Credit: Nic's events
Credit: Nic's events

When Memories Are Remembered, They Can Be Rewritten

It’s not often that scientists make people watch the first episode of 24 in the name of science. It’s even rarer that they pick Jack Bauer’s exploits because they wanted to show volunteers something “more true to life”. Then again, as Jason Chan dryly says, “Some of the earlier episodes were not as far-fetched as the later ones”.

Chan’s study is the latest to show how easy it is to disrupt our memories, and supplant what we think we know with misinformation. In this case, he and colleague Jessica LaPaglia from Iowa State University showed volunteers the pilot episode of 24 and then selectively rewrote some of their memories of the show’s events. For example, some of the volunteers came to believe that an assassin (Mandy!) knocked out a flight attendant with a stun gun, when she actually used a hypodermic syringe.

It wasn’t just a simple matter of saying Mandy used a stun gun. That wouldn’t have worked. Instead, Chan and LaPaglia fed their volunteers with false information immediately after they had actively remembered what they had seen. Then, and only then, did the new memories overwrite their old ones.

The trick relies on a quirk of memory that has come to light in recent years. I’ve written about it before:

Every time we bring back an old memory, we run the risk of changing it. It’s more like opening a document on a computer – the old information enters a surprisingly vulnerable state when it can be edited, overwritten, or even deleted. It takes a while for the memory to become strengthened anew, through a process called reconsolidation. Memories aren’t just written once, but every time we remember them.

This means, somewhat ironically, that the remembering something creates a critical window in which memories can be erased or manipulated. Many scientists have done this in rodents and humans using drugs or conflicting information. But these experiments usually manipulate single simple memories, such as a drug craving or a fearful association between a colour and an electric shock.

Chan and LaPaglia have now used the reconsolidation window to change declarative memories—facts and knowledge that we consciously recall. “We have people forming a very complex memory of a story that lasts 40-50 minutes and changing specific details within that larger context,” says Chan. “This is what’s new. It’s a pretty important step for demonstrating the fundamental importance [of reconsolidation] in humans.”

After showing the pilot episode of 24 to 146 volunteers, Chan and LaPaglia asked them to either play Tetris or answer memory-testing questions about the video. Twenty minutes later, they listened to a short audio recording that supposedly recapped the episode, but that secretly changed some details—for example, swapping Mandy’s syringe for a stun gun. Five minutes later, everyone took a final true-or-false test about what they had originally seen.

In this final test, the volunteers were worse at accurately recalling details that were changed in the audio recap, but only if they had previously answered questions that made them recall the video. Those who played Tetris were unaffected.

So, taking the quiz destabilised the volunteers’ memories of what they were quizzed on, paving the way for the false recap to mess with their knowledge. This worked even when volunteers correctly remembered what happened in the episode during the first quiz—the incorrect audio still changed what they thought they knew.

Through repetitions and variations of this basic experiment, Chan and LaPaglia showed that the effect lasts a long time, even if the final test followed the audio recap by a day rather than 5 minutes. But for the trick to work, the false information needs to come quickly and be very specific. If 48 hours passed between the first quiz and the audio recap, rather than 20 minutes, the original memories stay unchanged. And if the recap involved a different scenario—say, an assassin knocking out a flight attendant in the context of drug trafficking rather than terrorism—the new info never overwrote the original memory. This explains why we’re not constantly upsetting our old memories even though we’re constantly exposed to new information.

Chan and LaPaglia also suspect that people need to believe that the new information accurately represents the old set, and not if they consciously detect a factual discrepancy. “If they think there’s misleading information in here, they’ll be much less susceptible to that effect,” says Chan.

Other studies on reconsolidation have found similar results, but this one shows that memory manipulation isn’t limited to the simple products of basic conditioning, but also more complicated bits of knowledge. It supports the work of psychologists like Elizabeth Loftus, who have shown how easy it is to implant people with false memories.

It also fits with a growing body of evidence showing that, despite what people believe, eyewitness testimony is often seriously unreliable. “Say you’ve been questioned by an investigator and you recall the event,” says Chan. “In the next 15-20 minutes, you could run into another eyewitness or overhear investigators talking to each other. Some inaccurate information could update your memory.”

More positively, the study could have implications for treating conditions that involve unwanted memories, such as phobias or post-traumatic stress disorder (PTSD). As Chan and LaPaglia, “Humans are notoriously inept at suppressing unwanted thoughts.” If we try not to think about something, we usually end up thinking about it all the more. Instead, it may be more productive to actively remember what’s troubling us and reinterpret that in a new light, relying on reconsolidation to remake the old memories in a less disqueting way.

Acceptance and commitment therapies for PTSD work along similar lines, but it’s often assumed that they help people to put the past behind them or to disconnect their experiences from negative feelings. But Chan and LaPaglia suggest that such techniques might actually be exploiting the reconsolidation effect to actually rewrite the past, rather than just severing our connections from it.

Reference: Chan & LaPaglia. 2013. Impairing existing declarative memory in humans by disrupting reconsolidation. PNAS

PS: I love that reference 55 of this paper is “24 12:00 a.m.–1:00 a.m. [dvd]. Fox Television Studio, producer; 60 min, sound, color”. And reference 56 is “Neave P (2009) Tetris N-Blox (Tetris Holding, LLC, Hawaii).”

More on memory: