S3 Episode 2: Our Emotional Memory

Where were you on 9/11, and why do you remember it so clearly? Dr. Elizabeth Phelps, a Harvard neuroscience professor who studies the effect of trauma on memory, explains how highly emotional events get stored in our brains. Find out why we are so confident that these “flashbulb memories” are completely accurate, even though the evidence suggests otherwise. Plus… the “Michael Moore effect” that can influence what we think we remember.
Listen to the episode

Phil Steig: Hello. And welcome to Professor Elizabeth Phelps here to speak with us about how we form memories and their relationship to trauma. Dr. Phelps is currently a professor of human neuroscience at Harvard University and one of the lead scientists evaluating our memories for September 11, 2001. How consistent or accurate are our memories? How much confidence do we place in these memories and what kinds of memory are involved? And how do they interact with our brain?  Dr. Phelps, thank you so much for taking the time to be with us today.

Elizabeth Phelps: Thanks for having me. Please call me Liz.

Phil Steig: I will. Very nice to meet you. And just to start off. So everybody’s on the same page, can you describe to us how our memory works during a threatening event?

Elizabeth Phelps: When we talk about memory, we always talk about the hippocampus. The hippocampus gets its name for a Seahorse. So it’s a long, skinny structure in the middle of the brain. And this brain region is important for laying down memories in general.  Next to the hippocampus, sort of at the very top of it, there is sort of an almond shaped structure called the amygdala. And the amygdala is more activated by threats – by things that are salient to you, things that matter to you. And so the hippocampus can form memories perfectly well, when there’s no threat in the environment. When you are threatened and you have stress hormones – norepinephrine. adrenaline. And these stress hormones have the impact of laying down the memory more strongly. So what you take in is a little less detail because you’re focusing on just a threatening event. But then those memories are laid down a little more strongly due to the stress hormones. And so it’s really the interaction of the amygdala in the hippocampus that changes the memories for these highly emotional events.

Phil Steig: In your writings you talked a little bit about how others around you might remember the circumstances of the threatening event differently than you do. How is that? And why is that?

Elizabeth Phelps: Well, our memories are really guided by what’s important to us, and everybody differs, right? We’re all different people. We all care about different things. What you find particularly threatening may not be so threatening to me.  And so how we interpret emotional events has a lot to do with our emotional experience that we have. And so you’re taking in different information. Also, when you bring in information, you interpret it based on how you view the world. If you have different political views or different background than I do, you may see things differently, and that’s also going to affect how you remember them. So it’s not surprising that different individuals remember things differently.

Phil Steig: I couldn’t help but think about the witness of a crime. And again, experiences that occurred between the crime and the time of the court case and what the witness may be able to say, much less willing to say.

Elizabeth Phelps: The memory wasn’t designed for the legal system. Memories were not designed so that we can accurately recollect the past and all of its details. Memories were designed to allow us to do adaptive things in the future, to tell us the right things to do. So in that way, you may want a dynamic memory system, but in the legal system, we don’t. In a legal system, we want to know exactly what happened.

I think one of my favorite analogies for understanding how memory works and how memory works and how it can change over time is the game of telephone.

So if you remember this game, kids play it. I whisper something into your ear, which you whisper into the ear of somebody next to you. And if you whisper into the ear, and if somebody next to you and so on, and you go around the room, and then the last person has to say it right. So the first person knows what was said, but the last person got it sort of filtered through a bunch of different people. So I think of that as sort of the way that memory can be dynamic. You learn something and then you retrieve the memory, but in a different context with different information around you. And now that may shift a little bit, it may change a little bit. And then going forward, you may retrieve the memory another time in the future. And now the memory may shift a little bit. And so what you end up with may not be identical to what you started with. It’s going to be modified. It’s going to have some characteristics of the initial event, but it’s going to be modified by the things that are happening at the time.

Phil Steig: I’m only teasing, but I think that memory plays a large function in the role of being in love. But that’s a completely different topic…

Elizabeth Phelps: You know… and sort of how you interpret what people’s intentions were and things like that and can play a big role in whether you persist in that love. Right. It can be adaptive too.

