FLORA LICHTMAN, HOST:
This is SCIENCE FRIDAY. I'm Flora Lichtman filling in for Ira today.
(SOUNDBITE OF MOVIE, "GIGI")
MAURICE CHEVALIER: (as Honore) (Singing) We met at nine.
HERMIONE GINGOLD: (as Mamita) (Singing) We met at eight.
CHEVALIER: (as Honore) (Singing) I was on time.
GINGOLD: (as Mamita) (Singing) No, you were late.
CHEVALIER: (as Honore) (Singing) Yes, I remember it well. We dined with friends.
GINGOLD: (as Mamita) (Singing) We dined alone.
CHEVALIER: (as Honore) A tenor sang.
GINGOLD: (as Mamita) (Singing) A baritone.
CHEVALIER: (as Honore) (Singing) I remember it well.
LICHTMAN: Or do you? You know the phenomenon Hermione Gingold and Maurice Chevalier are describing: memories. They can be unreliable. And scientists are trying to figure out why, by distorting memories in mice. Reporting in the journal Science this week, one of my next guests says that he's able to create a false memory in a mouse's mind. But what does this mean for people? Does it suggest anything about why our memories are so easy to forget or confuse or about why our memories change over time?
If you want to talk about false memories, give us a call. Our number is 1-800-989-8255. That's 1-800-989-TALK or tweet us @scifri. Now, let me introduce my guests. Steve Ramirez is a graduate student in the Brain and Cognitive Science Department at MIT in Cambridge, Mass. He's one of the authors on that paper out in Science this week on false memories. Welcome to SCIENCE FRIDAY, Mr. Ramirez.
STEVE RAMIREZ: Great. Thank you for having me.
LICHTMAN: And Mark Mayford is an associate professor of neuroscience in the Scripps Research Institute at La Jolla, California. Welcome to SCIENCE FRIDAY, Dr. Mayford.
DR. MARK MAYFORD: Thank you for having me.
LICHTMAN: Steve Ramirez, let's start with you. What was the...
RAMIREZ: All right.
LICHTMAN: ...memory you actually implanted in these mice's heads?
RAMIREZ: Sure. So we basically started off by finding where in the brain a memory is located. The kind of memory that we were trying to find in this case was just a neutral memory of a safe box. So what we were able to do is actually reactivate the memory of that safe box and introduce new information into that memory in the form of a couple of mild foot shocks, so a sort of negative emotional event, so that the memory of that first box that was originally safe and neutral have now become a fear memory. And that's what we're calling a false memory.
LICHTMAN: So you have a mouse in a box. You find out which neurons are associated with that mouse's memory or feeling of being in that box, and then you bring it to a different box, shock it at the same as activating that first box memory. Do I have that right?
RAMIREZ: Exactly, exactly.
LICHTMAN: So is the premise here that when you recall a memory it can be overwritten or modified with new information?
RAMIREZ: Yeah. Absolutely. So we've known that this has happened. For a long time now we've known that this happens in mice and rats and humans. So memory itself is a reconstructive process. So we're used to thinking of memory as a sort of tape recorder of our past experience, and it's actually nothing like that. It's a sort of reconstructive process that rebuilds the world that we've experienced every time we recall a memory.
LICHTMAN: Does that mean that your most accurate memories are the ones that you never think about?
RAMIREZ: Ironically, yes.
(LAUGHTER)
LICHTMAN: Mark Mayford, if we see that these certain cells activate when a mouse is in a box, is that enough to know that if you reactivate them, you're actually recalling a memory?
MAYFORD: Well, before this paper and two papers from last year, we really couldn't say that, right? So much of the way we study the brain is looking at activity, right? So you're familiar with fMRI in humans. And they have people do things, and you see parts of the brain light up. And you can even get electrodes into human patients sometimes, and you can find neurons, brain cells that respond to very specific things. So a few years ago somebody described something called the Halle Berry neuron.
OK. This was a neuron that fired only when you showed pictures of Halle Berry. But this all falls into one area of trying to understand the brain, which is watching the brain, watching what the neurons do. And what this paper does is actually test whether the activity of these neurons is really doing what we think it's doing, OK? So the way I think about it when I talk about my work is I show people the Halle Berry neuron and I said, look, the thought experiment that we're going towards is if I could artificially make this neuron active, would Halle Berry pop into your head?
And I think this paper extends on some other work that was done last year by this lab and our lab, and says, yes, in fact, these neurons do represent what we think they're representing.
LICHTMAN: But hard is it to track down those neurons, the Halle Berry neuron? For example, I mean can - are we at the point where if I tell you I'm thinking of my college graduation, you can actually find the neurons that are associated with replaying that scene?
(LAUGHTER)
MAYFORD: Well, in people, most of the time in normal patients when we're doing some psychology experiment, you're not going to open their skull, and you really can't see with fMRI individual neurons, right? That's one of the reasons we like to use animal models, is it's very rare that we can look at individual neurons in humans. So we could see areas of your brain light up, but that area would contain hundreds of thousands of neurons. The Halle Berry neuron comes from studies of human patients awaiting epilepsy surgery.
