Thursday, October 17, 2002
Daniel L. Schacter, Ph.D.,
Professor and Chairman of the Dept. of Psychology,
Graduate School of Arts & Sciences,
Session 4: Remembering and Forgetting: Psychological Aspects
CHAIRMAN KASS: I'd like to turn this
session over to Professor Daniel Schacter.
Professor Schacter, the floor is yours.
DR. SCHACTER: Thank you. Thanks for having
me here today.
CHAIRMAN KASS: Push the button for your
DR. SCHACTER: Am I on now? Okay, great.
I think a lot of what I'm going to say today will hopefully
amplify some of the points that came up in Jim McGaugh's
First, let me just say a few general words about the perspective
that I'm coming from. In the program it listed my talk
as "Remembering and Forgetting Psychological Aspects," and
that is true, but I substituted a cognitive neuroscience perspective,
which is probably a little bit closer to the general approach
that I take to research, which is one that combines analysis
of psychological aspects of memory with analysis of brain
systems underlying those aspects. It's the kind of approach
that Mike Gazzaniga, among others, has advocated and has,
I think, become very influential as a way of studying memory
and many other aspects of cognitive function recently.
And this interface takes place at a relatively high level.
We're interested in the level of brain systems, how different
regions of the brain interact, how different regions of the
brain together underlie the various manifestations of memory.
So that's the perspective I want to illustration, although
fundamentally I am a psychologist and will focus more on those
Second, I want to just pick up on a point that came up a
few times in Jim's talk, which is that memory, and we
refer glibly to memory as if it might be one thing, but as
we heard mentioned a couple of times, it's not. It's
much more complicated than that.
And there are various ways in which we can divide up memory
so that we're clear in our own minds about which particular
kind or form of memory we're talking about.
For example, one way of looking at it is through the idea
that there are fundamentally different memory systems operating
within the brain, and this is an idea that has received a
lot of support in the last 20 years in cognitive neuroscience,
and there are many ways of making these distinctions.
One distinction I think widely accepted would be between
a short term or working memory, the kind of memory that's
capacity limited, as Jim mentioned earlier, holds roughly
seven plus or minus two bits of information, and typically
lasts for a few seconds; the kind of memory that you rely
on looking up a phone number in the phone book and then running
to the phone before you lose that memory.
That sort of memory seems to differ fundamentally from a
longer term memory that is the kind of memory that Jim focused
on. It's the kind of memory that most people are referring
to when they're talking about memory improvement. That's
the kind of memory that we really want to target with drugs
or other approaches.
Other distinctions have mainly been drawn within the domain
of long term memory, and there are many of them partially
overlapping, but just to give an example of what I'm talking
about, one distinction that we have found useful in our work
is between explicit and implicit memory.
So explicit memory is our kind of conscious recollection
of everyday events and facts of our lives, the kind of thing
that we ordinarily think of as memory. When we use the term
"memory," we're talking about our ability to
recall our past, personal episodes and experiences to access
general knowledge. That's all within the realm of explicit
The whole other domain of memory that usually doesn't
come up in these discussions, which we call implicit memory
others have called nondeclarative memory in opposition
to declarative memory and this is a type of memory that
operates largely outside the realm of consciousness. It's
the kind of memory that might be involved in acquiring skills,
how to ride a bicycle.
And interestingly, and there are many different manifestations
of implicit memory, nondeclarative memory, but interestingly
this type of memory can proceed pretty much independently
of the explicit form.
So one type of patient who people have studied intensively
over the last 20 or 30 years, and I'll talk about a little
bit later in a different context, amnesic patients who have
a devastating inability to acquire new memories, but this
seems largely restricted to new explicit memories. They can
learn new skills just fine and show other manifestations of
implicit memory even though they have no explicit conscious
recollection of the experiences that gave rise to those implicit
So we could do the whole talk just on this topic, but I just
wanted to mention that at the beginning because I think it's
useful to frame one's thinking about what exactly is it
we're trying to improve.
Another way to divide up different aspects of memory or forms
of memory is really the one I'm going to take today, and
that is looking at the various ways in which memory can fail
or various kinds of imperfections in memory.
So when we talk about trying to improve memory with drugs,
what precise aspect of memory is it that we want to improve?
What imperfection in memory is it?
And some of the background readings go into this, but in
a couple of recent articles in a book I published last year,
I've argued that if you look at the wide expanse of literature
on the topic of memory and you ask the question, which to
my surprise few people have really asked, what are the difference
ways in which memory can fail? What are the different kinds
of memory imperfections?
I've proposed that there are seven fundamental categories
of these imperfections, and by analogy to the seven ancient
deadly sins, I've called them the seven sins of memory.
I'm going to have to apologize if you're looking
at the slides. There is a slight glitch in some of the Power
Point translation from Mac to PC. So I don't ordinarily
use these arcane symbols here on the left in my slides, but
somehow they turn up when you E-mail a Mac Power —
DR. MCGAUGH: The yen.
DR. SCHACTER: Well, maybe that's it.
DR. MCGAUGH: You've used the yen for
all of them.
DR. SCHACTER: Hey, you've given me
a new idea for the next book. That's right. The seven
So there are a few glitches that will appear from time to
time, the result of Mac to PC Power Point Translation.
Well, let's just walk through these quickly, and then
I'm going to focus on a few that I think are particularly
relevant to your concerns. The first three of these so-called
sins are different kinds of forgetting, and it's really
forgetting that we are focused on for the most part in Jim
So transience I refer to as the decreasing accessibility
of memory over time, the fact that all other things being
equal, memories will tend to fade over time unless reinforced
through rehearsal or other means. This is probably the basic
fact of memory, and I think for most of us when we think of
improving memory, we are really thinking in terms of improving
this particular feature of memory. We want to stop information
from fading out of memory.
That's not the only kind of forgetting. Absent-mindedness
refers to lapses of attention that are associated with forgetting
to do things. Here it's not so much a question of information
fading over time as the information either never getting into
the system to begin with, operating on automatic and you put
your keys or glasses down. It's those kinds of everyday
episodes that fall under the rubric of absent-mindedness,
or failing to remember to carry out an action at the time
that it needs to be carried out, and I'll come back to
that with some everyday examples in a few minutes.
Sin number three I call blocking. This refers to the temporary
accessibility of storing information so that information hasn't
faded out of memory. You're paying attention, but you
can't get to the information at the moment you need it.
A slightly different sense of blocking than Jim McGaugh talked
about, and we'll talk about one common manifestation of
The tip of the tongue phenomenon we all know, and when information
is on the tip of our tongue, we can't get at it, and then
it comes to us some time later.
The next three of the sins I think are as important as the
first three, but they all refer to situations in which memory
is present but wrong. It's not forgetting, but rather
memory distortion, which we know is a fascinating and very
important feature of our memories; that when we remember,
we don't always remember accurately.
What I call misattribution occurs when we remember some
aspect of a past event, but we attribute that memory to the
incorrect source. Perhaps we think we really did something
when we only imagined it. We know what that is, but we get
the source wrong. We think we did it. In fact, we only imagined
Maybe we hear something from a friend, but we think we heard
it on the radio last week. We misattribute our knowledge.
This can have very important implications that we'll trace
out in a few minutes, sometimes leading to a phenomenon (cut
off here in that unfortunate Mac to PC translation) known
as false recognition that I'll elaborate on.
Suggestibility refers to situations in which implanted memories
arise as a result of leading questions or leading suggestions,
the kind of work associated with Elizabeth Loftus, now a colleague
of Jim's at UC-Irvine, and as we know, this is not only
a theoretically interesting foible of memory, but something
that can have grave consequences in the real world.
We know from the controversies that have existed over the
reality of suggested memories of childhood trauma and childhood
abuse. Paul McHugh has been a voice of reason in that very
Bias refers to I think an under-appreciated aspect of memory
that we all give lipservice to, but has really only recently
been studied systematically by psychologists, and this refers
to retrospective distortions produced by current knowledge
and beliefs, when what we know, believe and feel in the present
skews our reconstruction of the past.
I won't get into that in much detail today, but I think
it's pervasive in memory. There's some elegant demonstrations
of it. I think it as much as any of these sins emphasizes
the crucial point that memory is not a tape recorder, a playback,
a video recorder, but it's a reconstruction that uses
bits of information from the past and combines that with what
we currently know and believe.
And then the final of the seven sins I call persistence.
This is a little bit different from the other six, and these
refer to unwanted recollections that people can't forget.
We covered this, I think, in Jim's talk under the rubric
of PTSD, the kinds of traumatic experiences that result in
repetitive, intrusive memories that are associated with some
of the neurobiology that we heard about earlier.
So for the purposes of today's talk, what I want to do
is walk through some examples both from the laboratory of
everyday life, primarily of transience, absent-mindedness,
and blocking because when we talk about memory drugs, I think
these are the kinds of things that we're really concerned
about, and I also want to talk a little bit about misattribution
because I think there's some interesting issues there
that relate to concerns of a bioethical nature.
Okay. Let's talk about transience. Probably the best
known experimental finding coming from psychology that bears
on transience goes back 100 years to the German psychologist
Herman Ebbinghaus, and it was Ebbinghaus who for the first
time was able to document quantitatively the fact that memory,
all other things being equal, memory tends to get worse with
time, which people of course know from everyday experience
in a casual manner, but what Ebbinghaus did was to show that
the curve of forgetting has a particular shape.
