Scientists may have found a biological reason to explain why two people who witness the same event will, years later, often have different memories of what happened.
It seems that every time an old memory is pulled into consciousness, the brain takes it apart, updates it and then makes new proteins in the process of putting the memory back into long-term storage. The fact that new proteins are made means that the memory has been transformed permanently to reflect each person's life experiences o not the memory itself.
The finding is based on research involving a specific kind of fear memory in animals, but many experts predict that it may also hold true for other kinds of memories in humans. They also say that the discovery could lead to ways of altering or erasing people's memories. The research, carried out at the Center for Neural Science at New York University, was described in the Aug. 17 issue of the journal Nature. This is the first good neurobiological explanation of the way memories are updated, said Dr. Daniel Schacter, a Harvard psychology professor and a memory expert. "It's a mistake to think that once you record a memory, it is forever fixed," he said.
Dr. Elizabeth Loftus, a psychologist who studies memory at the University of Washington in Seattle, said: "This is very interesting research. We're on the brink of being able to figure out how you might accomplish something like memory engineering."
It may be possible to erase traumatic memories in people who are plagued by them, she said, and to better understand how false memories are implanted into people's minds when they are given suggestions that they want to believe.
It has been known for at least 100 years that newly formed memories are initially unstable, said Dr. Yadin Dudai, a neurobiologist at the Weizmann Institute of Science in Rehovot, Israel. A bump on the head, an electric shock or certain drugs can disrupt the process that gradually turns short-term memories into long-term memories through the production of new connections and protein synthesis in memory circuits. In the 1960's, researchers showed that certain drugs could interfere with the recall of memories, he said, but the research did not get very far because the drugs affected the entire brain and could not be traced to cellular mechanisms in memory networks.
Dr. Karim Nader and Dr. Glenn Shafe, research assistant professors at N.Y.U., carried out the new experiments on memory recall in ways that reveal those cellular mechanisms with much greater precision. In a process called fear conditioning, they simultaneously played a tone and delivered an electric shock to the feet of caged rats. Later, when the rats heard just the tone, they froze; they had learned to be afraid. Researchers know exactly how and where this fear memory is hardwired in the rat's amygdala, a part of the brain that processes emotions. If the rat's amygdala is injected with a drug that blocks protein synthesis shortly after fear conditioning, it does not acquire long-term memory of the fear, Dr. Shafe said.
But if the drug is injected six or more hours later, the memory is not blocked; the brain has made new proteins to consolidate and store the memory.
For six hours or so the memory is what scientists term "labile" o open or sensitive to some kind of manipulation. After this period, the memory is firmly in place.
"I was bored with these experiments," Dr. Nader said. "I began thinking, what happens to a memory when you remember it? It would be so cool if it became labile again." He proposed a new experiment: animals would be trained to associate the tone with the electric shock. The researchers would wait a day or more for the fear memory to consolidate. Then they would present the animal with the tone (to retrieve the memory) and a drug that blocks protein synthesis.
"I said the drug would have no effect" on past learning, Dr. Shafe said. If anything, the animal's fear memory should be stronger because the drug could deter the animal from learning that a tone was not necessarily associated with a shock o and that would reinforce the original fear memory. The two scientists bet a cocktail on the outcome. A few weeks later, Dr. Nader won a cosmopolitan.
"My jaw just hit the floor when I saw the result," he said. Instead of freezing at the tone, the rats scarcely reacted. It means memories become labile and open to revision every time they are recalled, Dr. Nader said. And new proteins have to be made before the memories are put back into storage.
Both researchers emphasized that this finding was only a first step in exploring the biology of how the brain consolidates and manipulates memories. It is not known if much older and more established memories are open to editing or if this mechanism is restricted to fear memories alone. Why evolution would choose a strategy that permits memories to be highly malleable is an interesting question. Memories need to be reliable to guide behavior, but they also need to be open to new information. In the long run, these findings may be used clinically to erase traumatic memories, Dr. Loftus said. A patient would recall the troubling event and be given a drug or other agent to disrupt the memory from being reconsolidated.
The research also sheds light on false memories, she said. If a recalled memory is open to revision, incorrect as well as correct information can be woven into the fabric of a memory. Once that happens, a person has no way of knowing what is true or not true. Yet people put faith in their memories to guide their decisions, she said.