WEBVTT

00:00:00.040 --> 00:00:02.460
The following content is
provided under a Creative

00:00:02.460 --> 00:00:03.870
Commons license.

00:00:03.870 --> 00:00:06.910
Your support will help MIT
OpenCourseWare continue to

00:00:06.910 --> 00:00:10.560
offer high quality educational
resources for free.

00:00:10.560 --> 00:00:13.460
To make a donation or view
additional materials from

00:00:13.460 --> 00:00:19.290
hundreds of MIT courses, visit
MIT OpenCourseWare at

00:00:19.290 --> 00:00:20.540
ocw.mit.edu.

00:00:23.210 --> 00:00:29.640
PROFESSOR: So when we talked
about learning, we said that

00:00:29.640 --> 00:00:33.280
everything we know comes from
learning except that which is

00:00:33.280 --> 00:00:35.410
given to us by our genes.

00:00:35.410 --> 00:00:38.550
Our knowledge of how to walk
around in the world, to read,

00:00:38.550 --> 00:00:41.530
to think, to talk, to feel, the
values that we have, the

00:00:41.530 --> 00:00:44.980
hopes we have, the fears we
have, they're all learned

00:00:44.980 --> 00:00:46.400
through experience.

00:00:46.400 --> 00:00:48.200
And then we talked about another
approach to thinking

00:00:48.200 --> 00:00:50.690
about how we learn things
through memory, and today I'm

00:00:50.690 --> 00:00:53.030
going to talk about what we
understand to be the brain

00:00:53.030 --> 00:00:55.380
organization of human memory.

00:00:55.380 --> 00:01:00.050
And Oliver Sacks, in The Lost
Mariner, chapter 2, talks

00:01:00.050 --> 00:01:04.480
about one of the two major kinds
of amnesia, the loss of

00:01:04.480 --> 00:01:07.050
memory due to a brain disorder
or a brain injury.

00:01:07.050 --> 00:01:09.960
But he has a very nice language
from Luis Bunuel

00:01:09.960 --> 00:01:12.100
talking about memory. "You
have begun to lose your

00:01:12.100 --> 00:01:14.780
memory, if only in bits and
pieces, to realize that memory

00:01:14.780 --> 00:01:16.450
is what makes our life.

00:01:16.450 --> 00:01:18.570
Life without memory
is no life at all.

00:01:18.570 --> 00:01:21.110
Our memories, our coherence, our
reason, our feeling, even

00:01:21.110 --> 00:01:21.740
our action.

00:01:21.740 --> 00:01:23.290
Without it, we are nothing.

00:01:23.290 --> 00:01:25.870
I can only wait for the final
amnesia, the one that erases

00:01:25.870 --> 00:01:26.890
an entire life as it did

00:01:26.890 --> 00:01:28.120
my mother's." Right?

00:01:28.120 --> 00:01:31.360
So our memories of our lives,
what we've done, where we've

00:01:31.360 --> 00:01:33.760
been, what mattered to us--

00:01:33.760 --> 00:01:36.370
that's a huge piece
of who we are.

00:01:36.370 --> 00:01:41.290
And he runs into a patient who
came into the hospital early

00:01:41.290 --> 00:01:45.280
in 1975 with a note saying
"Helpless, demented, confused,

00:01:45.280 --> 00:01:48.580
and disoriented." He
had what we call

00:01:48.580 --> 00:01:51.610
clinically Korsakoff's Syndrome.

00:01:51.610 --> 00:01:55.500
So these are individuals who
have severe alcoholism for

00:01:55.500 --> 00:01:59.170
many years, and a subset of
those people develop a severe

00:01:59.170 --> 00:02:02.260
memory disorder,
like this man.

00:02:02.260 --> 00:02:05.630
He's 49 years old.

00:02:05.630 --> 00:02:08.970
He tells Sacks about his life,
details of his life, going

00:02:08.970 --> 00:02:10.630
into World War II, that
he worked on a

00:02:10.630 --> 00:02:12.500
submarine, and so on.

00:02:12.500 --> 00:02:14.870
And Sacks says, "a full and
interesting life remembered

00:02:14.870 --> 00:02:18.650
visually in detail, but for some
reason his memories stop

00:02:18.650 --> 00:02:19.850
around World War II."

00:02:19.850 --> 00:02:21.270
Then he asked him,
what year is it?

00:02:21.270 --> 00:02:23.250
Remember the year is 1975.

00:02:23.250 --> 00:02:24.950
And he says, it's 1945.

00:02:24.950 --> 00:02:25.810
What do you mean?

00:02:25.810 --> 00:02:27.960
We've won the war, FDR--

00:02:27.960 --> 00:02:30.060
Franklin Delano Roosevelt-- is
dead, Truman's at the helm.

00:02:30.060 --> 00:02:31.700
These are great times ahead.

00:02:31.700 --> 00:02:33.880
And Jimmy, how old
would you be?

00:02:33.880 --> 00:02:35.820
He stopped and he said,
well, I guess I'm 19.

00:02:35.820 --> 00:02:39.400
That would have been his age,
you know, 20 years ago, right?

00:02:39.400 --> 00:02:40.620
30 years ago.

00:02:40.620 --> 00:02:42.920
So he's stuck in time.

00:02:42.920 --> 00:02:44.860
He thinks it's historically
a while ago.

00:02:44.860 --> 00:02:46.390
He has the completely
wrong age.