So he did a sort of unique study. He was his own subject.
He just memorized thousands of nonsense syllables and would
test himself at various times after learning and measure how
much of his earlier learning he was able to save at a later
So you could see when tested very soon, within a third of
an hour after learning, he showed 100 percent savings. Everything
that he had learned he had retained shortly thereafter, but
as time goes by, within the first hour there's a very
steep fall-off in forgetting, and then the rate of forgetting
slows down as more time passes.
Now, what's interesting about this curve is that although
it was by today's standards somewhat idiosyncratic, it
would be hard to get a study published when you're the
only subject and you're the experimenter and you have
hypotheses and so on, but this transience curve, as I would
call it, the basic properties turn out basically to hold up
across a wide variety of situations. You get variations in
how quick the fall-off is, and so on, but the basic idea that
the rate of forgetting slows down as time passes, I think,
holds over a remarkable variety of situations.
So one way you could think about attempts to improve memory
through drugs or other means would be, you know, moving people
around on this forgetting curve. We start off here. You
know, is there a drug that will keep us there? We don't
want to go down there. At least that's one perspective
Of course the risk is if we stay up here, we're going
to become like those unfortunate souls that Jim talked about,
Funes and the Borges story, Shereshevskii and the patient
he saw, which is perhaps we're going to have access to
too much information, which is a point that applies to each
of these memory's sins. I try to argue, and we may get
into a little bit later, I see them as kind of costs we pay
for benefits in memory that make the system work as well as
it does most of the time.
Maybe we don't want to be up here for every bit of trivial
information that we take in for the reasons you heard about
earlier. So you can think about transience with respect —
you can think about attempts to improve memory with respect
to this curve we're forgetting.
Now, what are the variables that affect where you are on
this curve? Well, we heard about one earlier consolidation.
Events that occur after an experience has been encoded can
have a very large effect on whether that experience holds
up over time or whether it's lost.
What I want to focus on, and this picks up on some of the
points that Jim was making in response to questions, has to
do with what goes on in the very first seconds when a new
memory is formed, when a new memory is born, the state of
memory process that psychologists refer to as the encoding
As it turns out, what goes on in the second or two or three
when an experience is encoded through the senses, taken in
related to things you already know, has a surprisingly large
impact on the subsequent durability of that memory, and it
has a lot to say about how quickly you're going to fall
down this curve.
Now, to give you an idea what I mean by encoding and how
psychologists have studied it and then more recently how we've
been able to look at it from a cognitive neuroscience perspective,
let me tell you about an old experiment, but I think an important
one, that was published some 25, 27 years ago by Craik and
Tulving, a couple of my old mentors at the University of Toronto,
who were interested in trying to get a handle on how people
encode new experiences and whether the nature of encoding
operations, the nature of the mental operations that transform
incoming sensory information into mental representations,
whether and to what extent those operations have an impact
on later memory.
Now, prior to the time of these experiments, psychologists
had not made much headway in this problem. The way they typically
studied memory is they brought people into the laboratory.
They gave them a bunch of words or nonsense syllables or pictures
or other material and said, "Here. Try to remember this,"
and they'd give them a test later on.
Now, that kind of an experiment can't really tell us
very much about how encoding processes influence memory because
I have no idea what you're doing with the material. You
could be repeating it. You could be making up images, making
up a story. I have no idea how you're encoding it.
So this is an early attempt to get control over the encoding
operations and to look at their influence on subsequent memory.
To do that people in this experiment didn't know that
they were in a memory experiment. They were just told, "We're
going to ask you some questions about words," and they
had no idea their memory would later be tested.
So at what we call the structural level of encoding, they're
shown a word. It could be in capital or in small letters.
You'd be shown words one at a time, and you'd be asked
the question is the word in capital letters.
So if you saw this word, you'd say yes. If you saw this
word, you'd say no. So now the level of encoding is focused
on the kind of perceptual level. What does this word look
like? That's what you're focusing on. That's
what you're encoding from the word on that particular
When they wanted to focus people on the phonemic level, they
would ask them a question like does the word rhyme with "wait."
So that's kind of the sound level.
So if I showed you "crate," you'd say yes,
and it wouldn't matter if it's in capital or in small
letters, and if I showed you "market," you'd
say no, and it wouldn't matter if it's in capital
or small letters. So now we've got you encoding at a
phonemic, phonological or sound level.
The final level of encoding was what they call the semantic
level. Would the word fit in the following sentence: he
met a, blank, in the street? So if I show you the word "friend,"
you'd say, yeah, that fits. It doesn't matter whether
it's upper or lower case. "Cloud," you'd
say, no, that doesn't fit. In either case, you have to
think about the semantic properties of the word to determine
whether it fits in that sentence frame.
So the subjects are sent away. They come back at a later
time and they say, "Oh, by the way, we didn't tell
you earlier, but we're actually interested in your memory
for the words that you saw earlier. I'm going to show
you these words now, and I want you to tell me was it one
of the words that you saw earlier when you were answering
And they'd mix in some words that hadn't been presented
previously in order to keep the subjects honest.
The question is: does the way that you encode the information
at study have an impact on your later ability to recognize
whether the word occurred earlier in the study list?
And the answer as you might guess is, yeah, it has a huge
effect. So here's the proportion of words recognized
on the memory test as a function of whether they earlier appeared
in the structural, phonemic or semantic condition, and every
word appears equally often in each condition. They're
counterbalanced in the way that psychologists usually do these
experiments, and you can see there's a huge effect on
level of encoding.
So if all you did was answer questions about upper and lower
case, you'd hardly remember any of those words later on.
You'd do somewhat better with the words when you thought
about sound, and then it's only with the semantic level
that you really get the robust level of memory.
So this speaks to the point that Jim made earlier that a
lot of what influences memory has to do with interrelating
new information with old information. Here you're interrelating,
in the case of semantic information, the word with lots of
semantic associations, things you know about the word, and
that provides you with a good basis for later memory.
There are more extreme demonstrations of this. Perhaps you
were thinking of this earlier. There's an interesting
study of a college student that was carried out at Carnegie
Mellon a number of years ago in which they were interested
in the question of whether you could increase memory span
beyond seven plus or minus two. So they want to see if you
could just do this with practice with an ordinary person.
So they brought a student into the lab, and they started
giving him strings of digits to remember, and he'd have
to repeat them right away, and most of the time he would just
give back seven plus or minus two, and they kept doing it
and doing it. Nothing much happened for a couple of weeks.
He would average seven as you would expect.
Then he started to show some improvements. He could do 11
digits. He could do 15. He could do 25. He could do 40.
I don't mean repeated digits. I mean random strings repeated
back right away.
After six months of practice, he could do 80 at once, and
what was going on here? Had they just increased some kind
of memory muscle or increased his memory capacity in some
The answer is no. Basically he had latched onto a very
effective form of semantic encoding. At the moment that he
started to show this improvement, the student was a runner
on the Carnegie Mellon track team, and he started to devise
a semantic encoding strategy in which he encoded these digit
strings with respect to his semantic knowledge of running,
and so if a string came along, you know, four, one, four,
five, he'd think to himself, "Oh, that would be a
pretty good mile if I had the wind at my back," and build
it, you know, starting very simple things like that. He'd
build a very complex semantic encoding network that he used
in order to achieve this remarkable memory performance.
And, you know, to really prove the point, they did a test
at the time when he was able to do 80 digits from one hearing.
They gave him a test of letter span memory, and for that he
can only do seven plus or minus two because he didn't
have this semantic encoding strategy developed.
So I think that shows (a) the power of this kind of semantic
encoding. You know, I'm not sure that you can get that
much of a boost from, you know, any of the drug agents that
we now know of, and (b) the specificity of it.
Okay. So that's some points about the psychological
aspects of encoding. One of the exciting things now about
being in the field of memory is that we're able to relate
some of these psychological and cognitive processes that have
been very carefully and thoroughly studied by cognitive psychologists
over the past 25 years, such as semantic encoding, to brain
And one of the ways in which we were able to do this in
the case of human beings is by using new neuroimaging techniques,
and there are a couple of these, really one now, that is pretty
much standing alone as the main way of doing these studies:
positron emission tomography, or PET, scanning and functional
magnetic imaging, or FMRI, which is really the technique of
In both cases basically what we're doing is measuring
hemodynamic or changes in blood flow or blood volume in the
brain, and the basic idea is that when a part of the brain
becomes very active, there's more blood rushing into that
area of the brain. So that if you can measure the blood flow
into a particular area of the brain and localize it very precisely,
which both PET and FMRI can do within a few millimeters, then
the inference would be that you could say something about
the neural activity in that region during a particular cognitive
act because that is associated with, related with, correlated
with the blood flow in that region.
So making that assumption that blood flow and these hemodynamic
factors associated with neural activity, a lot of people now
over the past ten years or more have been using imaging techniques
to look at a whole variety of cognitive processes, including
memory and including encoding.
So I just want to give you one example of an experiment.
This is one of several that show a similar result. One I
was involved in a few years ago with Anthony Wagner, Randy
Buckner, and a few other colleagues at Harvard and Mass. General
Hospital in which we adapted the Craik and Tulving type paradigm
So people would be in the scanner. They'd be looking
on a screen, and they'd be seeing a long series of words.