Phil Steig: In your 9/11 study, you focus on flashbulb memories, event memories and long-term memories. Can you clarify and define what those are? How are they similar? How are they different? How are they formed?

Elizabeth Phelps: So, the term flashbulb memories was first proposed by two researchers at Harvard in 1970s, Brown and Kulik. And they were interested in the fact that all these people, when they talked about consequential events that happened in the 1960s, the assassination of JFK, the assassination of Martin Luther King, and Malcolm X. When they talked about these memories, they didn’t seem like ordinary memories. People didn’t say, oh, yeah, I remember. He died. They said I was in my kitchen. I was making coffee; I was wearing my blue dress. And then my husband came in and he looked like something horrible had happened. They’re telling you this story, these details that aren’t really typical of everyday memories. Right? And so they came up with the term flashbulb memories because they said when people recollected these memories, it was as if it was a picture taken with a flashbulb, you know, print this thing now.  It was a moment in time, like a picture.

So they argue that there are special mechanisms in the brain for these types of events. And they also assumed that these highly detailed memories that people were very confident about were accurate. But that was the assumption they made. They had no way of knowing this. And so years later, other people started to study these types of memories as well. And there was a gentleman named Dick Neyser who studied the Challenger explosion, and he did something that Brown and Kulick didn’t do. He asked people to recollect these memories soon after the Challenger explosion, where they were, what they were doing, how they heard about it, what was the first person they communicated with.  And then he brought them back later and he asked them again. And what he found was that these memories for these highly consequential, traumatic public events were not necessarily consistent. You can’t know for anybody’s personal experience if the memory is accurate, if that’s actually what happened because you weren’t there. But you can know if it’s consistent. And if it’s inconsistent, what that tells you, is it’s accurate at some point in time, either initially or later on? The fact that memories can get worse over time is not surprising. Everybody knows that. What was surprising is that people were certain their memories were correct. They were highly confident their memories are correct. And so that’s sort of what we’ve known about flashbulb memories is that these highly consequential events, your personal experience for it. That’s what we call the flashbulb memories. Your autobiographical experience.  And then people tend to be highly confident in those memories

Phil Steig: Can you please describe for us what your ten-year study on the September 11, 2001 event was?

Elizabeth Phelps: Right after the terrorist attacks of 911. I knew about flashbulb memory research, but nobody really followed it for a very long time. And this seemed like the classic event to study something like a flashbulb memory. And I will also say I was in New York at the time, and it was really hard to focus on anything else. Soon after the attack, I got a call from a colleague at Stanford University, John Gabrielli, and we decided to launch this memory study. We brought in memory researchers from all over the country.  And within a week of the attack, we had started to collect. We put together a survey and started to collect data from people across the United States, but most of them in New York City asking them about their personal experiences of 9/11 and also what they remembered about the events that happened and how they interpreted the events and a number of other questions. And then what we did was about eleven months later, we surveyed as many of those people we could find again. We surveyed them again. We did that again about three years later and again ten years later. So we were looking at the qualities of the memories, the factors that may have played a role in the qualities of the memories for both sort of your personal autobiographical memories and also for the events of 9/11.

Phil Steig: And in that you had somewhat predictable but surprising results?

Elizabeth Phelps: Yes. So one of the things we knew from earlier work on flashbulb memories was that people are highly confident in these memories. They’re highly sure that they’re right even when they’re not consistent. So we follow the consistency of these memories over time.  What we found was when we looked about a year after the attack, about 40% of the details that people gave us about their memories had changed, so they weren’t the same thing they told us right after the attack. The details were things like, Where did you first learn about this? Where were you? What you were doing? How did you feel when you first became aware of the attack? Who was the first person you communicated with? And what were you doing immediately before you became aware of the attack? So these types of details about 40% of the time they changed.

When we measured again three years after the attack, what we found was they didn’t change as much from one to three years. Often if a change occurred in the first year, usually that persisted to three years. Right. So it seems as if within the first year the memory was changing a little bit. But once somebody had their story, (you know we all sort of have our stories of what we’re doing) once you had your story, that’s the one you told us three years later and the one you told us ten years later.

