Why We Can't Trust Our Memories

Brain researcher Elizabeth Phelps wants to understand why you think your memory's better than it actually is.

By Kat McGowan|Wednesday, July 23, 2014
RELATED TAGS: MEMORY & LEARNING
elizabeth-phelps
elizabeth-phelps
David Friedman

Early in her restless, inventive career, Elizabeth Phelps was trying to understand the deep structure of memory by showing word lists to people with amnesia — patients who’d survived a brain injury or stroke but lost the ability to remember. When she’d talk about her work with friends, somebody would eventually ask: What do word lists really have to do with memories — the vivid images and intense emotions that flood the mind? It’s a good question. Why do you remember best the things that really matter to you? That question, among others, led her to explore emotion and the fascinating terrain where memories, fears and stress come together. 

Ever since, Phelps — a past president of the Association for Psychological Science and a psychologist at New York University — has pursued questions that get to the heart of what it means to be human. She has investigated unconscious bias, decision-making and neuroeconomics, all while exploring how emotions influence the way we learn and remember: Why are memories for emotional events so vivid, and seemingly so hard to forget? What do we forget, and why? 

To find the answers, Phelps uses the tools and principles of cognitive neuroscience, a discipline that explores how the brain gives rise to the mind: how neuroanatomy and the electrical activity of nerve cells relate to thoughts and actions.


elizabeth-phelps-2
elizabeth-phelps-2

Elizabeth Phelps has long sought to understand how memories and emotions interact.

David Friedman

Discover: Many people say we have a “rational brain” and an “emotional brain.” Is there any truth to that? 

Phelps: The idea has been there since the early philosophers. It’s intuitive. But it’s not scientific. It doesn’t make any sense. It doesn’t fit with what we know about the brain.

The way it’s described — as if we have two separate tracts in the brain, one of which does emotion and one that does reason, with separate outputs — doesn’t make any sense. It holds us back to think about these things so simplistically. 

The role of emotion is to modulate your decisions, to prioritize. You should be influenced by things that evoke an emotional response in you. It would be unreasonable not to be. What emotion is telling you is, “This is something that matters to me.” 

I think we can be much more sophisticated and nuanced about how we understand the relationship between emotion and choices. 

Philosophers have a lot to say about the human experiences you’re interested in — emotions, memories, mental life. You were always interested in philosophy, and even minored in it. Why go the psychology route instead?

P: Studying philosophy, you’d spend forever talking about the structure of knowledge, the nature of human beings, or something like that. At the end of the semester, you’d be thinking, “OK, so what’s the answer?” I started doing research on memory as an undergrad and liked it. To me, psychology was like philosophy with answers. 

memory-study
memory-study

The memory studies in Phelps' NYU lab can include giving participants mild electric shocks. 

David Friedman

What answers did you begin to find? What was your approach in studying the link between memories and emotions? 

P: I went on two paths. One was to look at what emotion does with episodic memories — our ability to consciously recollect the events of our lives, what most people call “memory.” The other path was to study fear learning — simple associations that are expressed physiologically through a bodily response such as a change in heart rate, skin conductance, arousal or pupil dilation. These are part of the fight-or-flight response, and they are easy to measure in the lab. In my lab, our first 10 years or so of research was on how we acquire fear. The second decade has been: How do we get rid of fears? 

You and others have found that strong emotions enhance memory. But we often think our memory is more accurate than it really is. 

P: The biggest misconception is that the strong feelings we have for highly emotional memories indicate they are more accurate. We’ve known this since cognitive psychologist Ulric Neisser studied memories of the explosion of the space shuttle Challenger. But it’s so compelling. For example, people think they have totally accurate memories about 9/11. I wouldn’t be able to convince you that you don’t. But like any other memory, you actually forget most of the details.

You lived and worked in Manhattan on 9/11, and launched a study with others on how emotions influence memory. How did that happen? 

P: Like everybody else, I felt I had to do something. An old friend and colleague, [neuroscientist] John Gabrieli, called just to see if I was OK, and the idea of doing a study came up. We had little booths on the street, and we surveyed people, asking some of the canonical questions in memory research: How did you become aware of it? Where were you, what were you doing, who else was there? 