They would either make a semantic judgment about the words:
does the word refer to an abstract or concrete concept?
So if it was "democracy," they'd say abstract.
If it was "garden," they'd say concrete. On
some trials there's a semantic encoding trials, and then
on other trials they'd be asked the upper/lower case judgment,
the low level structural judgment.
The question is: what's going on in the brain in the
semantic condition relative to the non-semantic condition?
Can we isolate particular brain regions that are active and
related to these semantic encoding effects that we know have
such a big effect on transience?
And the answer was yes. Here are a couple of images from
that study. Let me just make a couple of points. What I'm
showing you here is a slice of the brain. If you can imagine
that you're taking a slice through the brain, you're
looking in from the top down, and we're going about halfway
down and making a cut, and that's the left side of the
brain or actually a little lower than halfway down. That's
the left and that's the right.
And these colors here, as I'm sure some of you know,
are basically a statistical map showing parts of the brain
that are activated in the condition of interest compared to
a controlled condition.
So in this upper strip what we're showing are brain
regions that in the semantic encoding condition, the abstract/concrete
task, plus the non-semantic encoding condition, the upper
case/lower case task, are showing activity relative to a low
level condition where you just fixate on an X and don't
The most notable thing here is this huge activity in the
back of the brain. That's the visual cortex showing a
large activation. There's also a big activation here
in part of the frontal lobe on the left, the lower left frontal
Now, if you move down, this lower strip is the more interesting
slide because now you're comparing directly activation
during that semantic encoding task to the non-semantic encoding
task. So now you're seeing words in both conditions.
All that varies are the mental operations, the encoding operations
they're carrying out.
You'll notice now that big yellow blotch in the back
is gone because that's occurring in both the semantic
and non-semantic conditions, that visual activation. So when
you compare the two together, it cancels out. There's
What's left over at this point is something that's
going to tell us something specifically about semantic processing,
and what we see here is that left frontal lobe activation
remains, and there's also an activation at the very front
of this visual activation near a part of the brain known as
the hippocampus, which we know is very important for explicit
memory. It's actually in the parahippocampal gyrus.
Those two regions seem to selectively activate more for semantic
than non-semantic encoding, implying that these regions are
playing some role in, and we don't yet understand exactly
what that role is, in carrying out these semantic encoding
operations that are associated with good subsequent memory.
Now, we and others have pushed this a little bit further,
tried to link up the brain activation and the memory effect
even more strongly and asked the question: suppose I put
you in the scanner right now. All right? You're listening
to me, and presumably if I tested your memory tomorrow, you'll
remember some of the things I said, and you're going to
Could I tell from your brain activity right now, could I
get any insight into whether you're more or less likely
to remember or forget a particular word or sentence that I
And that was basically the question we asked in a follow-up
study to this and more precisely was: would these two brain
regions that we saw activate more for semantic than non-semantic
encoding activate more when you're encoding some information
at the time of study that you're going to remember later
on as opposed to forget?
So in this study they were just shown hundreds of words,
and they made abstract concrete judgments about all the words.
It's semantic encoding for all of them, but we know if
I give you a test later on, you're going to remember some
and forget others. Maybe for some words you can hook it up
to what you know about the word or you get a better image
or something of that nature.
So the question is: can I tell now from the scanning patterns
and encoding whether you're going to remember or forget
the word later?
And the answer was yeah, and it turns out to be the same
brain regions again. What I'm showing here is that left
frontal, a few different views of the activation in the left
frontal lobe, the same slice I showed you before, and this
is looking in from the back of the brain.
And down here on these graphs what we're plotting is
the strength of the FMRI signal at encoding for words that
you would later remember. That's in red, and the stronger
signal versus for those you would later forget, and you can
see in all three cases these different regions within the
left frontal lobe are showing a stronger signal at encoding
for the words.
It turns out that, you know, a couple of hours later you
would remember compared to those it turns out that you would
forget. So this provides an even nicer link between the brain
activity and the neural activity, and this also shows up in
that region near the hippocampus, the parahippocampal gyrus
that I was telling you about, and those are the only two regions
in the brain that show the effect.
It's not just that on the trials where they forget the
subject falls asleep and they close their eyes and there's
no signal in the brain. Other brain regions show equal signal
for words you'd later remember and forget.
So this kind of work and a lot of other work like it, I
think, is starting to give us some clues now into the neural
processes underlying this very fundamental encoding process
that I think is central to any discussion of memory improvement
through drugs or other means.
Okay. Let me move from transience, some ethical questions
I was going to raise that I think we've already heard
a little bit of discussion about and we can get to in discussion
later, and say a couple of words about absent-mindedness,
lapses of attention that involve forgetting to do things.
Let me just give you a couple of everyday examples and throw
out a couple of points for your consideration. Here's
a picture of Yo-Yo Ma, the famous cellist, who a couple of
years ago had a kind of unnerving encounter with memory where
he put his cello in a taxicab, took a ten minute cab ride,
got out of the cab, and then walked away without his cello.
Fortunately for him, this episode of forgetting was nullified.
The New York City police got right on the case, and later
that day here he's shown reunited with his $2.5 million
Now, this would not appear on the face to be an example
of transience. Presumably the information hadn't faded
out of his memory within minutes that he put the cello in
the trunk. Rather, he wasn't reminded at the time he
needed to carry out the action that he needed some cueing
to remind him that the cello was in the trunk.
Presumably had he said, "Where's your cello, Yo-Yo?"
he would have immediately said, "Oh, it's in the
So the forgetting here, I would argue, is based on totally
different mechanisms than is the forgetting in the case of
transience, and one thing we know about this kind of absent-minded
forgetting is that it can take extraordinary forms when people
(a) are operating on automatic and (b) are not cued to carry
out an action at the moment they need to cue. So with memory
for doing things in the future, it doesn't help very much
to remind yourself, "Oh, I've got to pick up"
— or it may not help very much — "I've got to pick
up bread and milk on the way home," if you're not
reminded later on at the moment you need to carry out the
action. And there's a lot of interesting research that
backs up that point.
Just an extreme case with, I think, some interesting ethical
implications, so I couldn't resist talking about it with
you today, briefly was that of a — and there have been several
cases like that around the country in the last couple of years
— a woman by the name of Carie Engholm, a high level hospital
administrator who drove her son and her daughter to work one
morning, her seven month old daughter who she put in the back
of a van, dropped the son off as she did every day. She was
not accustomed to taking her daughter; drove on to work, forgot
about the existence of her daughter, and unfortunately the
daughter died in the van and was found later that day.
Now, there's a lot of discussion, you know, how could
Mom forget that baby. How is it possible for someone to forget
something of this nature?
I personally would argue that it's an extension of the
Yo-Yo Ma principle, that without cueing at the moment you
need to carry out an action and operating on automatic, it
is astonishing how far forgetting can do.
Eventually she was found not guilty because there was the
question of is someone responsible for their memory system.
Do you take responsibility for the fact that one of the foibles
of memory is that if you go on automatic and are not cued,
you can forget extraordinary things? Are you, therefore,
responsible for setting things up in advance to make sure
something like that never happens?
The judge eventually decided that she was not guilty. There
were various reasons for that, some having to do with the
exact nature of the charge which required proof that Engholm
knew she left her child in the van when she clearly didn't,
but he then concluded that forgetting is an involuntary process
that can't be knowingly or recklessly done. A person
either forgets or remembers.
So I throw that out to you in terms of maybe another angle
on some of the ethics of forgetting. When are we responsible
for our memories?
If I show you a 20 word word list and I give you a test and
you only remember five words, I don't accuse you of being,
you know, a flawed moral agent because you can't remember
them all. Does that logic apply here?
I think that's better for you guys to decide than me.
Memory sin number three, I want to say just a little bit
about blocking and then conclude with a few words about misattribution.
Blocking is the kind of memory failure that occurs when we
haven't lost anything from memory and we know the information
is there, and it's not a matter of paying attention.
There's an example of this, again, a flawed Power Point
slide of British Deputy Prime Minister John Prescott, who
was at a press conference a couple of years ago where he was
forced to justify the cost of the Millennium Dome, this extremely
expensive stadium that was built in London. I think it is
now being taken down.
And he was asked, "Well, where did the money for this
come from?" in his overrun budget so much.
And he said that money came from the — you know, what do
they call it? He just couldn't come up with this word,
and then he said, "Oh, was it the raffles? No, it wasn't
the raffles. What was it?" and then finally someone
came up and whispered to him, "the lottery." It
was the lottery.
So that's kind of a classic tip of the tongue blocking
phenomenon. Clearly he knew the word. It hadn't fallen
out of his memory, and this tip of the tongue state is quite
pervasive. It's something that is known in virtually
There's an interesting article published by the psychologist
Bennett Schwartz who noted that 88 percent of the languages
that he surveyed used the tongue metaphor to describe this
kind of temporary retrieval block, "on the tip of the
tongue," "on the tongue," "on the top
of the tongue," "on the front of the tongue,"
my favorite, "sparkling at the end of the tongue,"
"in the mouth and throat," all aspects of the TOT.
One thing we know about the TOT state, this blocking type
of forgetting, is that it's a very active state. One
comes up with incorrect items like raffles. You've got
to make decisions about whether those items are correct or
not. Are they just leading you down the garden path?
