scorecardresearch Skip to main content

The memory hackers

It’s not quite “Eternal Sunshine of the Spotless Mind,” but scientists are exploring what memories are made of — and how to dial down the pain of bad ones.

Memories feel ineffable, but we are getting closer to understanding them more concretely.Yifan Wu for the Boston Globe

You may be standing in your kitchen when your mind flashes to the road you grew up on, a memory of hot gray gravel. Turning down the sheets in the bedroom, you recall the bright joy of a birthday cake someone made for you once, long ago. Memories are often startling in what they bring back to us. And sometimes we wish we held onto them less, especially when what we have lived through is hard to keep living with.

What is happening in the brain in these moments? Memories feel ineffable, but we are getting closer to understanding them more concretely.


In the last decade or so, neuroscientists have gained the ability to permanently mark the neurons that flash on as a mouse relives a specific memory. These traceries reveal that a handful of cells is linked to each recollection. This insight has opened the doors to a new era in memory research. With more precision than ever, scientists are now able to seek answers to questions like: What does an individual memory consist of? Is it possible to lessen the pain of extreme memories with what one researcher calls “therapeutic forgetting”? And what does it mean, on the level of your cells, to forget?

We begin to lose our memories when we are very young, a phenomenon called infantile amnesia. Young children have easy-to-access, well-developed memories, as anyone who has spoken to a chatty 3-year-old about what they got for Christmas can attest. But these memories slide out of reach sometime during childhood. Few adults remember their lives before the age of 4 or so, notes Paul Frankland, a neuroscientist at the Hospital for Sick Children in Toronto. The person shaped by these experiences survives, while their ability to recall them decays.

One of the drivers of this phenomenon may be the furious development of the hippocampus, a brain region involved in learning and emotion. “If you train mice, for example, when they’re 17 days old, they will learn OK and they will remember for a day. But if you test them a week afterwards or a month afterwards, they show complete forgetting,” says Frankland. “Our idea is that continued growth of the hippocampus overwrites those memories.” Although the hippocampus develops very quickly in the early period of life for both mice and humans, the hippocampus in some animals does most of its growing before birth. And at least one such animal, the guinea pig, does not show this same pattern of forgetting.


Intriguingly, forgetting a memory doesn’t necessarily mean destroying it. In a recent review paper, Frankland and Tomás Ryan, a neuroscientist at Trinity College Dublin, describe experiments in which mice are conditioned with electric shocks to fear the room where the shocks are delivered. They eventually freeze upon entering the space. The mice are then given an amnesia-inducing drug. Having forgotten earlier events, they no longer freeze. However, if the scientists use light to activate the cells that encoded the fear memory, the mice appear to suddenly remember. “These memories may not be destroyed,” Ryan says. They are merely out of reach of the usual cues for recall.

Experiments using mouse models of Alzheimer’s and other kinds of memory loss have shown similar results: Memories can be moved in and out of the mind’s reach without wiping them away. That makes sense if, as seems likely, forgetting things is part of the process of learning. If a memory grows disconnected from one’s personal experience, if it is no longer useful for helping to predict the future, then it makes sense for it to retreat.


“We shouldn’t be assuming that memory suppression, for want of a better term, is due to a bug or problem in the system,” Ryan says. “It could be a feature.”

Less emotional oomph

“If we know one thing about memory from the last half-century of research, it’s not really like an iPhone video of the past,” says Steve Ramirez, a neuroscientist at Boston University. “It’s a reconstructive process.” As memories are retrieved and then reconsolidated — or put back into storage — they can be altered. This is likely a helpful quirk — we might like to think memories should be immutable, but in reality, the experience of living gives us new ways to interpret the past. A memory of a day at the beach with our parents may change over the years as we grow into the role of parents ourselves. Stressful afternoons in a college classroom might become imbued with nostalgia as years pass.

What if you could combine that insight about reconsolidation with recent discoveries about the basic structure of memories and how they are forgotten? Might it be possible to induce similar changes when a memory is too damaging to keep around, as in cases of PTSD? That is a question that Ramirez and his colleagues are investigating.


They are trying an unusual balm: positive memories. Mice in these experiments are conditioned to have a fear memory as well as neutral memories of their home cages and positive memories of being with a mate. The cells involved in each of these memories are identified, tagged, and, thanks to genetic engineering, made sensitive to light, so they can be activated at the researchers’ whim with a light source implanted in the animals’ heads.

To see if they could soften the impact of the fear memory, Ramirez and his colleagues put the mice in the chamber where they were shocked, and at the same time, they lit up the cells associated with the positive memory. In a paper still awaiting peer review but available on the preprint site bioRxiv, they write that the next time the mice were placed in the room, they were less likely to freeze than they had been before.

The researchers also note that this effect lingered for quite some time. Mice treated with positive memories or neutral memories at the moment of recall and left to their own devices for two weeks still froze less for less time the next time they entered the room where they had once gotten shocked. “It’s not like we’re turning negative memories good,” Ramirez says. “It’s more like we’re suppressing the emotional oomph.”

No one is envisioning artificially activating human memories by beaming light inside someone’s skull. But perhaps someday, humans’ own memories could be used like a drug to soften the anguish of their fear, Ramirez speculates. Protocols in which a therapist guides remembering could draw on this detailed biological explanation of memory. Or drugs to give positive memories more impact could be developed and used alongside existing treatments for PTSD.


Reconsolidation might indeed be the moment to alter a memory, says Daniela Schiller, a professor of neuroscience and psychiatry at the Icahn School of Medicine at Mount Sinai in New York City. She and her colleagues have found that interfering with reconsolidation can destabilize a recollection of fear in humans.

For instance, in a study published in 2010, human volunteers learned to associate a colored square with a shock on the wrist. Later on, the researchers reminded them of the painful association and then showed them the square many times without the shock, a process called extinction training. For people’s response to the square to fade, this training had to happen when the experience of the shock had just been recalled, during the window of reconsolidation. If the researchers waited too long, the training did not have the same effect.

In a twist that complicates the interpretation of Ramirez’s findings, the mice also seemed to have their fear memory deadened if the researchers activated a large, random assortment of neurons while the fear memory was being retrieved — neurons that were associated with no particular memory. That suggests that precision may not be necessary for future fear-softening treatments, says Ramirez. For example, deep-brain stimulation and transcranial magnetic stimulation, therapies where electric currents or magnets are used on the brain, have been successful in treating depression, Parkinson’s, and other neurological disorders, even though they are not manipulating individual neurons.

There’s a long road still to using these new insights into the malleability of memory to develop fear-deadening treatments. But as researchers continue to chart the constellations of memory cells, lighting them up in mice and recording how they respond to change, a new understanding of the nature of memory and forgetting may emerge.

Christine Denny, a professor of neurobiology and psychiatry at Columbia University Irving Medical Center and a pioneer of this approach, says that the next frontiers are at the molecular level, where genes influencing the encoding and retrieval of different aspects of memories are at work.

As Denny says: “Molecularly, what makes up a memory, what is driving the encoding and the storage, whether it’s inaccessible or not — that[will be] a huge leap forward.”

Veronique Greenwood is a science writer who contributes frequently to Ideas. Follow her on Twitter @vero_greenwood.