00:02:46.390 --> 00:02:48.580
And Sacks takes out a mirror and
throws it at him and says,

00:02:48.580 --> 00:02:49.640
what do you--

00:02:49.640 --> 00:02:50.440
what happened?

00:02:50.440 --> 00:02:52.220
And he says, Jesus Christ,
what's going on?

00:02:52.220 --> 00:02:52.700
What happened?

00:02:52.700 --> 00:02:53.200
Is this a nightmare?

00:02:53.200 --> 00:02:53.730
Am I crazy?

00:02:53.730 --> 00:02:54.680
Is this a joke?

00:02:54.680 --> 00:02:58.560
So it's as if he's lost the
last 30 years of his life.

00:02:58.560 --> 00:03:00.640
And he's living as if it
were 30 years ago.

00:03:04.230 --> 00:03:06.680
And then when he walks back in,
he's hardly recognized.

00:03:06.680 --> 00:03:08.770
Sacks is hardly recognized
by the person.

00:03:08.770 --> 00:03:12.830
On intelligence, he shows
excellent ability, but his

00:03:12.830 --> 00:03:15.710
memory-- very quickly,
he forgets things.

00:03:15.710 --> 00:03:17.560
What is this, I asked him,
showing him a photo in a

00:03:17.560 --> 00:03:18.450
magazine I was holding.

00:03:18.450 --> 00:03:19.590
It's the moon.

00:03:19.590 --> 00:03:21.240
No it's not, if it's a
picture of the Earth

00:03:21.240 --> 00:03:22.530
taken from the moon.

00:03:22.530 --> 00:03:26.110
I'm mean, don't forget, in 1945,
nobody imagined that we

00:03:26.110 --> 00:03:28.250
would be sending up people
to the moon, right?

00:03:28.250 --> 00:03:30.330
And taking pictures from
the moon of the Earth.

00:03:30.330 --> 00:03:30.970
You're used to it.

00:03:30.970 --> 00:03:33.630
It's nothing to you.

00:03:33.630 --> 00:03:34.150
No it's not.

00:03:34.150 --> 00:03:35.140
It's a picture of the Earth
taken from the moon.

00:03:35.140 --> 00:03:35.760
Doc, you're kidding.

00:03:35.760 --> 00:03:37.510
Somebody would have to get
a camera up there.

00:03:37.510 --> 00:03:38.000
Naturally.

00:03:38.000 --> 00:03:38.730
Hell, you're joking.

00:03:38.730 --> 00:03:40.940
How would you know that?

00:03:40.940 --> 00:03:46.920
So he has lost the last 30 years
of his life, for all

00:03:46.920 --> 00:03:48.650
practical purposes,
despite normal

00:03:48.650 --> 00:03:51.680
intelligence and normal thought.

00:03:51.680 --> 00:03:55.000
So we're going to talk more
about how this occurs in

00:03:55.000 --> 00:03:58.180
another patient, and more
generally, how the brain

00:03:58.180 --> 00:04:00.820
supports your ability
to learn things.

00:04:00.820 --> 00:04:02.730
So here's this remarkable
brain.

00:04:02.730 --> 00:04:06.520
And in one sense, you could say
it's almost all there to

00:04:06.520 --> 00:04:09.620
absorb things in life.

00:04:09.620 --> 00:04:12.910
You have very small insects
that can do a lot with the

00:04:12.910 --> 00:04:13.830
genes they're given.

00:04:13.830 --> 00:04:17.700
We have big brains to learn
the life we lead.

00:04:17.700 --> 00:04:20.320
And I'm going to come back at
the end to this metaphor, but

00:04:20.320 --> 00:04:23.740
the idea that in many ways, your
brain as you sit there

00:04:23.740 --> 00:04:27.150
right now, or any healthy
brain, is kind of like a

00:04:27.150 --> 00:04:31.700
symphony orchestra that has many
specialized instruments

00:04:31.700 --> 00:04:33.850
for learning different
kinds of things.

00:04:33.850 --> 00:04:36.650
And you feel like a unified
whole, and you are, and these

00:04:36.650 --> 00:04:38.270
instruments interact
with each other.

00:04:38.270 --> 00:04:40.980
But just like a symphony
orchestra, you know, together

00:04:40.980 --> 00:04:42.340
they make a sound.

00:04:42.340 --> 00:04:45.340
So imagine you're a Martian who
lands outside a symphony

00:04:45.340 --> 00:04:48.950
hall, and you hear a
symphony playing.

00:04:48.950 --> 00:04:49.790
It's something beautiful.

00:04:49.790 --> 00:04:51.420
It sounds really interesting.

00:04:51.420 --> 00:04:54.360
If you're a Martian on the
outside of the symphony hall,

00:04:54.360 --> 00:04:56.810
how many instruments are you
guessing, how many things

00:04:56.810 --> 00:04:58.870
might you guess, are
making that sound?

00:04:58.870 --> 00:05:00.860
Would you have any idea?

00:05:00.860 --> 00:05:02.430
One big one?

00:05:02.430 --> 00:05:04.220
Two big ones?

00:05:04.220 --> 00:05:05.490
50?

00:05:05.490 --> 00:05:07.000
It'd be very hard to guess.

00:05:07.000 --> 00:05:09.710
And it's only been in the last
20 years or so that we had

00:05:09.710 --> 00:05:12.240
some real clarification
on the instruments of

00:05:12.240 --> 00:05:14.380
learning in our brains.