William James — being in William James Hall I have to have
one William James slide for any talk I give, and he always
has something interesting to say about something in psychology
— describes it well. "Suppose we try to recall a forgotten
name. The state of consciousness is peculiar. There's
a gap therein, but no mere gap. It's a gap that's
intensely active, a sort of wraith of the name is in it beckoning
us in a given direction, making us at moments tingle with
the sense of our closeness and letting us sink back without
the longed for term. If wrong names are proposed to us, this
singularly definite gap acts immediately as to negate them.
They do not fit its mould."
So that really captures the phenomenology of the TOT state.
It's as if there's some cognitive monitor in there
that knows what we don't know. We're not quite sure
how that works, but, again, we've been starting to get
some insight into some of the brain activity during blocking
through neuroimaging studies.
We recently did an imaging study of blocking, of tip of the
tongue blocking, where we induced TOT states while people
were in the scanner by giving a cue such as War and Peace
plus author, London plus river, Aida plus composer. Those
may seem obvious to some. Can you think of the answers for
each one? Obvious to some, maybe not obvious to others, but
about ten to 15 percent of the time you'll get a TOT state
for Tolstoy or for Thames or for Verdi.
So we put them in the scanner. We run hundreds of these
by them, and interestingly what we find is that there are
certain parts of the brain that activate selectively specifically
during the TOT state, and by and large, these are regions
that in other studies have been implicated in kind of cognitive
monitoring processes. They're parts of the brain that
get going when there's conflict, when we've got to
make cognitive decisions between conflicting alternatives.
One was a region of the brain in the middle of the right
frontal lobe, and the other is in an interesting part of the
brain tucked away in the middle of the brain called the anterior
cingulate that gets activated in all kinds of imaging studies
and seems to be related to this monitoring of cognitive conflicts.
And in this study these brain regions were activated only
during the blocking state. They didn't show significant
activation when people knew the answer or didn't know
So there's some initial insights into some of the neural
activity associated with blocking. It's an interesting
question as to whether this kind of blocking is something
that merits attention and attempts at memory improvement.
Certainly it is one of the most common subjective complaints
of memory loss as we get older, is blocking particularly on
the name of familiar people. There are cognitive ways this
can be approached, and as far as I know it's really never
been looked at through drugs and, you know, maybe shouldn't
be looked at.
Finally, I just want to say a few words about misattribution
because I think the whole realm of memory distortion, though
it's really not a big issue when it comes to memory improvement
people are, whether they should be or shouldn't
be, people are generally not looking for ways to make their
memory less distorted. They're looking to remember more,
and I think people often take accuracy for granted when perhaps
Let's just say a few words about misattribution. We'll
go through those paintings and go directly to misattribution.
A number of you may remember — probably all of you remember
— that when the Oklahoma City bombing occurred back in 1995,
there was a search for two suspects, John Doe No. 1 and Joe
Doe No. 2. John Doe No. 1, of course, was Timothy McVeigh.
John Doe No. 2 was never found, although there were bulletins
issued by the FBI shortly after the shooting in a search for
a person who looked like this.
Well, it turned out he wasn't found because he didn't
exist. At least he didn't exist as a suspect in the case.
He existed as the figment of the memory of a guy by the name
of Tom Kessinger, who was a mechanic who worked in the body
shop where McVeigh rented the van that he used to carry out
the bombing. And he distinctly remembered McVeigh coming
in with this guy.
Well, it turned out he was also there the next day when
somebody, who the FBI later discovered looked like McVeigh,
entered the body shop with a guy who fit this description.
He was an innocent Army private by the name of Todd Bunting.
What had happened here was a classic memory misattribution
error. Kessinger was right. He had seen that face before,
but he misattributed his memory in this case to the wrong
time. He mixed up the time that he had seen the two and thereby
committed this error.
Now, this sort of memory error can have very serious implications
for eyewitness testimony. It is the kind of thing that is
often involved in eyewitness misidentification as, as we know
from studies of people who have been wrongfully convicted
and exonerated on the basis of DNA evidence.
Approximately 90 percent of these individuals who have been
studied were put in prison wrongfully, largely or entirely
on the basis of eyewitness identifications, often involving
this kind of memory misattribution. So it's certainly
something that we need to pay a lot of attention to.
One way psychologists have studied this recently is through
a very simple but effective paradigm that came originally
from the 1950s and then more recently in the 1990s, rediscovered
by Roediger and McDermott where if we had time we could easily
induce this phenomenon in most everybody here.
You present people with a bunch of words that are related
to one another, such as candy, sour, sugar, bitter, good,
taste, tooth, nice, honey, soda, chocolate, heart, cake, eat,
pie, and then you later give them a memory test where you
present words that were on the list like "taste,"
and most people will correctly say that "taste"
was on the list; unrelated words that weren't presented
like "point"; and then the interesting case which
is associatively related theme words or critical words like
"Sweet" wasn't on that list, but if you do
this experiment, if we had time to do it here, what you find
is that most people swear up and down that they heard the
word "sweet," and Paul knows this because he's
had this done to him.
DR. MCHUGH: Exactly.
DR. SCHACTER: Here's an example from
one study we carried out, Ken Norman and I carried out, a
few years ago with college students and old people who were
exposed to a bunch of associate lists like this, and this
shows you that both the young group and the older group in
their 70s about 75, 80 percent of the time are correctly saying
that words like "taste," which were on the list,
really were on the list.
The problem is they're also claiming with high confidence
that words like "sweet" that weren't there were
also there, and as you can see, the older group shows an increased
susceptibility to this. I'm sure that it had nothing
to do with you showing the effect.
DR. SCHACTER: And which we thought was
an interesting discovery at the time, and I can't resist
just giving you this quick quote that shows that we were scooped
by 100 years by none other than Mark Twain who said, "When
I was younger, I could remember anything whether it had happened
or not, but my faculties are decaying now, and soon I shall
be so I cannot remember any but the things that have never
happened. It's sad to go to pieces like this, but we
all have to do it."
What's going on here is this, again, illustrates the
fact that memory is not just a tape recorder or computer or
whatever; that what we're doing is we're kind of constructing
a mental representation of the general sense or the gist of
that list, and "sweet" really fits in well with
everything you heard.
So later on when you reconstruct what you heard, that seems
to fit so well that you're absolutely sure that you heard
Now, we've been interested in this for a number of reasons.
In part, we've been interested in what is going on in
the brain when people make this sort of memory error.
And through a number of approaches, we've gathered evidence
that the hippocampus, the region of the brain that I told
you about earlier that we know is so important for remembering
things that really did happen, also seems to be involved in
this memory illusion.
For example, patients who have damage to the hippocampus
and music patients that have great difficulty remembering
what really happened also how less of the memory illusion
than healthy people do. They show about half of the memory
illusion. They're not fooled like you and I are.
Why? Because we think the hippocampus has something to do
We've also done brain imaging studies that I wanted
to get to and conclude on because I think they raise some
issues that might be within the purview of this group, namely,
about the ability to use brain imaging to tell whether someone
is reporting an accurate memory of the past or a false memory.
So this is a study we published last year with Roberto Cabeza,
where we scan people while they were making judgments about
whether words were on the list they heard earlier. They had
been given a whole series of these associate lists, and they
false alarmed to words like "sweet" that really
What's interesting here is that you can see these are
the FMRI signals, that there's just as much FMRI signal
for the false words like "sweet" in the hippocampus
as there is for the words that were really there like "true."
There's a nice increase in signal compared to words that
had nothing to do with what were on the list and people easily
say no to.
So in both of these cases the hippocampus is lighting up
roughly equally. It's fooled into thinking that "sweet"
was on the list perhaps because it's remembering the semantic
gist of what was there.
But there's another part of the brain in this study,
that parahippocampal region we just talked to posterior to
the hippocampus that shows activity only for the true words,
not for the sweet words. We're not quite sure. We have
some hypotheses about why that occurred, and we've seen
similar things in a couple of other studies.
Now, the interesting — the broader societal question here
is if you want to know whether someone is remembering accurately,
can you just put them in the FMRI scanner and tell? And
the short answer to that would be no. This is data that comes
from a group study. You have to average across many people
in order to get these effects. If you look at any one individual,
you're not going to see very much, but I think it is a
technology about which we'll be hearing more, the use
of these imaging devices for distinguishing between truth
and deception in the context of lie detection, perhaps between
true and false memories, and although we're certainly
not at a stage we're even close to being able to use this
for practical purposes, it is something, I think that at the
interface of technology and memory that you might want to
I'll stop there.
CHAIRMAN KASS: Thank you very much.
Someone gets us the lights. Thank you.
Questions for Professor Schacter? robby George.
PROF. GEORGE:Yes, just very quickly, Professor.
On that false memory with the "sweet" are the Paul
McHughs of the world claiming to remember seeing the word
or are they simply reasoning that it must have been? Have
you somehow tested for that, controlled for that?
DR. SCHACTER:Yes, that's a good question.
There's a lot of data that speak to that and indicate
that they're not just reasoning and guessing that it was
there or just saying, well, probably it was there because
it fits with everything.
There are a number of studies that various people have done
and some of our own work that show how it really is for whatever
reason people are having this subjective experience of memory.
