Impossible even to broach this subject without reference to Eternal Sunshine of the Spotless Mind, in which an unkempt Jim Carrey enlists a legitimate if shoddily run medical firm to erase the memory of his ex-girlfriend. If that film had a message, it was almost certainly not “erasing memories is good,” but that is not what we are concerned with today; what we are concerned with today is whether erasing—or editing—memories is possible. As it turns out, this has been a robust field of inquiry since before the film’s release; and while Eternal Sunshine-ing a whole months-long relationship is probably out of the question, science-wise, there are plenty of other ways to alter or otherwise mess with memory. For this week’s Giz Asks, we assembled a panel of researchers to discuss these developments.
Professor Emeritus, Neuroscience, Columbia University
In principle, yes. But in practice, it’s very difficult.
The evidence from neuroscience right now suggests that a given memory is very sparsely encoded. What that means is that our cerebral cortex, where most of these memories are stored, has about 15 billion nerve cells, and a particular memory may involve a change in activity of only a couple of hundred of them. Finding those few hundred cells is very, very complicated.
What modern science is now able to do, however, is to take a mouse and to identify a group of cells that are participating in one of that mouse’s particular memories, and then manipulate those nerve cells and demonstrate behaviorally that the mouse has “forgotten” what it was supposed to do. But doing something like that in a human in any sophisticated fashion, a la Eternal Sunshine of the Spotless Mind—that’s just not possible. There are other limitations, but overcoming this one in particular is crucial.
That said, there are now therapies that are being used that seem to be able to work on preventing or reversing things like PTSD. With PTSD and disorders like it, the sufferer generalizes the traumatic event to extend to normal everyday stimuli. For instance, you get mugged near a mailbox on a dark street, and then the mailbox comes to represent the event, and every time you want to mail something, you start panicking. In principle it is now possible—although it may not be feasible—to eliminate that terror, through a variety of therapies and specific drugs.
What these drugs focus on are the different flavors of how a memory gets stored. A memory has four basic phases: it’s initiated, then consolidated, then stored, and then recalled. In both animals like mice and monkeys and in humans, each of these phases have been shown to share a certain mechanism. Scientists are now focusing on those shared mechanisms to see if maladaptive memories like the case of the mailbox can lose their anxiety-provoking aura. You don’t forget that you got mugged in front of a mailbox—instead, the mailbox stops meaning what it used to mean. The memory is still there, but its emotional context has been removed. This sort of therapy might be possible in a relatively short time.
They are also, I should note, studying the feasibility of giving drugs like this to soldiers before they go into battle, so that the trauma isn’t generated in the first place. I find this use of science highly problematic, and think that people should be aware that this kind of thing is going on. On the surface, it seems okay—if soldiers are coming back traumatized from activities that society approves of, we would want to prevent that. But it’s a very slippery slope.
Associate Professor, Psychology, Iowa State University, whose research focuses on human memory, among other things
Right now, in neurobiology and on the more clinical side of things, a lot of research is being done into a phenomenon called “memory consolidation.” Memory consolidates over time—we have known this for the better part of a hundred years. After you learn something, or experience something, your brain undergoes changes through protein synthesis that solidifies, or consolidates, the memory. In the ‘60s, some people suggested the concept of reconsolidation. The idea there is that when an already-consolidated memory has been retrieved, it re-enters a labile state and then needs to be reconsolidated. That work was reinvigorated by some researchers in the early 2000s and has been investigated extensively in the last 15-20 years. There are a lot of people trying to do work that looks at whether we can take advantage of this reconsolidation process and block it, thereby impairing the memory in question—making it weaker, or making it go away.
Most of the research has been done with non-human animals, and it’s been quite promising. When it comes to humans, things get much harder, partly because human memory is much more complex—episodic rather than conditioned memory. So the data with humans is more conflicting, and more controversial, but people are definitely working on it. On the clinical side, many people are trying to look at whether using, for example, propranolol can help this process: issuing it right before or right after people reactivate (and then reconsolidate), to see if that might weaken the emotional impact of those negative memories. In general, the results have been promising, but of course there are conflicting findings there as well.
Professor, Psychological and Brain Sciences, Texas A&M University, whose work investigates the brain mechanisms of emotional memory, among other things
Although we are a long way from handheld memory-wipe devices, a la Men in Black, there is no doubt that memory editing is now within our grasp.
Of course, memory editing is possible because memories are much more malleable than most people think. Indeed, the simple act of retrieving a memory changes it. This is a feature, not a bug—it allows new information to be incorporated into a memory when circumstances change. The fact that memories undergo “reconsolidation” reveals that they are not indelible traces etched into the neural networks of the brain; rather, they are dynamic representations that change with experience and time.