00:05:14.380 --> 00:05:16.260
So I'm going to talk about
three things today.

00:05:16.260 --> 00:05:19.880
Anterograde amnesia, the loss
of the ability to form new

00:05:19.880 --> 00:05:23.540
memories; retrograde amnesia,
the loss of memories you

00:05:23.540 --> 00:05:27.610
already had; and that idea of
memory systems, this idea that

00:05:27.610 --> 00:05:30.650
each of us has a symphony of
learning instruments in our

00:05:30.650 --> 00:05:31.990
brain that empowers us to learn

00:05:31.990 --> 00:05:34.610
different kinds of things.

00:05:34.610 --> 00:05:36.650
So anterograde amnesia is,
again, the inability to

00:05:36.650 --> 00:05:38.550
remember new information.

00:05:38.550 --> 00:05:40.830
And people mostly talk about
two kinds, but as you see,

00:05:40.830 --> 00:05:42.670
they're most of life, OK?

00:05:42.670 --> 00:05:45.220
Not all of that-- we'll talk
about that-- but a lot of it,

00:05:45.220 --> 00:05:48.190
such as events you experience
or facts you encounter.

00:05:48.190 --> 00:05:49.400
So think about this.

00:05:49.400 --> 00:05:51.380
All the events you've remembered
from your entire

00:05:51.380 --> 00:05:54.840
life, all the facts you ever
learned in school, at home, on

00:05:54.840 --> 00:05:55.860
the computer--

00:05:55.860 --> 00:05:58.170
that's a huge amount
of what you know.

00:05:58.170 --> 00:06:00.350
And a patient like we just
described or the patients I'll

00:06:00.350 --> 00:06:04.090
describe in a moment lose the
ability to remember any event

00:06:04.090 --> 00:06:07.180
from the day they become amnesic
or remember much in

00:06:07.180 --> 00:06:09.990
the way of new information
of any kind, in terms of

00:06:09.990 --> 00:06:11.570
facts of the world.

00:06:11.570 --> 00:06:13.740
So in what sense is memory
in the brain

00:06:13.740 --> 00:06:14.830
distributed or localized?

00:06:14.830 --> 00:06:17.760
In what sense is it in one
place or many places?

00:06:17.760 --> 00:06:20.430
And this has been a central
debate in all of neuroscience,

00:06:20.430 --> 00:06:21.430
but in memory, too.

00:06:21.430 --> 00:06:24.680
And Karl Lashley, a giant of
neuroscience, who worked just

00:06:24.680 --> 00:06:28.570
two T stops away at Harvard,
said this: "It is not possible

00:06:28.570 --> 00:06:31.540
to demonstrate the isolated
localization of a memory trace

00:06:31.540 --> 00:06:32.810
anywhere in the nervous
system.

00:06:32.810 --> 00:06:35.730
The engram is represented
throughout the brain." The

00:06:35.730 --> 00:06:38.700
engram's kind of a magic word
in memory and the brain.

00:06:38.700 --> 00:06:41.850
The engram is the thing in you
that changes in your brain

00:06:41.850 --> 00:06:44.110
that is the stuff of a memory.

00:06:44.110 --> 00:06:48.250
If you learn today that there's
more terrible news in

00:06:48.250 --> 00:06:51.770
Japan with the nuclear reactors,
and you remember

00:06:51.770 --> 00:06:53.740
that because it matters to you,
because you care about

00:06:53.740 --> 00:06:58.670
the people, there's a physical
change in your brain that is

00:06:58.670 --> 00:06:59.230
the memory.

00:06:59.230 --> 00:07:01.400
If you remember what I'm saying
now for more than a few

00:07:01.400 --> 00:07:04.070
moments, that is a physical
change in your brain.

00:07:04.070 --> 00:07:06.430
So all of us in neuroscience
are amazed by that.

00:07:06.430 --> 00:07:10.240
How can we physically change the
brain to be the biological

00:07:10.240 --> 00:07:12.230
record of a memory,
that we can keep

00:07:12.230 --> 00:07:13.730
and use in the future?

00:07:13.730 --> 00:07:16.350
And the engram, Karl Lashley
said, is represented

00:07:16.350 --> 00:07:17.550
throughout the brain.

00:07:17.550 --> 00:07:19.150
It's spread throughout
the brain.

00:07:19.150 --> 00:07:20.830
Here's why he thought that.

00:07:20.830 --> 00:07:23.740
So he was working with rats,
and they ran a maze.

00:07:23.740 --> 00:07:26.340
And you saw how quickly you
learned the maze to go from

00:07:26.340 --> 00:07:30.240
start to finish to get the
food if you're a rat.

00:07:30.240 --> 00:07:33.590
And here's what he found, and
the results won't amaze you,

00:07:33.590 --> 00:07:34.890
to start with.

00:07:34.890 --> 00:07:36.150
So what he did is he made--

00:07:36.150 --> 00:07:38.550
this is number of errors
during learning, so

00:07:38.550 --> 00:07:39.710
it's good to be low.

00:07:39.710 --> 00:07:42.480
And this is how much neocortex
they took out, ranging from

00:07:42.480 --> 00:07:44.860
10% to 80%.

00:07:44.860 --> 00:07:47.790
And he discovered that the more
neocortex you took out,

00:07:47.790 --> 00:07:49.130
the worse learner you were.