There's a lot of debate in the literature about what
the mechanism is, but they will claim if you give them a choice
of saying one of the two things, one of the following two
responses: you're saying, yes, this word was in the list
because you have the specific recollection of having seen
or heard it, depending on whether the presentation was auditory
or visual, or you're saying, yeah, it was on the list
because you just know it was there, and you don't have
a specific recollection.
And sometimes people will say, "Oh, I just know it."
They give the remember response just as much to these words
as they do to the words that were really there.
So in other words, they say, "No, I have a specific
recollection of it," just as much to the sweet words
that weren't there as to the words that really were there.
If you do an experiment where you have half of the words
presented by a man and half of the words presented by a woman
and you say to people, you say to the subjects, "Okay.
Tell me if you remember the word, and then if you really remember
who told you, a man or a woman, tell me man or woman. If
you don't really remember the source of the word, don't
write down anything," they'll write down man or woman
with high confidence just as often to these false words as
they will to the ones that were there.
So it really seems to be subjectively they're absolutely
convinced that it was there. That's my reading of the
CHAIRMAN KASS: Charles.
DR. KRAUTHAMMER: I just had a question on
intrinsic and extrinsic memories that you mentioned earlier.
Can a Korsakoffian learn to ride a bicycle?
DR. SCHACTER: We don't know the answer
to that exact question, but by implication the answer would
be probably yes, barring — yeah, barring any general cognitive
deterioration. They can learn a whole variety of motor skills
and laboratory tasks that, you know, for example, if you have
people trace a moving stylus they get better at keeping time
on target with practice, and these patients will show that
same kind of improvement.
We've even done experiments years ago where we have
shown that they're able to learn some basic computer programming
skills. Even though they come back to the lab, they have
no idea that they've done it. They don't realize
that they've ever worked on a computer, but if you get
them going on a test, they'll start to do it, and they're
kind of amazed that they can.
DR. KRAUTHAMMER: And have you studied brain
injured people whose injuries happen to be in the two areas
that you identified with encoding, that left temporal —
DR. SCHACTER: The left frontal and the
DR. KRAUTHAMMER: Left frontal, right.
DR. SCHACTER: Parahippocampal. You know,
there's a little bit of literature on the left frontals
that suggest an encoding deficit. It's hard to find people
that have a specific parahippocampal deficit. You just don't
see those patients very often. So I'm not aware of any
literature on them.
But that's one of the tasks in this general area, is
to try to map the brain imaging work onto the lesion work,
and they don't always go together. I wish I could say
that they did. They often do, but they don't always.
DR. KRAUTHAMMER: Because if I could just
add, because most of the knowledge and until this new technologies,
this stuff I learned 25 leap years ago from Normal Geshwin
was based entirely on pathology.
DR. SCHACTER: That's right.
DR. KRAUTHAMMER: The patient was missing
an area, and then you deduce what that area was doing. Now
you can see it in vivo. I'm just wondering if they correlate.
I mean, there's the whole entity you'd want to be
correlating with now.
DR. SCHACTER: Right.
DR. KRAUTHAMMER: Active areas versus pathology.
DR. SCHACTER: Yeah, they don't always
correlate. They correlate often enough in important ways,
I think, both, you know, to validate the technique and, you
know, give us some confidence that, you know, we're looking
at some of the same things that the earlier neurology types
But I think they're very different kinds of evidence,
and you probably need them both, and in the sense with bright
imaging we're looking at areas that are active during
a task, but it doesn't tell us that they're necessary.
They'd just be coming along for the ride or they may have
something to do with the subject's reaction to the test
where monitoring is in the case of blocking, whereas the neuropsychological
patients tell us, you know, more what areas are necessary.
So I think, you know, each one adds to the other.
CHAIRMAN KASS: Could I ask about — this
may be perverse of me to ask about something you chose not
to talk about today, but if I'm remembering rightly, persistence
was the last of the sins —
DR. SCHACTER: Right.
CHAIRMAN KASS: — is not somehow a fault
either of failure, kind of forgetting, or of distortion, but
it's rather a dismaying accuracy that we remember things
we would rather not remember.
First of all, am I right about that?
DR. SCHACTER: Yeah.
CHAIRMAN KASS: Is that what you meant by
DR. SCHACTER: Yeah, that's what I meant,
and I thought we'd touch on it a little bit earlier on,
but that's exactly right. Persistence, as with the other
sins, I wouldn't see as some fundamental shortcoming in
memory. It's the price we pay for having a memory system
that, you know, responds to emotionally arousing situations
in the way that Jim McGaugh described earlier and in a way
that presumably works to our advantage most of the time, and
then it's good to remember, you know, these terrifying
events and, therefore, avoid them in the future.
The down side is, you know, you can be plagued by such events.
CHAIRMAN KASS: Let me see if I could Ó-
and this would be to invite Jim McGaugh back into the conversation
as well to tie up this discussion with the earlier one.
As I think about the possibilities for altering memory, the
enhancement of memory whether it's as good as just working
harder or not, there are lots of our fellow countrymen who
would want it if it were easily available. It doesn't
bother me a whole lot.
But the thought that you could selectively block and erase
memories and particularly not just for the prevention of post
traumatic stress syndrome, but for what when Paul McHugh's
colleagues finish with the DSM-5 or 6 will include painful
memory disorder, shameful memory disorder, because we can
do something about that, and there will be reimbursement for
it, and therefore, it will be in the book.
And it's not just the really horrible things, but you
know, within the last week I'm sure all of us have probably
some things, you know, embarrassments, sorrows. The temptation
will be to have some kind of amnesia for these things or,
as we were talking about it at the break, amongst the young,
one could lose one's inhibitions for shameful behavior and
not have to remember it afterward. It's a kind of win-win
So I'm wondering —
PARTICIPANT: It's called alcohol.
CHAIRMAN KASS: It's called alcohol,
but it comes with disturbance.
In any case, I'm wondering about how you would sort of
talk through that aspect of memory and whether we could, in
principle, distinguish those things that persist which say
as a therapist one would be happy to somehow extirpate if
one could from the temptation that all of us would have to
our peril and, in fact, to our detriment, somehow wipe out
the things which we would rather not remember and make us
wince years and year after.
DR. SCHACTER: I think there's one distinction
that needs to be drawn between some of the work that Jim was
talking about with propranolol and alcohol in that in the
former case it's not a question of wiping out the memory
because people can remember what happened. It's taking
the sting, the extra emotional edge off of the memory so that
you remember it, but you don't —
CHAIRMAN KASS: It doesn't bother you.
DR. SCHACTER: You're not overwhelmed
by it, right.
So that line of work as far as I know has not really gotten
into memory erasure, which as he points out, you know, I guess
we already have alcohol and we can make our judgments about
the effectiveness of that strategy.
I'll turn that one over to Jim, but I don't see that
as a new issue that is suddenly being driven by new drugs
or new developments.
DR. MCGAUGH: No. I think I may have overstated
the ease of doing it in the first instance. That is, recall
that the only experiment with humans was not erasure with
propranolol, but it was dampening the development of PTSD
so that the subjects could still remember the event, but they
didn't have the emotional outflow, and it didn't take
over their lives.
Now, that takes a long time to do. So it isn't as though
I did something horribly embarrassing yesterday and I can
pop a pill and that's going to be gone. If it was really
so horrible, it's going to flash into my head, recur and
recur and recur so that it begins to take over my life; then
that would be a case, I think, in which a little propranolol,
if the studies bear out, might be of value.
But there's something else that I didn't emphasize,
and I should put in the equation. This thing works equally
well for good things that happen, as well as for bad things.
So I think when we think of the strong memory's persistence,
I'm sure that Nobel Prize winners remember what they were
doing — well, in the U.S. most of them were sleeping, but
where they were and what they were doing when they got the
call of winners and prizes. Anything as much coveted, you
remember birthdays and weddings and all of these kinds of
things; they stand out.
So the pleasant side of this works just as well as the unpleasant,
but there's never a quick fix. It's not as though
— well, there would be. You walk around with an ECT machine
in your backpack and something terrible happens. Deliver
yourself an ECT. That's the only thing I can think of
that would do the job.
Short of that, it's all going to be a dampening. Would
you agree with that?
DR. SCHACTER: Un-huh, yeah, I would agree
And you know, it's an interesting point to consider,
you know, when you consider some of the adaptive role of some
of these intrusive memories. Actually I was running out of
time, but had it in my last slide in case I had time to get
to the persistence issue a picture of a Holocaust survivor
from a real interesting series of art works that's been
done on this topic by an artist by the name of Jeffrey Wollman,
who interviews Holocaust survivors about their stories and
then takes the picture and does it in interesting ways.
And in this particular slide, if you can read the small writing
about the woman's story, I mean, she talks about how,
however painful it was, however critical it was for them,
you know, to keep that memory alive and, you know, the whole
issue from an adaptive perspective of how, you know, dealing
with a memory helps you adapt to the traumatic event.
And then the question of whether it's better or worse,
you know, to be dealing with it with the sting or less of
the sting, it's hard to address, but I think that's
one of the issues.
DR. MCGAUGH: Yeah, I think the critical
case is that of when it becomes incapacitating.
DR. SCHACTER: Right.
DR. MCGAUGH: That's the issue because
certainly I think in most cases having a very strong memory
of a horrible event has adaptive consequences. You want to
stay awake. You want to be on guard. You want to know what
to do, what not to do, and so on.