There was a lot of street traffic and the weather was beautiful, and people seemed to want to be around each other. We had a good response.

What did you learn when you followed up with the same people some years later?

P: In the study, people were about 50 percent consistent in remembering whom they talked to, but they were closer to 80 percent in where they were. It makes sense. Some details, because of their usefulness, are more important to remember than others. We’re set up to capture time and place. 

One of the more interesting things about emotion and memory is the fact that it gives us a heightened sense of confidence about the memory without increasing accuracy for a lot of the details: whom you were talking to, where you went afterward. I don’t mean people forgot that 9/11 occurred. But people were wrong about all the little details about 50 percent of the time. You forget most of the details — but you don’t think you forgot them. 

So why does it feel like these intense experiences are perfectly preserved in memory? 

P: The term for that is flashbulb memory. These are memories for highly emotional and traumatic events. The term is used because people have the feeling that these memories are so detailed and vivid that it’s almost like a picture taken with a flashbulb. 

In the lab, the way we measure this enhanced confidence is that I show you scenes that are emotional, such as car accident scenes, versus office scenes, which are not. Then I show them again later and ask you: Do you remember this scene with a lot of detail and context? Or does it just seem familiar to you? This is the subjective sense of remembering: Is it a strong memory or a weak one? Do I recollect all the little details? 

We see a difference in brain regions for emotional versus neutral scenes. If it’s a neutral scene, and you remember it in a lot of detail, we see more activation in an area called the posterior parahippocampus, which is involved in remembering places or physical layout. [On the other hand] we see more amygdala involvement when you say, “Yeah, I really remember it in a lot of detail” if it’s an emotional scene. This suggests that you’re making the same memory judgment — whether the memory is strong or weak — but different brain systems are involved if it’s emotional as opposed to neutral. 

What does this tell us? We know this posterior parahippocampus is important for memories for contextual details, such as location. We know the amygdala is important for enhancing your perception and attention for central threatening details. That makes sense: We should pay extra attention to threats in the environment. The cost is that sometimes we miss the other details. In eyewitness testimony, they call it “weapon focus.” At a crime scene, if someone has a gun, people are so focused on the weapon, they don’t remember the face of the criminal. 

Maybe what’s happening is that for highly emotional situations, you’re getting really strong memory for a few details, but worse memory for a lot of other details. The fact that the memory for these few details is really strong leads you to believe you have strong memory for all the details.

So the upshot is that we can’t trust our own memories?

P: Memory does work. It’s good for a lot of things, but it’s not perfect. People have a naïve view of what memory is supposed to be like, that it should always be accurate. That’s not the case. Our memory works for us in our lives, which doesn’t necessarily mean it’s 100 percent accurate. 

If there’s a threatening circumstance, you need the memory of it to avoid it in the future. Maybe something like whom you are with is not so important, but where it happened is really important. That’s what I think is going on. 

What about you? How’s your own memory? 

P: I don’t worry about it. When you do memory research, forgetting is a normal part of memory. Ever since I started studying amnesics, I realized I don’t have a problem with my memory because I met people who really did.

And I don’t believe people who tell me they do have a good memory. I don’t buy it. When someone says, “I absolutely remember you saying that,” I say — only in my head because it’s annoying — “There’s no way in hell you remember those details.” When people say, “Oh, I have a really good memory,” I think they actually have a bias about their memory. Usually they say that because they’re trying to prove that they’re right and you’re wrong!

After you studied how fears are acquired, you focused on how they’re forgotten. Will we ever be able to voluntarily forget our fears? 

P: Initially, when you learn to fear something, you form an association — in our lab, for example, we often pair simple images, like a blue square, with a shock. Now you have a learned fear response to the blue square. This very simple notion is what Pavlov first described. 

“Fear extinction” is when you learn that something that was in the fear domain is no longer aversive. In this example, the blue square is no longer paired with the shock, so you form a second association: The event (the blue square) that previously predicted the negative outcome (the shock) no longer does. To do this, you need activity in the [brain’s] prefrontal cortex, which signals the cells in the amygdala to inhibit the expression of the fear memory. 