What’s surprising about this is one that your memories change over time. That’s not surprising about memories in general, but people don’t think these memories change over time. So on a scale of one to five, where five is highly confident, people are almost always above 4.5 and how confident they rate their memories as being how accurate they rate their memories as being. I should say.  Now we don’t know if the memories are accurate or not. All I can do is measure whether your memory changed.  I don’t know what you were doing on 9/11 and who the first person you communicated with was right. But I do know that if you told me it’s a different person a year right after the attack versus a year later somewhere your memory’s wrong. But you just don’t think it’s wrong.

Phil Steig: Not wrong, but changed, but changed. Yeah. Maybe you got it wrong the first time. Perhaps you also talk about certain factors that affect your memory, such as media attention and swing conversation.

Elizabeth Phelps: So these effects were more apparent in the event memories than the flashbulb memories than the personal autobiographical memories. But for what actually happened. Right. So how many airplanes were there in which cities did the airplanes end up? Where was President Bush when the attack occurred? For those types of things, how much you paid attention to the media, how much you engaged in conversation? That type of thing really did seem to improve the memory performance. And that I can measure whether you’re accurate or not. Right. There is a factual answer to that question.  I can know it. So I would say a relatively small portion of people knew where George Bush was at the time of the attack. But interestingly, more people knew that in the third year.

So what happened between the first and the third year? There was a movie called “Fahrenheit 9/11” that came out where they actually showed in this movie that George Bush, when he was actually in a preschool or, I think, an elementary school reading a book and they came in and whispered in his ear.  So they showed that in this movie. And so we call this the Michael Moore effect, right. That was a Michael Moore movie. People somehow remembered three years after the attack, where George Bush was when they had no idea one year after the attack, and they didn’t necessarily remember where they learned it, right. But now it was something they remembered.

Phil Steig: It strikes me as though affective states, what’s my mood when one of these flashbulb  events occurs has to affect both our initial impression of what we think we’re seeing, but also our long term memory of that.

Elizabeth Phelps: So one of the things about our flashbulb memory studies that we wanted to do that other people hadn’t done was measure the emotional state of people. So we asked people their emotional state. Now, I’ll tell you, two interesting things emerge that we didn’t necessarily anticipate. One was that we asked people, do they remember their emotional state? And people actually are very bad at remembering how they felt previously worse than other types of details.  It was a little surprising to us, but people would often say they felt differently earlier, and usually how they said they felt earlier was something akin to how they felt now. So it’s very hard for us to imagine having another feeling at another time and have a memory for it. The other thing in this particular study. And I’m going to tell you we followed up on this. We didn’t see a strong relationship between how people reported their emotional state at the time and their later memory performance.  We expected to because there’s lots and lots of evidence from the laboratory and other studies that emotion influences memory.

A few things might be happening here.  So when we think about why these highly emotional events, these flashbulb memories, why we would be highly confident in these memories that aren’t necessarily accurate in their details. I think back to the function of memory, which is to guide our future decisions. And when we think about a traumatic event, like 911, it doesn’t really matter where I was. It doesn’t matter how I heard about this. What matters is that there can be a terrorist attack in the United States right near me, and I could be hurt, right. So I don’t forget that. But why am I highly confident about everything else? And I think the function of confidence might be to help us to act quickly and unambiguously based on that memory in the future. So if you’re highly confident about a memory, you’re not going to look for more information before you act.  You’re going to move forward and you’re going to act. So I think one of the reasons why confidence might have this enhanced sense for highly emotional memories might be because those memories, at least in terms of their gist, are accurate, and that’s really important for our future actions.

Interstitial Music


If we can alter our recollection of a traumatic event, is it possible that we could eliminate that memory altogether? And… would we really want to?

Sigmund Freud’s work with patients led him to believe that there was some process in the human mind that actively kept traumatic or unwanted thoughts at bay. This belief eventually grew into one of his main theories: we block painful memories as a defense mechanism.

Months after the terrorist attacks of September 11, a New York City deputy fire commissioner was still comforting grieving families and coordinating funerals and memorial services.