And as a matter of fact, they had a hell of a time getting
post traumatic stress syndrome classified as a psychiatric
disorder because it was matched up against malingering. You
know, it's not really true; it's not debilitating.
What you want to do is just take advantage of the VA system
and take all of the benefits, and so on.
And it took a long time for them to get to recognize that
in a certain percentage of the cases, their lives were really
debilitated by it, whereas in most cases it's not. So
that they have now matched cases of people who have severe
automobile accidents. All right? Twelve to 15 percent of
those cases, as my reading of the literature, will develop
some PTSD. All right? Some will have none even though they've
had the same experience. They certainly will remember a lot
Now, what do you do about those 12? Well, in most of those
cases or over half of those cases, those will resolve in about
eight months, and then you're left with a residuum of
a very small number of people whose lives have been damaged
over the long term. Now, that would be the focus group, I
think, right there.
CHAIRMAN KASS: Thank you.
I have Bill May, Alfonso, Rebecca, and Mike Gazzaniga.
DR. MAY: Well, in these last two sessions
we verge ruminatively on topics that in so many ways move
beyond the reach of medicine, but no less interesting for
I think about the two responses to the past and remembering
of nostalgia and remorse. I've just gone through a 50th
reunion, and so you ride down the moonbeams of nostalgia.
There is a faculty member who is now in his mid-80s who wrote
a lovely letter saying, "Affection grows as memory fades."
DR. MAY: That's a very nice comment.
Another professor on another occasion talked about the sickness
of nostalgia, that living in a lost world. That's a very
interesting human problem, and it's not just ideological
bias in that case. It's a kind of disconnect with the
present that afflicts people in the course of time.
And remorse, your category persistence very much relates
to that. We're not now talking about the medical danger
of eliminating past memories, but in fact, as I recall, the
Catholic Church in its sacrament of penance vividly understood
that remorse is very dangerous because it is that relationship
to a past which stings and you bite yourself all over again
in remembering it. So it's unavailing, and there's
no way of moving beyond it and moving forward into the present.
And perhaps just as dangerous as writing out memory is the
reliving of a past event that is so wincing in memory that
one engages in a kind of suffering all over again, which is
unproductive of a future.
And I guess it remains to us, sir, as to how any of this
relates to the questions before us in bioethics. Fascinating
CHAIRMAN KASS: Thank you, Bill.
DR. GÓMEZ-LOBO: This is on a slightly
different topic, I think, but I'm really delighted that
I have a chance to ask you these questions.
The topic of memory, of course, fascinated the ancients.
There's a lot of memory in the Platonic dialogues and
in Aristotle, and I'm thinking about — and I'm going
to formulate a question — the following.
When we talk about the blocking of memory, if I understand
it correctly, there is a question of retrieving something,
and if we retrieve it wrongly, if for, say, War and Peace
we say Dostoevski, we can do that because there's something
that we have not forgotten, right? So there is an actual
belief held by reference to which we match whether the memory
was correct or not.
Now, this is independent of whether it is factually correct.
So memory seems to entail this matching or making coherent
two beliefs of some sort.
Now, my question, what I'm really intrigued with is this.
In the contemporary study of memory, is there such a thing
as remembering something which we did not first acquire, for
instance, through a sensible experience?
I've noticed that memory and the response to robby George
is not restricted to the sensory experience.
Now, the example I have in mind is this. Is it correct to
speak of memory, for instance, in the retrieval of the set
of natural numbers, for instance? We've had no encounter
with that, and yet somehow that seems to be stored in our
DR. MCGAUGH: What was the word? What did
you say? The retrieval of what?
DR. GÓMEZ-LOBO: The series of natural
DR. MCGAUGH: Yeah.
DR. SCHACTER: Well, I think this partly
gets into the realm of the distortion and false memory, that
you could say we know from experiments that people can be
induced to remember episodes from their past that by all objective
accounts did not occur. College students, you just ask them
about it several times, and a certain proportion of college
students will claim to recover a memory of spilling punch
on the groom or bridge at a wedding when they're five
years old, and as far as we can tell, this event never occurred.
So, you know, it doesn't make sense to talk about that
as a false memory. Well, if you define memory or part of
memory as the subjective experience of what occurs at the
time of report or retrieval for whatever reason it's there,
then, yeah, that makes a lot of sense.
If you define it with respect to tying it to an event that
occurred in the past, then it would be an oxymoron of sorts.
DR. GÓMEZ-LOBO: Yeah, I think I
was taking care of that. What I'm curious about is the
kind of knowledge that in philosophy we call a priori knowledge.
In other words, the fact that we can say a lot and handle
a lot of problems in this series of natural numbers, which
we have never learned, and in the standard sense, never stored
in our minds, but maybe this question is a question in epistemology
DR. MCGAUGH: Yeah, we wouldn't refer
to that as memory even in the same way that we wouldn't
refer to our extraordinary capacity to learn language as memory.
It's a capacity that we have, but it's not memory
until we've learned something.
CHAIRMAN KASS: Rebecca.
PROFESSOR DRESSER: A couple of comments.
With regard to the ability to reduce the distress associated
with an embarrassing experience or a traumatic experience,
obviously as with many of these things we're discussing
the problem is line drawing because, I mean, we don't
want to remove that distress for social reasons in some circumstances.
For example, I think that's where a lot of empathy comes
from. That is, when we have an embarrassing experience, we
develop empathy for others who have a similar experience.
Also, shame or feeling of responsibility for consequences,
I mean, once you do something stupid or sloppy or, you know,
that's a lot of growing up, development. So we want some
of that sting. So the question is: what is dysfunctional
There probably is some sting that we would rather not have
as individuals, but it's good for the rest of us that
others have it in determining, you know, when sting ought
to be removed. This is, I think, a big problem.
The other point that I thought of in response to your presentation
was I think we've been discussing the use of these things
in situations where individuals might want to use them or
parents might want to use them and other people might have
reservations about whether that should happen.
But here with the question about whether something is an
actual memory, there might be cases where it would be in the
interest of society to force someone to have an FMRI or a
PET scan or some other intervention in order to discover whether
it's a false memory or not.
And the person might say, "Well, I don't want that."
And so there could be a coercive potential use of that, and
we would have to think about how should, you know, autonomous
choice fit into the situation.
DR. SCHACTER: I think you have all of these
same issues that surround lie detection today, you know, from
that sort of issue, but also the whole question of efficacy
and just how good the technique is.
PROFESSOR DRESSER: Right, yes.
DR. SCHACTER: That's why lie detectors
are not admissible. They're just not good enough, and
brain imaging is nowhere near the level even of lie detection.
So I think one would end up revisiting pretty much the whole
range of issues that has come up with lie detection.
DR. MCGAUGH: Could I comment on the first
part of that?
I think that this sort of was the view of the Veterans Administration
on the complaints of the GIs who were suffering so terribly,
you know. "Suck up your guts. Come on. Everybody has
But there's a difference between the level, the intensity
of this disorder in those people who really present as opposed
to people who just said, "I had a very bad deal in Vietnam,"
or on the road the other day or whatever it is.
But there's some point you can draw a line and say, "You
can cope with this and for some reason you can't."
And there is no treatment for long lasting post traumatic
stress disorder. I mean, it just sits there.
While you were talking, I was thinking about the case of
anxiety. It's good for us to be anxious about things.
It's good because it helps us prepare. We have to anticipate
And yet there are people who are so anxious that they can't
anticipate the consequences in an adequate way. And so benzodiazepines
are a drug of choice to deal with that end of the distribution.
People will become incapacitated because of the anxiety, but
they're not drugs, in my opinion, aside from Viagra that's
a most widely used drug. There isn't any rationality
for an ordinary person taking a benzodiazepine just because
they think they might be a little upset.
And yet that is the draw. That's what we see all the
time now on TV stations where they're now advertising
medicines. They're saying there's a pill for everything.
Of course, call your doctor, as though the doctor is going
to be standing by the telephone waiting for your call, right?
There's a pill for everything, and I think that — I
don't know what you can do about that with your committee
here, but that to me is a major concern, is the marketing
of drugs to the general public that are really initially developed
for and efficacious for a very small set of the public.
And it's the drugification of the society as a consequence
CHAIRMAN KASS: I have Mike Gazzaniga.
DR. GAZZANIGA:: I just wanted a point of
information from the two, Jim and Dan here. About a year
ago from Joseph LeDoux Lab there was this report of memory
erasure by bringing up an animal that's reminded of something
and then a protein synthesis inhibitor is injected, and then
the memory seems to be Ó-
CHAIRMAN KASS: Mike, could you speak up
DR. GAZZANIGA:: — the memory seems to
be erased, and this was an animal model repeat of the classic
ECT work that Jim was mentioning.
I know that talking to LeDoux that he was inundated with
phone calls from humans, not the rats he studied, who wanted
a pill because they wanted to get rid of certain memories
in their life.
Has that folded? Is there any biotech work on that? Is
that a viable concept still?
DR. SCHACTER: Well, they're continuing
with it in their lab, and I was actually at a conference a
few weeks ago, that Self Conference in New York. I was talking
to Karim Nader, who was one of the authors on that paper,
who claimed that there is some obscure clinical paper published
20 years ago where actually some form of this was done with
people. I haven't seen it yet. He gave me the —
DR. MCGAUGH: Larry Squire did that.
DR. SCHACTER: Larry Squire tried and couldn't
get the effect.