Recently, laboratory studies have revealed methods to disrupt reconsolidation, opening up new avenues to eliminate unwanted memories. Research in both animals and humans has found that drugs that interfere with reconsolidation, including the beta-blocker propranolol, can produce clinically significant reductions in pathological fear responses in patients with post-traumatic stress disorder, for example. These studies open promising new avenues of research to target and manipulate problematic memories. Ultimately, the question before us now is not whether we can edit memory, but whether we should.
Associate Professor, Physiology, University of Toronto, whose work examines the neural substrates of memory and applies these findings to study potential treatments for people with learning and memory disorders
Science has made some amazing progress in understanding how memories are formed, stored, and used in the brain. We now appreciate that memories are stored in groups of cells, often referred to as cell ensembles or engrams. Several researchers have now targeted particular engrams supporting a specific memory in mice and showed that it is possible to disrupt (or essentially erase) this specific memory (and not all memories). Of course, these experiments were conducted in rodents using very cutting-edge techniques. These techniques would not be applicable to humans just yet. But they do provide a proof-of-principle that memories can be modified by targeting “engram cells.”
This also brings up the important question of whether these types of experiments, even if they can be conducted with humans, should be. It is always important to learn from our past. Even some painful or humiliating experiences teach us something to help modify our future behavior for the better. I don’t think anyone in the memory community would be in favor of erasing or modifying the memory of the really embarrassing time when we tried to sing karaoke. But there are some conditions for which intrusive negative memories cause real distress to people and really impact their lives. In these cases of severe post-traumatic stress disorder (PTSD), it might be appropriate to attempt to erase or modify the intrusive negative memory.
The findings from very basic, fundamental studies examining memory are our best hope for really understanding memory so that we can help guide the development of targeted treatments for the variety of memories that afflict humans. It is only through this type of discovery research that we’ll get closer to therapies for PTSD, as well as other memory disorders in people, including Alzheimer’s disease and potentially other brain disorders.
Adjunct Professor, Psychiatry and Behavioral Sciences, Northwestern University
The brain edits memories all the time anyway. It updates, modifies, and re-distributes memories across neuronal networks as time goes by and with each new experience. By better understanding how these processes work, we might be able to edit memories in a controlled fashion.
As for deleting memories—this is not a testable question. One cannot prove that something does not exist. And with memories, one can never know whether they have been deleted or are simply inaccessible for a variety of reasons (suppression, impaired retrieval, etc).
Distinguished Professor of Physiology & Pharmacology, Anesthesiology, and Neurology, State University of New York Downstate Medical Center, who has studied the molecular mechanisms of long-term memory storage. His laboratory discovered the enzyme, PKMzeta, the first molecule shown to be necessary for maintaining long-term memories.
As we understand more about how memories are stored, we will likely be able to erase specific long-term memories.
Memories are thought to be stored as persistent changes in the strength of the synapses connecting neurons in networks in the brain. The persistent strengthening of the synaptic connections occurs during or shortly after learning and is believed to then be maintained for the lifetime of the memory, which can be decades. The enhanced synaptic strength persists even when we are not recalling a memory (i.e., during memory storage). Therefore, the key issue is the biological mechanism maintaining this persistent strengthening of synapses over time.
Until recently, we had no information about the persistent memory maintenance mechanism. Over the last few years, evidence has accumulated that the maintenance is due to an unusual, persistently active enzyme, termed PKMzeta. PKMzeta is synthesized and persists selectively in the networks of neurons involved in learning. Even though the PKMzeta enzyme is constantly being replenished, its overall level in the neurons of a memory-encoding network remain elevated for at least a month in experimental animals. This increase is likely due to its interaction with other molecules at synapses that stabilize the enzyme and maintain its high level. Remarkably, inhibiting the action of PKMzeta or its interaction with its stabilizing molecules erases long-term memories. But because most memories are stored by the same molecular mechanism involving PKMzeta, multiple memories are erased—like reformatting a hard disc of a computer—without regard to the specifics of the memory.
Specific memories might be erased by taking advantage of what happens to memory-storage molecules during recall. The specific content of a memory has to do with which neurons are held together in a network. Specific memories are retrieved when a sensory cue, which stimulates some of the neurons in the memory-encoding network, are activated, which in turn, through the enhanced synaptic connections, activates the rest of the memory network.
When the specific synapses connecting a neuronal network fire during the retrieval of a specific memory, the memory-storage molecules, such as PKMzeta, may be rapidly degraded and resynthesized in a process known as “reconsolidation.” Blocking all new synthesis during the period right after a memory is recalled prevents the specific, recalled memory from returning into long-term storage, leaving the other non-retrieved memories intact. But blocking general protein synthesis is toxic. If PKMzeta or its stabilizing molecules could be selectively prevented from reforming, one would predict a specific memory should be erased.
Do you have a burning question for Giz Asks? Email us at firstname.lastname@example.org.