On July 26, 2013, the 341st volume of the world-renowned journal Science delivered news that seemingly took the neuroscience field by storm. Steve Ramirez and colleagues reported that they had created “a false memory in the hippocampus.”
Every major news outlet followed suit. “Scientists Trace Memories of Things That Never Happened,” reported The New York Times. “Scientists... cause rodent-style Inception,” wrote Engadget. Gasp, Leonardo Dicaprio can now really mess with our memories!
This discovery meant great news for treatment of memory-related disorders. Memories can be replaced in people with Alzheimer’s, or memories can be erased from those who suffer from post-traumatic stress disorder.
On one hand, I agree that the capacity to implant memories holds therapeutic promise in the clinic. On the other hand, I propose that we need to step back and critically evaluate what “creating a false memory” means. No doubt, the report is interesting. However, I contend that the authors generated false associations rather than truly false memory, two entirely separate phenomena.
Just as Hollywood stars leave behind their handprints on the Walk of Fame, our experiences leave behind a physical trace known as engrams, the biological representation of memories. Yet where these engrams are located remains a mystery. After all, how are we going to look for traces of discrete memories among 100 billion neurons contained by the three-pound mass of spongy brain tissue? Research suggests that during a memory-forming experience, certain neurons become activated and undergo physiological changes, beginning to carry memory engrams. Reactivation of these engram-containing neurons will result in the playback of a specific memory.
Undoubtedly, it is challenging to prove and investigate the biological mechanisms underpinning how neurons generate memories. Scientists need a way to turn on a select population of neurons and observe how activation of these neurons contributes to memory formation. How could we make sure that only a small subset of neurons are activated while keeping the billions of others dormant?
Optogenetics is a new handy tool in neuroscience that can shed light on the mystery surrounding the biology behind memories. In essence, optogenetics allows scientists to activate any neuron they want by shining a light on cells labeled with special proteins, while other untagged neurons remain undisturbed. Because of the ability to modulate neural activity, numerous labs have gotten their hands on optogenetics, leading to new important discoveries about mental illnesses such as obsessive-compulsive disorder.
Ramirez and colleagues have applied this technology to study the formation of memory. First, they let a mouse run free in a blue chamber. As the mouse familiarizes itself with the environment, certain neurons become activated to store the mouse’s experience. Through optogenetics, these same neurons could be reactivated, leading to the mouse’s recall of its previous experience in the blue box.
When the mouse was moved to a different red chamber, the animal received shocks to the foot while being forced to recall the memory of being inside the blue chamber. Afterwards, when the mouse was returned to the blue chamber, it froze — a sign of fear from being shocked, even though nothing happened. Instead of linking the red chamber to being shocked, the mouse incorrectly associated the blue chamber with painful fears. The researchers thus called this erroneous linkage a “false memory” implanted into the mice by artificial means.
What is the process underlying this incorrect pairing of experiences? In the brain of the mouse, three memories existed: 1) Being inside the blue chamber 2) Being inside the red chamber 3) Being shocked in the foot. These are the three real, independent events that occurred and were stored as discrete memories by the mouse.
These memories are next arranged so that being shocked in the foot is matched with either the red or blue chamber. What should happen is that the mouse correctly pairs the red chamber with being shocked. However, forcing the mouse to recall the memory of the blue chamber while being shocked in the red chamber interfered with the association process, causing the discrete memories to be mismatched. In other words, the researchers reshuffled pre-existing memories that led to false associations, not false memories. As the mouse returned to the original blue chamber, it recalled two pre-existing memories: being in the blue chamber and being shocked. True, the mouse was never shocked in the blue chamber. Yet the incorrect associations of two real, distinct memories tricked the mouse into thinking it was shocked, even though nothing bad happened.
Since association may be defined as a type of memory, it may be argued that creating false association is tantamount to implanting false memory. However, the association process is conditional upon the existence of unrelated events. These events are the true fundamental units of memory that are then later connected together in a coherent sequence by association.
I argue that manipulating this process (as Ramirez and others have done) does not change the original pieces of memory, thereby making it impossible to establish false association as false memory. The authors did not necessarily implant “memories of things that never happened.” They showed that it was possible to override the association system through forced recalls of memories in the wrong context.