So during that, for clinical reasons, they're going to have some wires put into their heads so that the surgeon can try and find out where the focus of the epilepsy is, so he knows, you know, a very circumspect area to remove. And while these patients are basically sitting in the hospital, some psychologists and scientists come in and just show them thousands of pictures, just have them flashing across the screen. And every now and then you'll hit a picture where a neuron starts flying(ph). So that's how we find that. We're not going to find that kind of specificity in you or in normal human patients - with the current technology, at least.
LICHTMAN: Steve Ramirez, I mean it seems like this study is less about false memories and more about creating a new association between two memories.
RAMIREZ: Right. So I guess another way to look at that is one of the questions that we were asking is, is this brain activity in the brain region, the hippocampus, that we're looking at, is the hippocampal activity associated with a memory, is it sufficient to reinstate that memory? And that's more or less what we see.
LICHTMAN: Right. And also modify it, right? I mean, you know I know this is in mice, so you know, we don't want to jump the gun, obviously. But do you think that your research suggests any therapies for people with post-traumatic stress disorder? I mean if you can re-enact a memory and change the emotional valence, could you use that therapeutically?
RAMIREZ: Right. So this is actually - what I'm hoping is that our work sort of sets a foundation to now ask those kinds of questions, but at an unprecedented level of specificity, which is namely at the level of individual memories. So one can speculate, for example, that the principles are now there, that you can isolate part of a memory in the brain. You can activate it, and like you said, you can change the valence of that memory. So imagine if you just ask the other kind of question, where in a patient with post-traumatic stress disorder, what if you could find that traumatic memory and get rid of the emotional valence, or maybe even update that negative valence to tilt it to be either slightly more neutral or slightly more positive. I do think that that soon will be the stuff of reality, and I'm hoping that this work now begins to sort of inspire those kinds of questions.
LICHTMAN: Mark Mayford, what do you think?
(LAUGHTER)
MAYFORD: You know, I think this is basic science. And yes, of course it's giving us insights into this kind of sort of traumatic memories and how they work. But I view it as even much more important and much broader than post-traumatic stress disorder. I mean we're really talking about how thoughts are structured, all right? And this is some of the first attempts to probe how thoughts are structured at the cellular level. And I think that's going to have ultimately broader implications for diseases of thought like schizophrenia and diseases of mood and degenerative diseases in the brain.
I mean we're really at the early stages of trying to put together the pieces of how the mind is put together. So I think it will eventually have very profound implications for human health. But it's still a fairly early stage.
LICHTMAN: I wonder too how hard it is to go from mice to humans if we're talking about complex mapping of thoughts.
MAYFORD: So...
RAMIREZ: It's interesting you mentioned that because one of the - so the brain region that we're looking at is actually one of the most evolutionarily conserved brain regions across mammals. So a lot - not everything but a lot of what we know about how human memory is structured, we've gotten from - we've learned from animal models. So it's our hope in that sense that our kind of work now can begin to make predictions on a - for a higher order mammals, such as humans.
LICHTMAN: Steve, I mean I think a lot of people have had the experience where they misremember something.
RAMIREZ: Yes.
LICHTMAN: Does this work suggest anything new about why our memory isn't perfect?
RAMIREZ: Well, so we've known that it's an imperfect process. I think what this work together with Mark's work from a couple of years ago I think that collectively what those - what our work show is that we can begin to actually dissect those little cognitive hiccups now at the cellular level and at the level of individual memories.
LICHTMAN: What do you think, Mark?
MAYFORD: Oh, yeah, I think I mean both this work and a lot of other work shows that - I mean, essentially when you think about it, when you recall a memory, you're always modifying it. And also when you just are laying down not thinking about anything, your brain isn't silent, right? There are patterns of activity in your brain. And if you have a thought and it interacts as shown in this paper with something that new that you're learning, they will both modify each other. So I think somebody said earlier, yeah, maybe your most accurate memories are the ones you haven't thought of for a while because each time...
LICHTMAN: I was...
MAYFORD: ...you recall, you're going to be modifying it.
LICHTMAN: Yeah. That's kind of amazing. If you want to keep it safe, don't ever think about it.
(LAUGHTER)
LICHTMAN: I wonder are there some memories that are just more durable than other memories that are less likely to be modified or changed?
MAYFORD: Well, I mean traumatic memories, the kinds of - I mean, PTSD is essentially memories of experiences that were so traumatic that they're intrusive. You can't not think about them, and they come on at inappropriate times, and they cause, you know, a great deal of distress. So those are extremely strong kinds of memories in a way that you don't want.
LICHTMAN: It's really interesting work, and that's where we'll have to leave it today. Thank you both so much for joining me today.
MAYFORD: It was our pleasure. Thank you for having us.
RAMIREZ: Thank you.
LICHTMAN: Steve Ramirez is a graduate student in the Brain and Cognitive Science Department at MIT and an author on that paper this week. And Mark Mayford is an associate professor of neuroscience at the Scripps Research Institute in La Jolla. Stay with us. When we come back, we're teeing off in the science of golf.
(SOUNDBITE OF MUSIC)
LICHTMAN: This is SCIENCE FRIDAY from NPR. Transcript provided by NPR, Copyright NPR.