Science Brings Us One Step Closer to Implanting False Memories

Illustration for article titled Science Brings Us One Step Closer to Implanting False Memories

Researchers say they've succeeded in giving mice a "hybrid" memory that blurs the lines between fantasy and reality. They forced the mice to recall memories while they were already busy forming new ones, by reactivating specific neurons in their brains. This could even represent a step towards implanting full-blown false memories.


And this could mean that you're even more of an unreliable narrator than you ever realized. Learn to how to hack a memory, below.

Memory formation and memory recollection are incredibly complex neurological processes, and there is much about them that we still don't understand. One thing we do know, however, is that the neurons that fire when a memory is taking shape will activate again when we go to retrieve that memory later. If you could keep track of which neurons were triggered during the creation of a specific memory (the first time you saw Star Wars, for example) and regulate their activity, you could — to some extent — force yourself to reflect on this past experience at a later date.

Now, Ph. D. candidate Aleena Garner and her colleagues at UC San Diego have used a mouse model to achieve a similar level of neuronal (and memory) control. By labeling the neurons that fire in the brain of a mouse when it is exposed to one environment, and forcing those same neurons to reactivate at a later date in an entirely different environment, Garner's team has gained unprecedented insight into just how much the formation of a new memory can depend on an old one.

How to Hack a Memory

The experiment hinges on the use of some downright brilliant genetic and cellular manipulation.

Garner and her colleagues engineered mice with neurons that produce a receptor — dubbed hM3Dq — in an activity-dependent fashion; that means that even though the genetic information necessary to produce hM3Dq could be found in neurons throughout the mice's brains, only those cells that were sufficiently active would actually express the protein. This allowed the researchers to label specific patterns of neural networks that became activated when the mice were exposed to a particular environment — a test chamber with a unique coloration and scent, for example.

But the hM3Dq receptor does more than tag specific neurons — it can also be used to activate them. By dosing the mice with a drug called clozapine-N-oxide (called "CNO" for short, the drug interacts specifically with the hM3Dq receptor), Garner and her team could trigger activity across very specific, hM3Dq-labelled neural networks.

Illustration for article titled Science Brings Us One Step Closer to Implanting False Memories

In other words, the big advantage of CNO-mediated hM3Dq receptor activation is spatial specificity. What do I mean by spatial specificity? If you place an electrode directly onto a brain, it will cause every neuron in the electrode's vicinity to activate. The brain of an hM3Dq-labeled mouse, on the other hand, is a little like this maple tree. The single branch of off-colored leaves is not unlike a specific network of neurons that has been labelled with hM3Dq receptor. If Garner and her colleagues want to activate only those neurons associated with a specific memory, they need only administer CNO to the mouse, and a specific branch of neurons will light up. The rest of the brain's "branches" are unaffected, and continue functioning as they normally would.


Forgetting Fear

To see what effect internally generated brain activity would have on the formation of new memories, Garner and her colleagues exposed each hM3Dq mouse to a test chamber (we'll call the first chamber "experience A"), and allowed it to become familiar with the unique sights and smells of its new environment. This allowed time for the mouse's active neurons to be labeled with hM3Dq receptors.


Twenty-four hours later, the mice were dosed with CNO (causing the neurons associated with the first test chamber to activate), and placed into a second, physically distinct test chamber (experience B), with markedly different sights and smells. The mice were then allowed to explore their new environment, albeit at a price; once each mouse had become familiar with the second chamber, it was subjected to a series of mild electrical shocks.

These shocks are part of a procedure known as fear conditioning. Typically, a mouse subjected to fear conditioning will become afraid of the environment in which it received the shocks. If a mouse is later placed back into the environment in which it was conditioned, it will stop dead in its tracks, "freezing" out of fear. This is taken as a sign that the mouse remembers not only the environment, but the uncomfortable events that transpired there.

