Memories are the result of a lifetime of experiences and, consequently, regulate patterns of behavior. The ability to modify or create records of memories could have major implications including cutting through the need for therapy to bring about changes to the way we feel, think, and behave. Now, researchers have constructed an artificial memory to elicit attraction or avoidance to a smell never previously encountered—essentially, they performed ‘inception’ in a mouse.

In the future, this work can be applied to technologies like brain-computer interfaces to bypass time-consuming training, like in The Matrix where Neo, the protagonist, learns many forms of tactical skills by having various training programs uploaded directly into his brain in a matter of moments. More importantly, it provides a treatment platform for people who experience psychological trauma, such as post-traumatic stress disorder (PTSD), by modifying or erasing painful memories.

Thanks for the memories 

Animals readily learn to associate cues present in the environment with events relevant to life and commit these to memory for survival. For instance, the stimuli related to the presence of a drool-worthy burger or a rotten egg and the memory for these associations guide decisions in the future by promoting approach or avoidance, respectively, related to these environmental cues that once were considered neutral.

Key components of memory traces underlying several types of associations have been localized to specific brain regions and neural circuits along with the patterns of neuronal activity that correspond to specific experiences. With this grasp of how memories are located and encoded in the brain, it seems more and more feasible to reverse engineer and artificially implant a memory tied to a stimulus never faced with or an experience lived through.

Performing inception—in theory

To prove that you have successfully artificially implanted a memory, presentation of a real-life external stimulus must elicit a behavioral response based off a never experienced memory, such as approaching or steering clear of stimulus depending on whether it was associated with reward or harm, respectively.

What is more, the memory retrieval for this behavior must be limited to the ‘trained’ cue only and not irrelevant or unrelated cues—analogously, you wouldn’t want a robot trained to compact trash (or weld plates onto truck bumpers) wrongfully triggered by a person.

So, artificially implanting a memory would be best suited to a region of the brain that is well characterized and stereotyped across a species by anatomy and pattern of neural activity associated with a stimulus.

The ol’ smell factory system

The region of the brain devoted to sensing smell—the olfactory system—of the mouse is well characterized and the neural response to individual smells is intrinsically consistent across animals. Brain cells that sense odors—olfactory sensory neurons—are activated only by a single smell because they uniquely express receptors for a single scent.

Groups of olfactory sensory neurons with the same odorant receptor have cell projections like cables that converge onto spatially distinct regions in the olfactory system called glomeruli. For example, the chemical acetophenone—which occurs naturally in many foods including apple, cheese, apricot, banana, beef, and cauliflower—is recognized by a specific receptor unique to a group of olfactory sensory neurons that all project to the same glomerulus. In theory, by simply examining the different groups of activated olfactory sensory neurons in a mouse, you could decode the identity of the encountered smell.

So, these olfactory system properties make possible the artificial implantation of a behavior conditioned to an odorous stimulus without ever consciously and physically coming across the aroma by simultaneously stimulating the brain regions devoted to smell (i.e. olfactory glomerulus) and reward. If the model is right and all goes to plan, the mice will approach or avoid an otherwise neutral odor that they have never smelled before. 

Unrecognizable woman holding torn picture of couple in love. Ended relationship. Crying.Valentines day composition. Studio shot on brown wooden background.
Halfpoint / Getty Images

Creating scentimental behaviors

Researchers did exactly just that—they paired stimulation of glomeruli with either rewarding or aversive artificial brain stimulation and then examined whether these direct conditioning procedures had resulted in artificial memory implantation by presenting an aroma that the mouse had not encountered previously.

Vetere and colleagues concurrently stimulated regions implicated in aspects of award and aversion—the lateral and medial ventral tegmental areas, respectively—and a glomerulus to artificially associate an olfactory stimulus with either reward or aversion.

To do so, they pinned down the olfactory glomerulus containing terminals of sensory neurons responsive to acetophenone. Then, Vetere and colleagues artificially stimulated the olfactory neurons that sense acetophenone in combination with stimulation of approach or avoidance circuitry.

What they saw was that, even though mice had never smelled acetophenone before and had never experienced rewarding or aversive stimuli in the presence of acetophenone, mice spent increased or decreased in a test chamber compartment perfumed with acetophenone depending on whether the reward or aversion brain areas were targeted!

“It’s the stuff of sci-fi movies: Inception, The Matrix, and, my favorite, Eternal Sunshine of the Spotless Mind,” said lead researcher Dr. Gisella Vetere.

On top of that, they found that the presentation of acetophenone corresponded to the targeted brain region devoted to odor-induced natural memory recall and, importantly, not by another smell. In other words, the mice either approached or avoided this aroma depending on whether the stimulation of the associated glomerulus was paired with a reward or penalty.

There is no spoon

Vetere and colleagues next examined whether these artificial memories possess neural circuit properties that bear resemblance to those of real-life, experience-driven behavioral memories. Also, they found similar profiles of neuron activation in mice that underwent behavioral conditioning to acetophenone by shocking the feet of mice compared with mice that experienced paired glomerulus activation and stimulation of the aversive pathway (i.e. projections to the medial ventral tegmental area).

What’s more, quieting the activity related to the amygdala—the brain region that plays a role in linking unfavorable or hostile cues with fear or avoidance behavior—during testing disrupted avoidance of acetophenone following both behavioral conditioning with a shock of the foot and artificial memory construction with stimulation of the aversive pathway in the brain. These findings indicate that the artificially constructed memories use brain circuitry comparable to that used by memories formed from actual, real-life physical exposure to environmental stimuli.

What is The Matrix?

“This raises uncertainties about the nature of ‘experience’ in memory—does stimulation of a brain region lead to experiencing a sensation?” Dr. Veteres asks. “Does stimulating olfactory neurons induces a sensation of smelling a particular scent? The same goes for taste, touch, and sound—even sight! Are these even ‘experiences’? They certainly didn’t occur—did they?”

In the words of Morpheus of The Matrix, “What is ‘real? How do you define ‘real’? If you’re talking about what you can feel, what you can smell, what you can taste and see, then ‘real’ is simply electrical signals interpreted by your brain.”

Eternal sunshine of the spotless mind

It will be exciting to see how artificial memory manipulation will be applied medically in the future. For instance, it is conceivable that memories could be constructed while under general anesthesia, a state without physical stimulation or sensation. This may also provide an opportunity to probe other aspects of circuitry for memory including modification or obliteration of memories, which could have major consequences for people with severe trauma.

Dr. Vetere predicts, “I think that the next steps for the use of artificial memory manipulation for treatment in people will be understanding how to modify unpleasant memories linked to psychological disorders tied to experiences, such as those for people with PTSD and depression—imagine if we could take that which causes us the most pain, choose to dissolve the memory, and forget it forever.”