This is the same idea as exposure therapy for phobias. Say you’re deathly afraid of snakes. They’ll first do something minor, like show you a picture of a snake, until you can deal with that. Then they’ll show you a fake snake, or something like that, until your fear is diminished. You’re repeatedly exposed to the thing, and you slowly diminish your fear response. 

But the original fear doesn’t go away. It’s really a competition between the fear memory and the extinction memory. With the passage of time, the fear can come back, we think, because you never actually change the original fear memory. This may be why exposure therapy doesn’t work all the time. A possibly more effective method is to try to change the original fear memory. That’s why reconsolidation is so exciting.

How does reconsolidation work? 

P: Every time you retrieve a memory, it’s once again vulnerable. The idea with reconsolidation is that maybe we can target the actual fear memory — change it as opposed to inhibiting it. 

So we could really get rid of phobias or fearful memories this way?

P: Part of the problem with talking about this is people who are truly suffering can get excited about it. It’s not that we aren’t excited about it. We are. But we’re very far away from using this in treatment. 

Ultimately, our goal is to be able to translate this and a lot of our other work to new methods of treatment that are more effective. We are nowhere close to that right now. That’s why we do the research. 

You also have researched how stress influences thinking. What have you found so far? 

P: Stress has very uneven effects on brain systems. It impairs the prefrontal cortex, enhances the amygdala, and it might impair the hippocampus — all depending on the level of stress. So under stress, I’m going to get all these imbalances in decision and learning systems, and in social response systems. For example, we have one study showing that if you’re under stress, you’re more likely to attribute somebody’s bad actions to who they are, as opposed to the circumstances.

What’s underappreciated is how much the way we think about a situation influences our emotions. This is part of the inherent intertwining of emotion and cognition. Cognitive-emotional regulation is the idea that deliberately changing your thoughts about a situation can change your emotions. So you might have a predisposition to see the glass half-empty, but if you really decide to think about it differently, you can convince yourself that the glass is half-full. You’re applying a cognitive strategy effortfully to change how you appraise a situation. 

You make it sound easy, but for people who tend to see the glass as half-empty, it’s very hard to start seeing it as half-full. Why is that so difficult? 

P: We know from lots of work, not just mine, that the prefrontal cortex is required to effortfully change the way you interpret a situation, and the prefrontal cortex is highly sensitive to stress. When you’re highly stressed, you’re going to have problems with this form of regulation.

So how do we make these reappraisals depend less on the prefrontal cortex? For anything you learn that takes effort, the more you practice it, the more automatic it becomes and the less the prefrontal cortex is involved. It may be the case that when you’re highly skilled at using a cognitive strategy to reinterpret, it becomes automatic. 

For example, if you’re afraid of parties, you have to go in there thinking about all the positive things that could happen, rather than the negative things. By reinterpreting what you think might happen, you might get better feedback from other people, and that might become how you naturally view parties. The more you practice it, the more it works. As it becomes more habitual, the less likely it is that stress would interfere with it — we haven’t shown this yet, but it fits in with a lot of other science out there.

A lot of our work may be relevant to things like trauma and post-traumatic stress disorder because the neural circuitry of PTSD is built on the circuitry of fear learning and fear regulation. If we can identify ways that stress impairs us, we can think about strategies that might be more effective in situations where people are not adjusting so well. 

Exposure therapy is the best treatment we have for phobias, but it doesn’t work for everybody, and it may work less well in people who are highly stressed. 

That’s why reconsolidation is so exciting, because we have shown it diminishes involvement of the prefrontal cortex, which is impaired by stress. The hope is we can identify a faster, more effective, more long-lasting treatment. Clinically, we still have to prove that — but hopefully, we’ll get there. 

[This article originally appeared in print as "Hold That Thought."] 

Comment on this article
ADVERTISEMENT

Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

ADVERTISEMENT
ADVERTISEMENT
Collapse bottom bar
DSCOctCover
+

Log in to your account

X
Email address:
Password:
Remember me
Forgot your password?
No problem. Click here to have it emailed to you.

Not registered yet?

Register now for FREE. It takes only a few seconds to complete. Register now »