Illustration for article titled Science Brings Us One Step Closer to Implanting False Memories

But with Garner's mice, something extraordinary happened. The mice would only freeze in fear if they were both:
1) Placed in the second chamber (i.e. the chamber where they were fear-conditioned), and
2) Given CNO to activate the neurons associated with the first test chamber


If either condition were presented independently, the mice would carry on about their business — as though they had forgotten to be afraid. The mice had formed what the researchers call a "hybrid memory representation, incorporating elements of both the CNO-induced artificial stimulation and the natural sensory cues from the [fear-conditioning scenario]." When the mice were subjected to the electrical shocks in experience B, their memory of the event was being sculpted not only by the physical context of the second chamber, but by the imaginary, internally generated representations of the first chamber, as well.

Holy crap — So the Researchers Implanted an Artificial Memory into the Minds of these Mice?

Sort of. It's a tempting conclusion to draw, yes — but it isn't an entirely accurate one. That said, it isn't completely wrong either. Stick with me… we'll get to memory implantation, I promise.


Strictly speaking, what the team's findings do best is support the hypothesis that the neural activity associated with making new memories is constantly interacting with pre-established neural activity. In other words, explains Garner, the process by which we form new memories is "linked to pre-existing notions"; a memory that you forge today based on your physical surroundings will be shaped in the context of your past experiences as you imagine them.

This has big implications, not only for the reliability of our memories (if our recollection of an event is built around our previous experiences, then how objective is the newly formed memory, really?), but in our brains' ability to call on these memories across a variety of situations. Garner provided us with an example: Activating the neurons associated with a specific memory, she explains, will not necessarily elicit recall of that memory if other neurons that don't support that memory are also active. She continues:

Illustration for article titled Science Brings Us One Step Closer to Implanting False Memories

For example, if you are looking at a hot oven and the last time you saw a hot oven you ate a delicious cake, then you may become hungry and happy because you are thinking about the cake.

However, if the last time you saw a hot oven you burned your hand, then you may cringe at the thought of the pain you suffered. Or you may have had both experiences in the past. Our results suggest that the memory [you recall] depends on the reference frame that is also active at the time of memory retrieval.


One big difference between Garner's example and her team's experiment, however, is that the brain activity associated with these memories is triggered organically, in a context that warrants their recollection (ovens make cakes; they also burn hands). In the experiment, however, the brain activity shaping the mouse's new memory has been induced artificially? What, then, distinguishes the creation of a hybrid memory representation (i.e. what Garner and her colleagues claim to have done) and the implantation of a memory?

This question, says Garner, brings up an interesting point. After all, she asks, "how does one define
what a memory is, exactly?" She continues:

We do not know what the mice are specifically perceiving (a hybrid context?), but we have created a different memory trace, that is, the cellular correlate of the memory, in these mice compared to normal control mice and thus affected their behavior.


This, she explains, gets at the idea of memory as perception. Does she believe her team has had a hand in altering the perception of memory in its mice (even though she cannot say specifically what it is that the mice are perceiving)? Yes. But this, in and of itself, she says, is nothing new — entirely imaginary sensory perception often occurs in humans in the form of hallucinations, and we can induce these imaginary perceptions through the use of drugs like psychedelic mushrooms.

"However," Garner explains, "the difference between general drug administration and our genetic manipulations… is cellular specificity." She continues:

As we are able to more precisely target specific cell populations, we will be better able to generate a specific perception… depending on how one defines memory, it is not improbable that in the near future we will be able to ‘implant' an entirely imaginary memory perception.


Garner's team's findings are published in the latest issue of Science.

For a more detailed overview of the team's results, check out the following Perspective article, also published in the latest issue of Science: The Imaginary Mind of a Mouse, written by Drs. Richard Morris and Tomonori Takeuchi.


Oven and cake via; experimental figure via Morris and Takeuchi; Maple Tree with false branch by the author, photographed at MASS MoCA


It's pretty easy to implant false memories now. Go up to a friend or cow-orker. Say, "I really like that blue shirt you wore yesterday." He may say thank you, he may say he didn't wear a blue shirt yesterday or whatnot. Doesn't matter.

Now, go back to this friend several times over the course of a day and mention something about a blue shirt — that blue shirt looks good, we should all get blue shirts, whatever.

At the end of the day, ask your friend what color the shirt he wore yesterday was. More often than not, he will say blue even if you know for a fact that he wore a red shirt yesterday.

Try it sometime.