Erasing Memory: No Longer Science Fiction


This may seem like old news.  Thanks to Will Smith’s “neuralizer” blasting away horrific memories of alien attacks in the 1997 movie “Men in Black,” and the quest to bury the heartbreak of a broken romance in the 2004 flick “Eternal Sunshine of the Spotless Mind,” the concept of erasing memory has become commonplace.  More recently research to relieve PTSD using propranolol and other drugs to quell traumatic memories has edged fiction closer to fact, but the method has produced mixed results.  Certainly memories can be obliterated in experimental animals, but the drugs used on lab rats cannot be given to people because of serious side effects and other issues related to the more complex cognitive processing in humans.  The fact is, there is no convincing evidence that the memory of specific experience (called declarative memory) can be erased in humans–until now.

A new study published this week in the journal Proceedings of the National Academy of Science reports that memory of a specific event was impaired by targeting it selectively.  What’s more, the study used no drugs, electroshock, or other invasive measures to wipe away the memory.  They used a cleaver deception to hack into the biological process at a critical point when experiments on laboratory animals indicate that memories can become vulnerable to tampering.

When neuroscientist Jason Chan first presented his findings together with Jessica LaPaglia at a scientific meeting he received a sharply polarized response.  “Some liked it a lot, but some did not…they didn’t believe the results,” he said.  “I didn’t believe it myself three years ago when we first got the data.”

So Chan and LaPaglia, working at Iowa State University, repeated the experiments and tested the results in several other ways.  The results are now clear:  the researchers had blotted out a specific memory by intervening at the precise time in the multi-phase biological process that makes memories stick.

“What’s nice about his particular study,” memory researcher Daniela Schiller, who was not involved in the research explained, “is that they used a memory that is very similar to real-life memories (a movie in this case).”  Moreover they demonstrated that the memory interference was very specific; rather than causing some sort of global amnesia, a specific memory was targeted for ablation, and the method only worked when launched at a particular phase in the process of memory storage and recall.  “The memory has to be reactivated and thus active during the interference” to suppress the memory she explained.

This fits perfectly with the results from laboratory animal research identifying an aspect of memory called “reconsolidation.”  New experiences can be held in mind temporarily but soon forgotten (short-term memory), or they can be stamped into long-term memory through a complex biochemical process that requires the synthesis of new proteins to make long-lasting connections between neurons.  Block the synthesis of new proteins with a drug right after learning something new and the event will never get stored in long-term memory.  Timing, however, is critical.  Wait too long to block protein synthesis, generally a few hours after training, and it is too late–the memory has already been cemented into long-term storage.  Interestingly, animal research has found that when a well-established memory stored in long-term memory is later recalled, the cycle of memory encoding and storage begins again, making the recalled event vulnerable to being forgotten.  Protein synthesis inhibitors given right after recalling a long-term memory blocks its reconsolidation and the memory is lost.

Chan and LaPaglia had subjects watch a movie: a pilot episode of a TV drama “24” involving a terrorist attack in which the villain jabs a hypodermic needle into a flight attendant to disable her with a drug injection.  Later they quizzed the viewers about details in the video, allowing only 25 seconds to answer each of 24 questions about specific events in the 40 minute thriller.  The purpose of the quiz was in fact to make the subjects recall specific events from memory.  Now, according to reconsolidation theory, the memories that were retrieved could be disrupted as they were re-encoded into memory after recall.  Rather than give the subjects drugs to block protein synthesis, they instead manipulated reconsolidation by having the people listen to an 8 min audio recap of the movie they had seen, but some of the facts were altered in the recap.  For example, the recap stated that the flight attendant had been rendered unconscious, not by a hypodermic needle, but rather by using a stun gun.  When the test subjects were tested later, only 17% recalled the hypodermic as the weapon used, instead of 42% who had been similarly misinformed by the false recap, but they heard the recap after playing a computer game rather than after taking a quiz to force them to reactivate the original memory.   Disrupting the memory required that it first be recalled.