00:07:49.130 --> 00:07:52.140
Now if it didn't happen in some
broad sense, we'd say,

00:07:52.140 --> 00:07:54.150
well, what part of us
is learning it?

00:07:54.150 --> 00:07:56.030
You take out more brain, you're
less of a good reader.

00:07:56.030 --> 00:07:58.210
10% out, 10% percent
bad memory.

00:07:58.210 --> 00:08:01.360
50% of brain out, 50%
of bad memory.

00:08:01.360 --> 00:08:04.030
80% of the brain out, 80%
bad memory, right?

00:08:04.030 --> 00:08:05.260
OK, that makes sense.

00:08:05.260 --> 00:08:08.040
But what he was really impressed
by is it didn't seem

00:08:08.040 --> 00:08:12.360
to him to matter where he took
the brain out of the rat.

00:08:12.360 --> 00:08:13.360
It just didn't matter.

00:08:13.360 --> 00:08:15.450
It's just the percent
that he took out.

00:08:15.450 --> 00:08:17.480
Whatever percent he took
out is how bad

00:08:17.480 --> 00:08:18.730
the learner you were.

00:08:18.730 --> 00:08:20.415
And we'll come back to that, but
that's why I said, it must

00:08:20.415 --> 00:08:22.880
be spread everywhere, because
wherever I take out brain,

00:08:22.880 --> 00:08:26.430
part of the memory is weakened
or disappears.

00:08:26.430 --> 00:08:29.210
And sometimes people call that
mass action for distributed

00:08:29.210 --> 00:08:30.950
memory, that's spread
throughout the brain

00:08:30.950 --> 00:08:32.000
in this mass way.

00:08:32.000 --> 00:08:35.280
It's just spread all
over the place.

00:08:35.280 --> 00:08:36.760
And a single clinical case--

00:08:36.760 --> 00:08:38.990
I'll spend a few moments
on him for two reasons.

00:08:38.990 --> 00:08:42.320
It is the historical case that
turned this around, and I have

00:08:42.320 --> 00:08:44.059
a particular perspective on
this because when I was a

00:08:44.059 --> 00:08:46.680
graduate student, I did a lot
of work with this single

00:08:46.680 --> 00:08:49.560
patient whose case turned around
our understanding of

00:08:49.560 --> 00:08:51.350
the mammalian brain
and memory.

00:08:51.350 --> 00:08:53.250
So as I start to
talk about it--

00:08:53.250 --> 00:08:55.740
I can't tell you what Phineas
Gage was really like, beyond

00:08:55.740 --> 00:08:56.580
what I read, OK?

00:08:56.580 --> 00:08:58.580
Or Broca's patients
from the 1900s.

00:08:58.580 --> 00:09:00.860
I can tell you what HM, the man
we're about to talk about,

00:09:00.860 --> 00:09:02.980
was like because I spent many,
many hours with him, two

00:09:02.980 --> 00:09:05.530
blocks from here, when I was a
graduate student in Suzanne

00:09:05.530 --> 00:09:06.790
Corkin's lab.

00:09:06.790 --> 00:09:09.800
So the structure we'll focus
on is something called the

00:09:09.800 --> 00:09:10.530
hippocampus.

00:09:10.530 --> 00:09:12.060
It looks sort of big
in the picture.

00:09:12.060 --> 00:09:15.300
It's about the size of about 2/3
of your thumb, one on the

00:09:15.300 --> 00:09:19.020
left and one on the right in the
insides of the medial part

00:09:19.020 --> 00:09:19.880
of the temporal lobes.

00:09:19.880 --> 00:09:23.460
A pretty small percentage
of the brain.

00:09:23.460 --> 00:09:25.910
And you can see it's
located here.

00:09:25.910 --> 00:09:28.380
Here it is sort of drawn out.

00:09:28.380 --> 00:09:30.100
And here's the story
of this man.

00:09:30.100 --> 00:09:31.760
He died a couple years ago.

00:09:31.760 --> 00:09:35.640
There's lots of plays and movies
and books in the works.

00:09:35.640 --> 00:09:37.870
I can also tell you that when I
was a graduate student here,

00:09:37.870 --> 00:09:40.390
my fellow graduate students
would say, tell us who HM,

00:09:40.390 --> 00:09:43.320
what his initials really stand
for, as if it really mattered,

00:09:43.320 --> 00:09:44.570
particularly.

00:09:44.570 --> 00:09:47.270
But I felt privileged to have
the secret knowledge of what

00:09:47.270 --> 00:09:47.710
his name was.

00:09:47.710 --> 00:09:48.980
When he died, they
published it.

00:09:48.980 --> 00:09:52.190
Everybody knows now, so
my secret is out.

00:09:52.190 --> 00:09:53.450
Or our secret is out.

00:09:53.450 --> 00:09:54.650
He was born in 1926.

00:09:54.650 --> 00:09:58.520
At age 16 he had his first
major epileptic seizure.

00:09:58.520 --> 00:10:03.140
By age 27, he was having many,
many seizures a day.

00:10:03.140 --> 00:10:05.060
You can have petit mal seizures,
where you tune out

00:10:05.060 --> 00:10:06.030
from the world.

00:10:06.030 --> 00:10:07.990
You can have grand mal seizures
that lead to physical

00:10:07.990 --> 00:10:09.240
convulsions.

