets of memory, but how the mnemonic processes act. A lot is relatively known about the moment when an initial memory is being formed, but less about the retrieval of memories.
"I know it's a little early to think about human applications," Berman said, "but maybe in ongoing research we might plan new protocols to erase aversive fear memories. Very theoretically, for example, someone who fears height, you take the person to a high place. The person is afraid. At that moment you are reactivating his fear - maybe at that specific moment when the fear of height is being reactivated. Perhaps you can give them some drug and erase that memory forever.
"That's one way of thinking about it. The other way around, we could in the future even develop drugs that increase the acquisition of new memory, so the learning time would get shorter. It depends," Berman concluded, "on how you look at it."