Now it might have been that futzing with the facts in this way could have simply confused matters in the minds of participants.   That is, people might have remembered the hypodermic needle being jabbed into the flight attendant perfectly well, but they were uncertain whether that particular memory was the correct answer to the question.  The subjects might have remembered events in both the video and audio recap perfectly well, but they simply got confused about which murder weapon went with which context.  Also difficult to say is whether the original memory had been erased or instead the process of retrieving it had failed.

The researchers were able to greatly discount these alternative hypotheses by clever experimental design.  When they were latter quizzed, participants could only give true or false answers to the probing factual questions–and, the scientists never mentioned a stun gun.  Instead they asked questions like “The terrorist used a hypodermic syringe on the flight attendant, true or false?”  Or, the terrorist used a chloroform rag on the flight attendant, true or false?”  This is a test of recognition, not recall.  Why, if they had simply been confused by the disinformation conflicting with their memory of events, would anyone pick the chloroform rag scenario as fact unless they had lost the true memory of what they had originally observed?  In another experiment the researchers asked participants to respond “old” or “new,” to the questions, to indicate whether the events were recognized in relation to the video or the audio recap afterwards.  Memory probed this way still failed if the participants had been given the quiz before the recap to force them to recall the specific memory compared to people who were asked the same questions, but had not been forced to recall the event before the final test.

Also very telling were experiments that spaced the misinformation and recall at different times after watching the video.  As predicted by the time line of the biological process of memory encoding into long-term memory and memory reconsolidation, the method did not work if applied at the wrong times.  If the false recap was given 48 hours after watching the video instead of 20 minutes after, the original memory remained intact, exactly as happens when drugs are used too late to block protein synthesis.  They only work if given before the memory is reconsolidated.  The researchers probed their findings with 6 different kinds of experiments, all of which fit the pattern of disrupting reconsolidation of a well-established memory.

“There is a time-window for the interference to occur–beyond a certain time, when reconsolidation is complete and the memory is safely stored, the opportunity to interfere is lost,” explains Schiller, whose own research on reconsolidation of fearful memories conducted at the Mount Sinai School of Medicine is leading to the similar conclusions.  Most importantly, she emphasizes “Not any new information at the time of retrieval could modify the original memory.  It has to be directly relevant but novel or contradictory for it [the misinformation] to have an effect.”  If the false recap involved a flight attendant getting knocked out on a flight in the context of a different plot– a drug trafficking scenario rather than terrorism, for example– the new information did not overwrite the original memory.

This new research suggests that memory reconsolidation is an update mechanism.  After all, we continue to learn and add new relevant information to existing memories all the time.  Your memory of the name “Obama,” for example, has no doubt changed considerably from the time you first heard the word.  Indeed, this is why we are so often delightfully surprised when we view an old photo of a loved one.  Even though your mother, for example, is seared into our memory, you have in fact forgotten how she looked back in the days of polyester leisure suits and disco.  That memory is gone.  Memory is not a recording; memory is a construction.  To be useful to us in the present and future, that reconstruction must constantly change.

This research no doubt leads to some troubling issues related to witness tampering by authorities and the reliability of eye witness testimony.  That reality, now backed by science, is indeed very old news–something that should never be forgotten.


Chan and LaPaglia (2013)  Impairing existing declarative memory in humans by disrupting reconsolidation.  PNAS

More to explore

R. Douglas Fields  (2005)  Making Memories Stick, Scientific American

R. Douglas Fields (2005)  Erasing Memories, Scientific American Mind

R. Douglas Fields (2010)  Nightmares in PTSD:   Don’t get your blood pressure up.  Huffington post


  1. Hiraga Saito on October 5, 2013 at 12:50 am

    Please hurry up with this. There is a number of certain Anime shows which I want to forget so I can re-watch 🙂

Leave a Comment