00:10:09.240 --> 00:10:12.830
By age 27, the medications were
not helping him, in 1953,

00:10:12.830 --> 00:10:14.940
very much, that were
available.

00:10:14.940 --> 00:10:17.280
And he was just sitting in his
house just waiting for the

00:10:17.280 --> 00:10:18.210
next seizure, basically.

00:10:18.210 --> 00:10:19.700
That was his life.

00:10:19.700 --> 00:10:22.650
He was a man of normal
intelligence and fairly

00:10:22.650 --> 00:10:23.980
typical, average background.

00:10:23.980 --> 00:10:25.920
Nothing remarkable.

00:10:25.920 --> 00:10:28.710
And in order to treat the
epilepsy, William Scoville,

00:10:28.710 --> 00:10:32.980
the surgeon, resected, or took
out surgically, the tissue.

00:10:32.980 --> 00:10:36.960
Here's this cartoon of
what he took out--

00:10:36.960 --> 00:10:39.160
the hippocampus, also the
amygdala just in front of it,

00:10:39.160 --> 00:10:41.010
also tissue that surrounds it.

00:10:41.010 --> 00:10:43.860
And the reason he took it out
is the single most likely

00:10:43.860 --> 00:10:48.360
place where epileptic seizures
start is the hippocampus.

00:10:48.360 --> 00:10:51.660
It's kind of a signal, in many
ways, that the hippocampus

00:10:51.660 --> 00:10:53.570
does something very dangerous.

00:10:53.570 --> 00:10:56.530
It's operating in some sort
of fast speed lane of

00:10:56.530 --> 00:10:57.550
neurobiology.

00:10:57.550 --> 00:11:01.310
It's a very sensitive part of
the brain, for lack of oxygen,

00:11:01.310 --> 00:11:02.850
for diseases.

00:11:02.850 --> 00:11:04.070
We'll talk about Alzheimer's
today.

00:11:04.070 --> 00:11:06.840
It's a very vulnerable part
of the brain, but it does

00:11:06.840 --> 00:11:10.490
something magical and essential
for memory.

00:11:10.490 --> 00:11:12.690
And so they couldn't locate
where the seizures were

00:11:12.690 --> 00:11:15.110
starting, very well, from him in
ways they typically do, so

00:11:15.110 --> 00:11:15.720
they said, OK.

00:11:15.720 --> 00:11:16.750
He has all these seizures.

00:11:16.750 --> 00:11:18.040
We don't know what to do.

00:11:18.040 --> 00:11:20.480
Let's take out the left and the
right hippocampus and the

00:11:20.480 --> 00:11:21.890
surrounding tissue.

00:11:21.890 --> 00:11:23.970
And in terms of seizure
control, it was pretty

00:11:23.970 --> 00:11:25.040
successful.

00:11:25.040 --> 00:11:29.040
He had to stay on medications
from that day 'til a few years

00:11:29.040 --> 00:11:30.960
ago, when he passed away.

00:11:30.960 --> 00:11:33.480
But he had very, very
few seizures

00:11:33.480 --> 00:11:34.350
compared to before that.

00:11:34.350 --> 00:11:35.800
So it was medically
very successful in

00:11:35.800 --> 00:11:37.500
terms of seizure control.

00:11:37.500 --> 00:11:39.490
It had a giant side effect.

00:11:39.490 --> 00:11:45.150
From that day 'til a couple
years ago, HM never formed a

00:11:45.150 --> 00:11:48.320
new memory for any event, or
for all practical purposes,

00:11:48.320 --> 00:11:52.210
for any fact that he was exposed
to ever again, despite

00:11:52.210 --> 00:11:56.350
his normal intelligence and his

00:11:56.350 --> 00:11:58.050
otherwise intact abilities.

00:11:58.050 --> 00:11:59.720
So they took an MR of HM.

00:11:59.720 --> 00:12:02.530
It took a long time to decide
they could even take a MR

00:12:02.530 --> 00:12:06.770
picture of HM, because they were
very worried that if you

00:12:06.770 --> 00:12:10.010
put him in a MR scanner, there's
magnetic fields, and

00:12:10.010 --> 00:12:12.050
often they put in clips
after surgery.

00:12:12.050 --> 00:12:14.490
Regularly they put in clips
after surgery, to keep the

00:12:14.490 --> 00:12:16.940
vessels from bleeding.

00:12:16.940 --> 00:12:18.570
And they weren't sure
what the clips were

00:12:18.570 --> 00:12:19.980
made of for a while.

00:12:19.980 --> 00:12:21.630
And so they said, well, we can't
put him in, because we

00:12:21.630 --> 00:12:26.650
could move the clips and kill
him, which would be bad.

00:12:26.650 --> 00:12:27.520
So they thought it
was dangerous.

00:12:27.520 --> 00:12:30.340
And they did a huge detective
process of going to factory

00:12:30.340 --> 00:12:32.490
that made them and talking to
people who built it before

00:12:32.490 --> 00:12:34.980
they decided-- it took, like, 10
years of research to decide

00:12:34.980 --> 00:12:38.700
that the clips were not metal
and they could take an MR.

00:12:38.700 --> 00:12:41.010
And even here-- here's the top
of the brain, normal person.

00:12:41.010 --> 00:12:43.070
Here's the hippocampus, the
small structure in you--

00:12:43.070 --> 00:12:44.320
pretty small.