Sfx – distant sirens

All the while, Michael Regan was berating himself: he should have been at the World Trade Center during the aftermath of the attack. He finally unburdened himself to a Fire Department colleague, who told him: “you were there. You helped transport bodies of firefighters to the morgue. You don’t remember?” He didn’t – he had completely blocked out those terrible memories.  Regan’s amnesia “was a safety mechanism. I saw horrible things that day, and I didn’t want to think about those things.”

But is that a good thing? Turns out there can be a price to pay for suppressing bad memories. Some research suggests that it can block memory formation in the here and now. It explains why those suffering from post-traumatic stress disorder often have difficulty remembering recent events.

Authors of a 2016 study described how trying to forget past incidents by suppressing our recollections can create a “virtual lesion” in the brain that casts an “amnesiac shadow” over the formation of new memories.

The study was prompted by a report from a student who had witnessed the 1999 Columbine high school massacre. She admitted to suffering bouts of amnesia after the event. That may have allowed her to cope with that traumatic experience, but when she returned to school, she could not remember anything from her classes.  Dr Justin Hulbert of Bard College describes the effect this way: “If you prevent yourself from reliving a trauma, anything that you experience around the period of time tends to get sucked up into this black hole as well.”

Interstitial Music Fade Out

Phil Steig: Can we rewrite memories in our minds?

Elizabeth Phelps: There’s more and more evidence suggesting that we can. The question is, how exactly does it happen? When does it happen? When doesn’t it happen? Memory is changing. It’s dynamic. We can reinterpret something that happened at a later time. Maybe we had therapy and we now see things differently and we can start to reinterpret things and that can change our memories for those things. We can call this rewriting. I don’t think it’s completely overwriting. Some parts of the original memory may remain intact. But really understanding the details of what’s changed when under what circumstances?  That’s where the science is taking us right now. How does this occur? Why is memory changing? And how precisely is it changing both in the brain and in our memories? You know, in sort of the psychological construct of memory. And how do we understand that? And then ultimately, we want to know, how can we use that to our advantage, perhaps in the treatment of traumatic disorders?

Phil Steig: Actually, that was my next question that I wanted to ask you is, what have you learned? What results have you gotten from this study that you now think can be applied in the treatment of PTSD?

Elizabeth Phelps: I want to start with sort of the sort of part of the story that I don’t love, which is I don’t yet know precisely how it can be applied to the treatment of disorders. I feel like that’s the big question. What we’re finding out now is really that memories can change and maybe how they can change. And people are trying to use this to do things like enhanced therapy. I would say the data is moving along, but we aren’t there yet. And I say this in part because every time I talk about this research, I end up getting emails from people that are suffering, truly suffering. They think I have some magic way to help them by changing their memories. And I don’t and I really would love to, but I don’t – yet. But we are moving in that direction. We’re trying.

Phil Steig: I’ve gotten the impression in talking with Neuropsychologists however, with continuous repetition of the memory of a particular event that causes PTSD. they’ve been able to derive some superb therapies that have gotten very good results in the treatment of this problem.

Elizabeth Phelps: Yes. So there’s two things that happen. The standard treatment for something like PTSD would be a form of exposure therapy with some cognitive therapy as well. And the idea behind exposure is if you encounter something you’re afraid of and nothing bad happens, eventually, you’re going to learn that is something that you shouldn’t be so afraid of.

In exposure therapy for something like PTSD, you’ll ask people to retrieve the traumatic memories, to sort of live with the traumatic memories and realize that the memories themselves are not the things that are threatening.  And hopefully the threat responses start to diminish and do this over and over again with sort of thinking about how you can cope in that circumstance and hopefully the memories themselves becoming, you know, they’re still there, but they become not as evoking of all the trauma that is typical of PTSD.

The question is, can we learn about how we might change memories in the laboratory in a way that actually facilitates this process, changes the memory more quickly or perhaps more effectively because this process works.  But it’s difficult. It’s long, it’s difficult for the patient to participate in. They don’t want to become up with these traumatic memories a lot. And I will say that even when it’s successful, it’s not successful for everybody. And even when it is, there can be a lot of relapse as well. So we’re trying to think about ways to manipulate memories purposefully to increase the effectiveness of this treatment.

