Your Life your connection to The Boston Globe

Pick your own memories?

Research raises the far-off prospect of editing memories in the brain, deleting unpleasant ones and even inserting experiences you never had

A rat scurries into the dark half of a cage. Ouch! The floor is electrified.

Minutes later, Massachusetts Institute of Technology scientists detect the formation of a memory of that painful event -- that floor over there hurt! -- in the animal's brain.

They didn't actually see these words recorded, of course, but used supersensitive electrodes to measure brain electrical activity, combined with techniques that measure changes in brain chemicals, the researchers reported in the journal Science last month.

Those same researchers speculate that some distant day, work like theirs could translate into what some brain scientists call a ``Marilyn Monroe" experiment: A man might have implanted in his own brain an artificial memory of spending the night with Marilyn. Or he might choose to wipe out excruciating memories of combat.

Recent progress like the research reported in the Science paper ``tentatively raises the possibility that we can dissect out memories on a molecular level and add and subtract them," said Dr. Barry Gordon , a memory expert at Johns Hopkins. ``Are you going to do it in humans anytime soon? No way. On the other hand, we almost certainly have, at the biochemical level, the same equipment as a mouse, and so this makes it extremely plausible."

And though editing complex human memories is still science fiction, editing simple rodent memories is starting to appear to be within reach. Or, at least, worth trying to reach for.

Another paper in the Aug. 25 Science described how injecting a chemical into a memory center in a rat's brain could make it forget the location of a danger zone with an electrified floor. The memory could be wiped out 22 hours after it was formed, or a full 30 days later, researchers at the State University of New York found.

The Science papers were by no means aimed at memory-editing. Rather, they attempted to nail down what brain scientists have used as a working theory of memory for a generation: that it involves the strengthening and weakening of connections between neurons.

The patterns created by those changes, researchers believe, may be the basic stuff of which memories are made: a mix of strengthening, known as long-term potentiation, and weakening, known as long-term depression.

A memory begins with an experience that registers as a pattern of electrical pulses in the brain. Those pulses trigger chemical changes that alter the strength of the connections between neurons -- changes that are thought to constitute memory.

Every unique memory should produce a different pattern of change. The Science papers identified what researchers believe may be signatures of particularly strong memories, the kind necessary for survival, which may have made them easier to detect.

Of course, even without an understanding of the biological nuts and bolts, drugs, such as certain types of anesthesia given patients during surgery, can prevent memories from being formed. Psychiatrists are experimenting with drugs that, taken after a trauma, may blunt the memory of it and prevent long-term psychic damage.

But as researchers begin to gain a more sophisticated understanding of memory, the prospect arises of more-sophisticated manipulation, said Mark Bear , a professor at MIT's Picower Institute for Learning and Memory and senior author of one of the Science papers.

For example, he said, his lab's latest findings on fearful memories in rats suggest that these memories evoke particularly strong long-term potentiation. Perhaps, he said, a drug that wipes out these strong connections ``would selectively get rid of the negative memory without affecting the positive memory -- wouldn't that be nice?"

Members of Bear's lab are already thinking about one natural next step: If they can really pinpoint the brain changes that occur when a rat learns to fear a place, might they not be able to induce those same changes artificially, using electrodes that simulate the electrical activity associated with memory formation that they have picked up in their experiments? And if they did it well enough, could they make a rat fear a place even though it had never been there?

It sounds like a simple experiment compared with a simulated night with Marilyn Monroe, but even the enthusiastic researchers themselves acknowledge that it is a long, long shot.

``Consider that there's like 1.1 trillion cells in the brain, 100 billion of which are neurons, each one of which has 10,000 connections to the other ones, each one of which can be increased in its strength or decreased," said MIT researcher Marshall Shuler. Further consider, he said, that memory is believed to be broadly and sparsely spread out across brain regions, making it even harder to detect and replicate.

For all the difficulties of such research, Bear predicts progress on understanding memory may lead to better treatments for anxiety disorders and post-traumatic stress disorder -- and could even help people with memory impairments like Alzheimer's disease.

He is quick to acknowledge, however, that drugs meant to enhance memory have not panned out yet. Even with recent progress, researchers say, memory is hideously complex.

``It is just remarkable how much we've learned in the last 20 or 30 years," said John H. Byrne , director of the Neuroscience Research Center at the University of Texas Medical School at Houston. Much has come to light about the genes and proteins involved in memory, he noted.

But ``this is an extremely complex problem -- memory, the brain. This is the most complicated machine in the universe, and we've been thinking about this for thousands of years, but we've only really been doing anything about it for the last 50."

The great challenge now is to come to understand more about the circuitry that underlies memory, he said. The connections involved in memory are like the pixels on a television screen: ``The pixel is not the memory, it's the whole screen, and we don't know what makes the screen."

So, implanting a memory, even a simple rat memory, is ``going to be very, very hard," Byrne said. ``You'd have to tap into the exact circuit that's activated in the process. We don't know that circuit, and even if we did, we don't have the tools to manipulate it at that fine a grain."

A daunting challenge, to be sure. But, Shuler of MIT said, "Maybe we'll get lucky."

Carey Goldberg can be reached at