00:12:44.320 --> 00:12:46.390
It's removed in him, but you
can see he's got a lot of

00:12:46.390 --> 00:12:51.240
brain left, except for these
removed structures.

00:12:51.240 --> 00:12:55.860
But what happened was he had
such a fantastic impairment of

00:12:55.860 --> 00:12:57.180
learning new information,
it got the

00:12:57.180 --> 00:12:58.780
label of a global amnesia.

00:12:58.780 --> 00:12:59.970
Global amnesia.

00:12:59.970 --> 00:13:02.750
It didn't matter whether tests
were hard, like remembering

00:13:02.750 --> 00:13:04.610
stuff or multiple choice.

00:13:04.610 --> 00:13:07.050
He failed practically every
standard test of memory you'd

00:13:07.050 --> 00:13:08.070
ever give a person.

00:13:08.070 --> 00:13:11.120
Or if it's words or nonsense
syllables, faces, clicks,

00:13:11.120 --> 00:13:14.800
mazes, public events,
assassinations of presidents,

00:13:14.800 --> 00:13:18.660
wars, going to the
moon, whatever

00:13:18.660 --> 00:13:20.490
you think of as famous--

00:13:20.490 --> 00:13:22.420
he didn't know they
happened at all.

00:13:22.420 --> 00:13:23.430
Personal events.

00:13:23.430 --> 00:13:25.300
The death of his own parents.

00:13:25.300 --> 00:13:27.720
If you asked him how his parents
were doing, he would

00:13:27.720 --> 00:13:29.830
say something like, I haven't
seen them for a while.

00:13:29.830 --> 00:13:31.860
He would not know they
had passed away.

00:13:31.860 --> 00:13:34.710
So no matter how publicly
famous something was, no

00:13:34.710 --> 00:13:36.860
matter how personally important
it was, that

00:13:36.860 --> 00:13:37.610
knowledge--

00:13:37.610 --> 00:13:39.860
he could remember it for a few
seconds, and we'll talk about

00:13:39.860 --> 00:13:41.480
that more, but it
would drift away

00:13:41.480 --> 00:13:43.030
completely within seconds.

00:13:43.030 --> 00:13:45.930
No matter how often he heard it,
no matter how personally

00:13:45.930 --> 00:13:48.380
important it was.

00:13:48.380 --> 00:13:51.360
And you could say, what does he
say his experience is like?

00:13:51.360 --> 00:13:53.520
So Brenda Milner, the
neuropsychologist, who did a

00:13:53.520 --> 00:13:55.990
lot of the critical original
research with him, describing

00:13:55.990 --> 00:13:59.220
his case said that he's said,
"Every day is alone in itself,

00:13:59.220 --> 00:14:01.460
whatever joy I've had and
whatever sorrow I've had." I

00:14:01.460 --> 00:14:05.040
mean, it sounds very poetic, and
it's true, except that for

00:14:05.040 --> 00:14:08.200
HM, his memory lasts seconds.

00:14:08.200 --> 00:14:11.730
So he doesn't remember any day
since 1950 or something.

00:14:11.730 --> 00:14:14.220
So I think his experience is
more like this. "Right now I'm

00:14:14.220 --> 00:14:17.310
wondering, Have I done or
said anything amiss?

00:14:17.310 --> 00:14:19.700
You see, at this moment,
everything looks clear to me."

00:14:19.700 --> 00:14:21.050
His mind is clear.

00:14:21.050 --> 00:14:22.070
He's smart.

00:14:22.070 --> 00:14:24.720
"But what happened just before?"
What happened a

00:14:24.720 --> 00:14:27.160
minute ago, or two minutes
ago, or five minutes ago?

00:14:27.160 --> 00:14:28.270
"That's what worries me.

00:14:28.270 --> 00:14:30.740
It's like waking from a
dream; I just don't

00:14:30.740 --> 00:14:32.280
remember." All right?

00:14:32.280 --> 00:14:35.160
So when we worked with him here,
a few blocks from here,

00:14:35.160 --> 00:14:36.640
he wouldn't know who I was.

00:14:36.640 --> 00:14:37.960
He wouldn't know why
he was here.

00:14:37.960 --> 00:14:40.670
He was an extremely pleasant
and wonderful research

00:14:40.670 --> 00:14:42.670
participant.

00:14:42.670 --> 00:14:44.980
A piece of that was-- he was
very nice and very pleasant.

00:14:44.980 --> 00:14:47.720
A piece of that was he didn't
know time was passing.

00:14:47.720 --> 00:14:49.840
So if any of you are involved
in experiments as

00:14:49.840 --> 00:14:53.690
participants, sometimes people
enjoy the experiment for about

00:14:53.690 --> 00:14:56.150
five minutes but not for the
next 55, because they're

00:14:56.150 --> 00:14:58.390
getting a lot of measurements
to get a good measurement.

00:14:58.390 --> 00:15:01.540
For HM, it was never an
hour-long experiment.

00:15:01.540 --> 00:15:03.840
It was always a minute and
a half experiment.

00:15:03.840 --> 00:15:08.470
So you had to be very careful
not to abuse that situation as

00:15:08.470 --> 00:15:10.160
a researcher.

00:15:10.160 --> 00:15:12.030
So here's his name,
Henry Molaison.

00:15:12.030 --> 00:15:15.660
Here's his pictures from his
youth and from this.