There was this idea that has come up called reconsolidation, that sometimes when you retrieve a memory, it’s once again in a state where it needs to be restored and it’s plastic, and you need new neurons forming to restore that memory.  And that’s your opportunity to change that memory. And so the question is, can I take advantage of this re-storage process? Have you come up with a memory in such a way that it’s plastic and somehow help maybe remove some of the emotional reactions that you have to the memory or change the memory and the way you view it so that it doesn’t elicit the same emotional responses the next time you retrieve it. And that’s where the science is going.

And there’s some evidence we’re doing this.  Some researchers have taken advantage of some of these techniques to reduce the “fight or flight” response when you come up with these memories later. I think we’re now trying to say, how can I reduce the negative feelings, which is slightly different than the “fight or flight” responses? So we’re trying to move in that direction. How can I make this memory feel less bad to you by actually manipulating your memory in the neurobiology of your memory?

Phil Steig: What surprised you most about the results of your study?

Elizabeth Phelps: Well, my research is on emotion and memory. So two things surprised me. No one has ever looked before at how we remember our emotions. And so I was surprised at how bad we are at remembering our emotions. Right. Like we are really bad at remembering how we felt at different times. We have a study that we did during the pandemic, where we actually did a lot of online research during the pandemic because we, like everybody else, had to go home.  So we’re doing all our research remotely, and we have a study that we’ve done that we haven’t published yet showing the same thing. People are really bad at remembering at different points in the pandemic, how they were feeling at that time. Our memories for our emotional experiences are often clouded by what our experiences are now more so than our memories for what actually happened. I mean, not that our memories for what actually happened aren’t inconsistent over time – they are – but our memories for the emotional experience is really bad. And that surprised me.

Phil Steig: For the average bloke like me walking along the street, who I don’t think has had an event that has created PTSD in my life. My overall memory. Does it affect my wellbeing, or is it other things?

Elizabeth Phelps: Well, of course, your memory affects your wellbeing, Memory is clearly important.  My mother has dementia, and for her, she can’t remember that she doesn’t walk well. So she tries to get up and go to the bathroom in the middle of the night and falls because she doesn’t remember that, right. These are the kinds of things we use memory for. Your memory can change your wellbeing, there’s lots of evidence suggesting that choosing to recall positive things can actually reduce stress.  It can put you in a better mood. So you can use your memories in purposeful ways to help yourself feel better. You could use it in purposeful ways to help yourself feel worse.

Phil Steig: So that leads to my final question, then which is, can we, as humans proactively nip it in the bud. We see something. Oh, this could be a bad memory. What can I proactively do at that moment? To change the way it’s going to affect my life in a positive or negative way.

Elizabeth Phelps: When you retrieve something, you can choose to reinterpret that event in a way that is more helpful for you. It’s a classic “The glass is half full / The glass is half empty”. Right. Let’s say something. You think of something negative in your life. You can choose to think the worst could happen, or you could choose to think this is what I learned from this. You can choose to have that reaction.

And interestingly, we did a study asking people to do that in related to covid related events that happen to them. And when they do show a benefit, we’ve asked them to come up with the event and come up with positive things that happened to them as a result of it.  Even if it was a bad event, most people would say you can think about the positive, and that has led to, at least in our study, some lasting benefit for those individuals in terms of how they feel about these covid related things that happened that were negative.

The other thing is you can use memory strategically in some ways to help you change your mood. So we know if you retrieve positive memories, just think about good things that happen to you, you become in a better mood. You’re more optimistic. It reduces stress hormones. So there’s good evidence that retrieving positive memories is actually a way to help manage stress. You can also, if you want, choose to think about all the bad things that happen to you, and it’s going to increase stress responses. Right.  But this is a choice you can make sort of how you use your memories when you’re feeling in a place where you might actually could use some boost to how you’re feeling. You could use some stress reductions. How you use your memories can change that, or it can actually make it worse.

Phil Steig: I want to thank Dr. Elizabeth Phelps for taking the time to be with us to talk out how we form memories. But most importantly, how we can positively affect even negative memories and have a more positive effect on our overall brain health. Elizabeth, thank you so much for being with us today.

Elizabeth Phelps: Thanks for having me.