DR. MCGAUGH: Yeah, he did the —
DR. SCHACTER: With ECT.
DR. MCGAUGH: With ECT, did the right experiment.
DR. SCHACTER: Right.
DR. MCGAUGH: He gave subjects new material
to learn or material that they learned a long time ago or
material they learned, I think, just the day before, gave
them an ECT, and the only thing that they forgot was material
that they had just learned.
And so reactivating, that is, just doesn't do the job.
DR. SCHACTER: I guess the question would
be it's not clear whether it's this new research that
would be pointing toward a possible new avenue for this, but
given now that there's reawakened interest in this because
of what it might tell us about some of the basic, you know,
mechanisms of memory, might that point the way towards some
other way of approaching this than ECT?
You can comment on that.
DR. MCGAUGH: But even then it has constraints.
I was one of the first people to publish on this many years
ago, and it didn't work in 1970. We now have another
paper in which we did the same thing, but to the hippocampus
rather than the amygdala, no effect.
The effect is supposed to be that if you bring up information,
you retrieve information that you know pretty well, that makes
it susceptible to erasure by some treatment, and now I know
of several other studies that have completely failed to do
So I think the phenomenon itself is in serious question.
I wouldn't get too upset about its applicability because
in my view it doesn't exist, but we'll just have to
wait and see.
DR. KRAUTHAMMER: But what about application
to immediate erasure?
DR. MCGAUGH: To which?
DR. KRAUTHAMMER: Immediate erasure you
say happens —
DR. MCGAUGH: On immediate erasure.
DR. KRAUTHAMMER: Right.
DR. MCGAUGH: We've known about that
for 50 years.
DR. KRAUTHAMMER: Right. But the question
as I understood it was other than ECT, are there any other
agents that can do it?
DR. MCGAUGH: Sure.
DR. KRAUTHAMMER: Erasing?
DR. MCGAUGH: Sure. Protease synthesis
inhibitors can do that.
DR. KRAUTHAMMER: And these have been tested
DR. MCGAUGH: No, but scopolamine can do
it and atropine, and that's been tested in humans.
DR. KRAUTHAMMER: So you have an experience.
You get atropine. The memory is erased.
DR. MCGAUGH: It's not formed. It's
DR. KRAUTHAMMER: Right. I mean, right,
it never Ó-
DR. MCGAUGH: Yes.
DR. KRAUTHAMMER: — it never takes.
DR. SCHACTER: It's not fully formed
DR. MCGAUGH: Yeah.
DR. KRAUTHAMMER: And what's the lag
between the experience and the administration?
DR. MCGAUGH: It would have to be for something
like that probably minutes.
DR. KRAUTHAMMER: So you have a memory for
a minute or two, right? I mean, until it's —
DR. MCGAUGH: No, no. You can have that
memory for a longer period of time because that memory is
based on a different system than the memory that you use the
This is what he talked about in the first instance, the difference
between this immediate memory and the long term memory. It's
not the same brain mechanism at all. Underline that.
So what's happening with these treatments that are affecting
long term memory is that they're leaving short term memory
alone. You can have that memory for hours, let's say,
and then it will finally disappear, but what's happened
is you haven't made another stage of memory.
Maybe that's something to bear in mind, that not only
are we talking about information for different kinds of things
that are learned, like the motor skills, but we're talking
about information that's in a different temporal domain
having a different substrate.
DR. KRAUTHAMMER: I guess my question is:
if I could just follow up, I mean, is it conceivable that
you could market something that would tell people if you take
this immediately after a terrible experience, you'll wipe
it out and you won't suffer from it?
DR. MCGAUGH: Yes, that's conceivable.
That's more doable than the other side of it.
DR. KRAUTHAMMER: Right, and that would seem
to me to be a rather interesting question as to whether you'd
want to market a drug you carry around in your wallet to be
administered upon extreme shame.
DR. MCGAUGH: Well, let me give you an example
of that from the 1978 PSA crash in San Diego where they made
the horrible mistake of sending out desk people and baggage
handlers to clean up body parts after the crash, and then
there was a report in I believe it was the L.A. Times maybe
five years ago, a follow-up, and a very high percentage of
those people were never able to work again. They had been
permanently disabled because of the trauma.
Now, that's PTSD to the nth degree, and it probably for
them, I would make a guess, was worse than a soldier. These
things are happening all the time in the battlefield, but
imagine you as a baggage handler and for the first time you
have to pick up body parts and put them in bags.
Now, there would be a case in which something like that I
think would be of value. I don't think that you could
say, "Well, you know, it's really adaptive to be
able to know how to handle body parts. So it's a good
idea to keep that memory strong because you may have to do
that again and know how to cope with it."
I think that's a low probability. So that would be a
case in which — a clear-cut case — in which it would be
nice to say, "Take this. Weaken your memory of that."
DR. KRAUTHAMMER: I would agree with you,
and I would say that this is not just an odd event like a
crash of an airplane, but in Israel they're experiencing
that every week, and there's huge reports of post traumatic
stress syndrome. So in a society like that, you might want
people walking around with that in their wallets.
DR. MCGAUGH: Yeah. I just gave the other
one as a, you know, clear-cut stand-alone case.
DR. KRAUTHAMMER: Right.
DR. MCGAUGH: But the situation in Israel
and Palestine are just unbelievable. I mean it's just
hard to imagine living. I mean if we think it's bad enough
to worry about whether we should cross the street in Washington
or Maryland or Virginia, where we should be thinking the odds
are very high that we're going to get hit by a car, the
probabilities are just vastly different, the sniper versus
dying of food poisoning, for example.
CHAIRMAN KASS: Frank, we've got a few
people on the list, and then we are going to move to wind
Please, Frank, Gil, and the two Bills.
PROFESSOR FUKUYAMA: This is just a question
on a somewhat different subject, but both of you have talked
about fairly low level cognitive processes like, you know,
memorizing a seven digit string of numbers and so forth, but
there seems to be another much higher cognitive level that
involves memory that's also quite socially and politically
important, which is that, you know, you develop a paradigmatic
way of understanding the world at a certain point in your
life cycle, and usually that stops happening past — I don't
know — the age of 25, 30 or so, and once you've got that
paradigmatic way of looking at the world, you know, no amount
of disconfirmatory experience then can shake you from it,
which is why politics, you know, proceeds in generational
So if you lived through the Depression, you know, you think
that big government is the solution to, you know, out of control
markets, and you think that you've got to save, you know,
and saving is a great virtue because a rainy day may come
back, and so forth.
Whereas if you grew up, you know, in the '60s and '70s
or through the sexual revolution, I mean, you have just very
— you know, so you get these cohorts, age cohorts that have
basically imprinted memories on them, and basically until
they die they're never going to be shaken, you know, from
I mean, do we know anything about the physiology of this
kind of, you know, memory at this cognitive level? Because
it seems to me it's actually quite, you know, politically
relevant and important. But it seems to be at such a higher
level than the kinds of issues that you're dealing with.
DR. SCHACTER: I think within the realm
of memory discussions what you're referring to would probably
fit under the rubric of what's called semantic memory
at another subdivision, semantic and episodic. Semantic is
kind of general knowledge of the world. Episodic memory for
And you know, we do know something about semantic memory,
albeit studies in the laboratory context and, you know, in
a more modest way than full flown political beliefs and whatnot,
but we know something about the structure of semantic memory
and some very interesting work on how it may be organized
categorically. You know, you have patients who can lose access
to one category of knowledge and not another, you know, fruits
and vegetables versus living things versus tools, and there's
new imaging work on that.
So there's nothing that I'm aware of that directly
speaks to your point, but there's certainly a lot of cognitive
literature on what I would assume would be the basic building
block processes that would be relevant in your case.
DR. MCGAUGH: Well, I would add that the
assumption that is made in the neuroscience of memory is that
the building blocks may be the same for many kinds of memory,
but they build into different systems, and then they can build
into higher order systems, such as concepts.
There was a very famous book by Donald Hebb from McGill University,
Organization of Behavior, which you dealt with. How do you
get a concept? And his notion was you get a concept out of
individual small units which then interact, interface with
each other, and you get a concept because these things hook
Now, once those things are built up from a neurological point
of view, they're all the same. That is, this one is like
this one over there, but the content of it is different.
But one thing is sure. You're never going to build that
for somebody. That is, you're not going to put a gene
in and say, "I believe in globalism," or put a gene
in and something else. These are all slowly developed over
a long period of time.
PROFESSOR FUKUYAMA: But if I could just
say, I mean, but there does seem to be an age related ability
to put those concepts together and then to deconstruct them
and replace them with other concepts that would seem to have
to — I mean that just intuitively would have to have some
genetic basis because, you know, that ability to reconceptualize
things in very fundamental ways does seem to decline with
DR. MCGAUGH: Well, I'm not sure. I'm
not sure. Let's put in the motor skills. All right?
You can learn to ski at any age, at any age, and even if you
have been an ice skater, which is very different motor skills
involved, you can still learn to ski.
So you're not prevented from doing that. I would make
the counter argument that you get comfortable and then avoid
changes. It's not that it's genetically fixed in
that you're going to be stuck here for a while and then
I think that rather it's a consequence of what you're
doing, what you're surrounded by, what opportunities there
are that provide that. Because we even see people shifting
political parties and saying, "All the time I was wrong."