00:15:15.660 --> 00:15:18.100
But you see how he had
a complete reversal?

00:15:18.100 --> 00:15:21.600
All of a sudden, the removal
just two small regions-- the

00:15:21.600 --> 00:15:24.510
left and the right hippocampus
in the brain--

00:15:24.510 --> 00:15:27.400
stopped the creation, of
formation of new memories

00:15:27.400 --> 00:15:30.230
altogether and in any
practical way.

00:15:30.230 --> 00:15:33.440
So that shows you that memory's
not spread throughout

00:15:33.440 --> 00:15:36.280
the brain in a sort of
undifferentiated mass-action

00:15:36.280 --> 00:15:40.120
way, but that particular parts
of the brain play very

00:15:40.120 --> 00:15:41.800
particular roles in how
memories are formed.

00:15:41.800 --> 00:15:44.210
And the hippocampus plays
a huge role in that.

00:15:44.210 --> 00:15:44.700
Can't see it.

00:15:44.700 --> 00:15:48.900
So they had a big project where
they brought him from

00:15:48.900 --> 00:15:50.700
the nursing home in Connecticut
where he was.

00:15:50.700 --> 00:15:52.630
They drove him by ambulance to

00:15:52.630 --> 00:15:54.470
Massachusetts General Hospital.

00:15:54.470 --> 00:15:58.670
They overnight did a something
like an eight hour MRI scan on

00:15:58.670 --> 00:16:00.770
him, to have the world's most
detailed MRI scan you

00:16:00.770 --> 00:16:02.650
practically have on anybody.

00:16:02.650 --> 00:16:04.760
And then they removed his
brain and sent it to a

00:16:04.760 --> 00:16:09.090
laboratory in San Diego, where
they took his brain and they

00:16:09.090 --> 00:16:12.895
sliced it into many, many, many
incredibly thin slices

00:16:12.895 --> 00:16:14.850
that they're going to
post on the web.

00:16:14.850 --> 00:16:17.790
And on the web for quite
a while was the

00:16:17.790 --> 00:16:19.470
slicing of his brain.

00:16:19.470 --> 00:16:24.530
He consented to this, all this
research effort, and finally

00:16:24.530 --> 00:16:27.850
his brain is a gift to research
and understanding the

00:16:27.850 --> 00:16:29.100
brain and memory.

00:16:39.220 --> 00:16:41.800
So let me tell you a couple of
experimental results from him,

00:16:41.800 --> 00:16:44.270
and then you might think of
questions you have about what

00:16:44.270 --> 00:16:46.290
a man like this is like, OK?

00:16:46.290 --> 00:16:49.370
So we talked about last time
about that if you give a list

00:16:49.370 --> 00:16:51.830
of words to people, and you try
to do it yourself, that

00:16:51.830 --> 00:16:53.870
people remember the
first words best--

00:16:53.870 --> 00:16:56.290
the primacy effect, a long-term
memory effect that's

00:16:56.290 --> 00:16:59.150
not affected by adding a
few seconds of delay--

00:16:59.150 --> 00:17:00.320
and a recency affect--

00:17:00.320 --> 00:17:02.650
superior memory for the last few
words of the list that's

00:17:02.650 --> 00:17:06.640
strongly affected by adding in
just 10 seconds of delay from

00:17:06.640 --> 00:17:08.900
the last word until
you recall.

00:17:08.900 --> 00:17:12.310
And this is done with other
amnesic patients, and not HM,

00:17:12.310 --> 00:17:13.790
but a similar disorder.

00:17:13.790 --> 00:17:15.710
You can see that their memory
is worse overall--

00:17:15.710 --> 00:17:16.829
there, the dark line--

00:17:16.829 --> 00:17:19.220
but they have some
benefit from the

00:17:19.220 --> 00:17:21.680
beginning of the list and--

00:17:21.680 --> 00:17:22.295
but it's small.

00:17:22.295 --> 00:17:24.040
It's not the same
as the control--

00:17:24.040 --> 00:17:27.040
but they're completely normal,
completely normal for the last

00:17:27.040 --> 00:17:28.450
two or three words.

00:17:28.450 --> 00:17:31.430
So this is what we mean by
impaired long-term memory and

00:17:31.430 --> 00:17:33.110
spared short-term memory.

00:17:33.110 --> 00:17:34.830
But the short-term memory's
just moments.

00:17:34.830 --> 00:17:37.500
But it's not just an impression
you have.

00:17:37.500 --> 00:17:41.450
It's a scientific experiment
you can do.

00:17:41.450 --> 00:17:42.170
Here's another--

00:17:42.170 --> 00:17:43.110
we did this last time.

00:17:43.110 --> 00:17:47.030
We said, if you read digits to
people and ask them to repeat

00:17:47.030 --> 00:17:51.350
them back aloud, that we have a
limited short-term memory of

00:17:51.350 --> 00:17:52.570
seven plus or minus two.

00:17:52.570 --> 00:17:56.860
People get to about seven
or eight, typically.

00:17:56.860 --> 00:17:59.870
So there's a mean test that's
given to look at something.

00:17:59.870 --> 00:18:01.420
So here's what they do.

00:18:01.420 --> 00:18:03.640
Imagine that you do this test,
and you were correct for two

00:18:03.640 --> 00:18:06.360
digits, correct for three,
repeating them back after you

00:18:06.360 --> 00:18:06.960
heard them.

