Can We Erase False Memories?

False memories cause real problems. A false identification sends an innocent person to prison. A false childhood memory can disrupt a family. But what if there are ways to reverse false memories?

What research reveals about false memories

In a new study, Aileen Oeberst and colleagues (2021) reported techniques for erasing false memories. If you want to erase false memories, you first must create some. Oeberst and colleagues did that using an interesting technique. Like other researchers, they asked the parents of research participants to supply short descriptions of true events. But they also asked the parents to provide two possible false events — events that are plausible but that did not happen. Asking the parents to provide the false events was a nice innovation. In most studies, the researchers use a single false event with all participants. But this method provides individually plausible events.

Trained interviewers, unaware of which events were true and false, then conducted repeated sessions asking people to try and recall all the events. This basic technique always works — some people will create false memories in this type of paradigm. And Oeberst and colleagues created a lot of false memories. In their more robust methodology, just over 50% of participants constructed false memories they claimed as their own.

But could the memories be erased?

In trying to reverse the memories, Oeberst and colleagues used two techniques. First, they told people that everyone has some memories that have come from other sources. Maybe a family story or photos that we’ve adopted as our own memories. Second, they explained that repeatedly being asked about a false event can lead some people to construct false memories.

Using these techniques resulted in some people decreasing their confidence in the false memories they created. Many no longer claimed to recall the false events, although they still accepted the events as true. In this way, the techniques had some ability to erase claims of remembering. The erasing effects were small but reliable.

Oeberst and colleagues also conducted a follow-up months later. This was another brilliant innovation. I don’t know of other studies with such a long-term follow-up of false memories created in an experiment.

At the one-year follow-up, very few participants claimed to remember or even accept that the false events happened. You might wonder if the work at erasing memories simply took more time. But there was another manipulation involved. At the end of the first part of the study, the participants were debriefed. We always explain our experiments when we finish — it is part of our ethical requirements. Oeberst and colleagues informed the participants that some memories were false. They most likely informed them exactly which ones. Erasing the memories required explicitly informing participants exactly which memories were false.

So is this an effective demonstration of erasing false memories? Probably not. Let me explain.

Non-believed memories

When informed that a memory was a false construction created in the experiment, participants accepted this information. But they probably did not erase the images that came with the memory construction. They didn’t suddenly forget the narrative they had constructed. Those images and stories were still in their heads.

Instead, they re-evaluated that information. They decided they no longer believed those images and that story were their personal memories. Alan Scoboria and his colleagues have called these non-believed memories. We all have some. Things that feel like memories, but that we are fairly confident never happened.

The efforts to erase memories left the images and stories but removed belief in the memory. We can consider this a partial success in erasing false memories.

Non-believed memories leave people in an interesting and awkward position. They have something that feels like a memory. But they no longer trust that memory is true. They may even be absolutely sure it isn’t true. But the images and narrative remain.

Let me give you two examples. One is a well-known instance of false identification. Through the use of repeated pictures of an innocent man and lineups, Jennifer Thompson identified the wrong person as her rapist. Ronald Cotton was sent to jail for a crime he never committed. He was finally exonerated with the use of DNA evidence.

False Memories Essential Reads

Online Misinformation Produces False Memory

Why Some People Believe They've Been Abducted by Aliens

The disturbing part is that when Thompson was informed that she had identified the wrong man, her memory didn’t change. Her belief in that memory changed, but the memory of who raped her remained. She developed a non-believed memory. Their story is worth reading (Thompson-Cannino, Cotton, & Torneo, 2009).

And my second story of a non-believed memory comes from my own research. My students and I conducted some of the original research on the creation of false childhood memories. Our participants constructed memories of spilling a punchbowl on the parents of the bride at a wedding reception (Hyman & Billings, 1998; Hyman, Husband, & Billings, 1995; Hyman & Pentland, 1996). I have always talked about this line of work when teaching about both memory and research ethics. Once a young woman talked to me after class about these studies. She told me she had been in the study, which was cool. She said she could still see herself spilling the punchbowl. I reminded her that it didn’t happen. She agreed. But she repeated that she could still see it.

The memory remains. We may no longer believe the memory is true. But the memory remains. We can erase the belief in the memory. But the images and the stories may stay in memory. This new work by Oeberst and colleagues is nice. But what they have erased is a belief. I am not sure that we can remove the memory itself. And this is an additional risk of false memories. They remain, even when we know they aren’t true.