And we see Lula in Brazil having an epiphany about world economics
that he didn't have a few weeks before.
So the constraints that are put on one can cause one to see
the world in a different way. And then we use the same mechanism
to make a change.
Another thing that happens, I think, because I am growing
old, is that you can see the stages in your own life. I grew
up in the Depression. So I am one of the people that you
talked about, but you can then as you get more information,
you can put that phase of your life in a very different context.
So I'm not tied to what I learned and did during the
bad old Depression days, but I can use that information in
a flexible way to change my views about what our relations
are with Mexico, for example, and what our relation to Canada
and whether Schroeder is a nut or whether Schroeder is onto
Because we learn more. We incorporate more information,
and I think that's one of the great things about the human
brain is we're not fixed in developmental stages, but
rather we retain a large flexibility. That would be my take
It's just artificially that we get forced into it and
you see we respond to our own stereotypes, and this certainly
is the case in the political arena.
CHAIRMAN KASS: Gil.
PROFESSOR MEILAENDER: This is really just
a factual question though depending on the answer there might
be sort of more to be made of it, and it probably relates
Different people at least as they age, the effect of aging
on their memory is different, more pronounced in some than
in others, and so forth. What I want to know, if this isn't
a silly question, is do people in the very early years, the
early years of childhood, do some people just form less transient,
more vivid memories than others?
And I ask it because anecdotally, it seems to me to be the
case that they do. In other words, just among my friends
and acquaintances, some people remember a great deal more,
claim at least, seem to remember a great deal more about when
they were four years old than others do.
Are there differences at the beginning of life the way there
are at the end? And if so, do we know anything about why?
DR. SCHACTER: Yeah, well, presumably, I
mean, there are individual differences in memory though they're
quite poorly understood throughout, you know, life, and you
know, for the early stages I'm not aware of anything that
would really speak to an understanding of (a) the nature of
the difference and (b) whether one can really localize it,
as it were, to transience, to the actual forgetting over time
There might be many possible subprocesses that are responsible
for a difference between one individual and another in memory.
Some of those might have to do with basic processes operating
within the memory system or it just might be different contents
being operated on by those processes.
I suspect that a lot of time when people compare their memories,
and you have a good memory for this and I have a good memory
for that, that it's less about the basic process and more
about the particular contents of the memory.
But I'm not aware of anything on the very early stages
that would speak to that.
DR. MCGAUGH: But you also have to be concerned
about confounds in that. How much did the family talk about
it? How many photo albums are there around? How many times
were they reminded of it in various things? And these would
play into that.
PROFESSOR MEILAENDER: I even have in mind
rather pronounced differences among siblings in this matter.
I mean, I know of cases like that and so I was just curious.
DR. SCHACTER: I would suspect that post
event rehearsal is probably playing a bigger role there than
people realize, and you know, why it is that one person talks
about and thinks about and is reminded of, you know, an event
than another is, for example, within families where you'll
get siblings, you know, one claiming to have no memory for
the same event that the other remembers in detail is unclear,
but I would think of that, you know, in some sense as a confounder
and in come sense as part of what's interesting about
individual differences in memory, just the big role of post
CHAIRMAN KASS: I have two people left,
and I ask them to be brief, and we have as a procedural matter,
we have one person who wants to make public comment. So if
you'd be willing, we'll not take a break and go straight
to that and we should be done probably in ten to 15 minutes.
I have Bill Hurlbut and then Bill May. And then we'll
go to the public comment.
DR. KRAUTHAMMER: I just want to ask a question.
CHAIRMAN KASS: Why don't you wait at
the end then. Thank you.
Please, Bill Hurlbut.
DR. HURLBUT: Returning to the question
of forgetting, to Frank's question about the political
convictions and affiliations, there's an interesting relationship
between learning, memory, and our larger sense of place within
There was a flood in Leningrad after Pavlov had conditioned
his dogs, and the dogs were floated up and almost drowned,
and afterwards their learning, their conditioning had been
erased. Do you know about that?
DR. SCHACTER: No.
DR. HURLBUT: And this was picked up on
and used as some of the basis for the Soviet ideas on brainwashing,
and I think this is called transmarginal inhibition. Have
I got that right?
Anyway, here's the interesting —
DR. SCHACTER: Outside of my domain.
DR. HURLBUT: Here's the interesting
question. How do we and how does memory relate to the dissolution
and resolidification of the self in these very crucial domains
And this comes back to William May's comment that affection
grows as memory fades. What I'm really getting at —
and there's a question in here — it seems to me at least
looking at animal models that certain types of memory and
certain types of structuring of the self are tied in not just
with memory in the broad sense, but with very specific systems,
and here I'm thinking of things like oxytocin where the
memory of the sheep, for example, for its nursing offspring,
is with the second offspring completely erased. It doesn't
recognize or acknowledge at least its first offspring.
And there have been suggestions that maybe oxytocin plays
this same role in solidifying and bonding. It's a kind
of a memory conviction state.
Do you know anything about this? And what I'm getting
at here is are there any systems you know of that suggest
that there might be ways to specifically enhance certain kinds
of memories and, therefore, certain kinds of affectionate
structures or convictions of personal identity?
One of the suggestions, for example, has been that oxytocin
be used in marriage therapy to make the marriage bond stronger.
Do you see what I'm saying?
DR. SCHACTER: I think I see it. I don't
know anything relevant to it. Do you?
DR. MCGAUGH: The only thing that I know
is mother-child attachment is associated with the release
of oxytocin by the mother. That's all I know.
DR. HURLBUT: And this kind of bonds a memory
DR. MCGAUGH: It is proposed that it plays
a role in the bonding. What we have is a correlation, but
no one has done any critical experiments on that.
Now, oxytocin has been studied in Holland and DeWied's
laboratory as a memory enhancing drug, if you like, peptide
about 20 years ago, but it faded from fashion. So I don't
know of anyone who's working on it now.
CHAIRMAN KASS: Bill May.
DR. MAY: In thinking about the medical
— the warrants for use of medicine, I guess an old distinction
would be between existence and developed existence, being
and well-being, survival and flourishing. And in some of
your writings there's some suggestion, well, memory is
very important to survival. You know, we need it to survive,
and obviously medical warrants for doing something for those
who are impaired.
We get a little bit more nervous if what it seems to be
used for is economic survival, I mean, to help SAT scores
and so forth. On that level the medical warrants, other than
helping the impaired, the basic argument is no longer survival,
The problem of impoverishment and the Alzheimer's patient,
we've got various social ways of keeping them going and
so forth, but the human impoverishment is huge. It's
the flourishing of the human which is our concern there, and
we see objective warrants for trying to solve that problem
it seems to me.
Now, Frank is now gone, but he worried about was there any
application whatsoever at the level of politics or the level
of flourishing of the society at large, and there it seems
to me we may move beyond medicine and so forth because really
we happen into the arena where we are talking about ritual
retrieval of the past and the retrieval of decisive and defining
And that's a dimension of education which is very important,
but gets lost where education is justified simply in terms
of what it will give you by way of economic survival. The
problem of safeguarding the treasures from the past operates
at the level of politics and religion, and through education,
the transmission of tradition.
CHAIRMAN KASS: Thank you.
Charles, do you want a brief comment and then we will —
DR. KRAUTHAMMER: Well, before your question,
I would like to ask Dr. Schacter. I was intrigued by the
hints that you were giving about possibilities of using these
technologies to distinguish real memory from misattributed
memory, and I was just wondering since we are talking about
enhancement and thinking about whether we should worry about
enhancement in this area, if you could just speculate on what
the field would look like in 20 years and whether there's
anything about the direction of the field and the power it's
acquiring that ever troubles you.
DR. SCHACTER: It's an interesting question.
So far I would say I'm not too troubled mainly because
I'm all too familiar with the ins and outs of data and
how far away we are in this realm from really having anything
that works at an individual level.
So as a practical matter right now, I'm not particularly
troubled. You know, where it might be in 20 years it's
not clear. It's interesting to look back on, for example,
lie detection, again, as an example, something that is related
to this even though we're talking about people who were
doing their best to tell the truth in a memory situation.
I'm not a careful student of that literature, but it seems
to me that pretty much the issues have been the same at least
as practical legal issues, you know, for the last 30 years.
We may make faster progress in brain imaging that will bring
these things onto the front burner more quickly than they
have in lie detection, where you know I think if you turn
the clock back 20 or 30 years, I don't think you would
have — I don't think we've advanced all that much
in that amount of time.
Now, whether, you know, the same applies to this kind of
work is not clear to me. Most people, for example, our group
and others who are doing this are doing this not with any
pragmatic aim in mind. We're interested in basic memory
processes, and so you know, I think from the perspective of
the field, people who are working on this stuff in the field
aren't by and large doing it from a practical viewpoint.
Whether the advances, you know, in the name of basic science
in the next 20 years would be sufficient to where we have
to start worrying, I just don't know. It's just hard
CHAIRMAN KASS: Well, thank you both very
This is one of those occasions, I think Mike Gazzaniga commented
at the break, this is one of those occasions where in addition
to learning some interesting and wonderful things about memory,
memory research, that a certain kind of reassurance has been
provided with respect to at least the immediate concerns that
some people have, and it's good to be able to report things
like that, as well, and to a culture that's rather nervous
about just about everything that's coming.