00:18:06.960 --> 00:18:10.930
Four, five, six, seven,
you did well.

00:18:10.930 --> 00:18:13.410
And then you made mistakes at
eight, which would be very

00:18:13.410 --> 00:18:16.670
typical for a human.

00:18:16.670 --> 00:18:19.240
So your span, therefore,
is seven.

00:18:19.240 --> 00:18:21.170
That's how much you can hold
in short-term memory.

00:18:21.170 --> 00:18:24.090
It varies a little bit from
person to person.

00:18:24.090 --> 00:18:26.670
And now they're going to do all
the rest of the experiment

00:18:26.670 --> 00:18:28.610
at one beyond your span.

00:18:28.610 --> 00:18:30.510
One beyond what you can do.

00:18:30.510 --> 00:18:32.260
One beyond what your
short-term memory--

00:18:32.260 --> 00:18:34.090
here's the limit of your
short-term memory.

00:18:34.090 --> 00:18:35.770
Let's add one thing.

00:18:35.770 --> 00:18:40.490
And one thing that we know helps
memory is repetition.

00:18:40.490 --> 00:18:43.070
It's not the best way to learn
things, but it is a powerful

00:18:43.070 --> 00:18:44.400
way to learn things.

00:18:44.400 --> 00:18:47.680
So now they give the entire
experiment with eight digits

00:18:47.680 --> 00:18:51.190
at a time for this person, one
beyond what you could report.

00:18:51.190 --> 00:18:53.150
So they give you this, and
they give you this.

00:18:53.150 --> 00:18:56.570
But they give you one boost,
which is every third one, it's

00:18:56.570 --> 00:18:58.370
the same sequence over
and over again.

00:18:58.370 --> 00:19:00.970
So some you hear once
and once only.

00:19:00.970 --> 00:19:03.410
And one, you hear over
and over again.

00:19:03.410 --> 00:19:04.580
Is that OK?

00:19:04.580 --> 00:19:06.380
Now what would happen if you
think, if you hear eight

00:19:06.380 --> 00:19:08.460
digits over and over again?

00:19:08.460 --> 00:19:09.090
What would happen?

00:19:09.090 --> 00:19:10.320
You know this in
your own life.

00:19:10.320 --> 00:19:13.590
Your Social Security number,
telephone numbers, things--

00:19:13.590 --> 00:19:14.330
what happens?

00:19:14.330 --> 00:19:16.340
Can you learn eight digits
if you do it over and--?

00:19:16.340 --> 00:19:17.240
Yeah.

00:19:17.240 --> 00:19:21.020
It wouldn't be easy, but you
get it with repetition.

00:19:21.020 --> 00:19:24.360
So here's what happens
in typical people.

00:19:24.360 --> 00:19:25.380
Here's the repetition.

00:19:25.380 --> 00:19:26.980
Here's how well they do for the

00:19:26.980 --> 00:19:29.720
non-repeated super-span ones.

00:19:29.720 --> 00:19:30.970
They never get that
good because they

00:19:30.970 --> 00:19:32.470
basically can't do that.

00:19:32.470 --> 00:19:34.510
Occasionally they get lucky,
but mostly not.

00:19:34.510 --> 00:19:37.110
But the ones that repeat, they
get better and better.

00:19:37.110 --> 00:19:39.790
So repetition is another way
to break the bounds of

00:19:39.790 --> 00:19:43.480
short-term memory and register
something in long-term memory.

00:19:43.480 --> 00:19:47.110
But in HM, they did this for
hundreds of trials, and he

00:19:47.110 --> 00:19:49.890
never got better at the
repeated series.

00:19:49.890 --> 00:19:52.880
So you could do endless
repetition with him, or it

00:19:52.880 --> 00:19:53.870
could be personally important.

00:19:53.870 --> 00:19:55.330
He wouldn't remember it.

00:19:55.330 --> 00:19:58.500
So let me add one more
piece to the story.

00:19:58.500 --> 00:20:01.460
If I talk to you and say, here
are the digits, tell that back

00:20:01.460 --> 00:20:03.670
to me, that's an auditory-verbal
experiment.

00:20:03.670 --> 00:20:05.690
So they're words, and
you're hearing them.

00:20:05.690 --> 00:20:07.960
You can also test short-term
memory and long-term memory

00:20:07.960 --> 00:20:10.870
using a visuo-spatial
test of blocks.

00:20:10.870 --> 00:20:14.260
So you're seeing what the
examiner sees, with numbers on

00:20:14.260 --> 00:20:16.490
the back of each of these.

00:20:16.490 --> 00:20:19.700
The person taking the test
just sees these blocks as

00:20:19.700 --> 00:20:22.180
black boxes without
any number.

00:20:22.180 --> 00:20:24.930
You need that because you have
a piece of paper with numbers

00:20:24.930 --> 00:20:26.830
on it that tell you
what to tap.

00:20:26.830 --> 00:20:27.490
OK?

00:20:27.490 --> 00:20:30.310
So when you give this test,
you might see five, seven,

00:20:30.310 --> 00:20:33.190
eight, and you would tap with
your hand, five, seven, eight,

00:20:33.190 --> 00:20:35.970
and the person would
tap that back.

00:20:35.970 --> 00:20:37.640
And I can tell you what made me
pretty nervous when I gave

00:20:37.640 --> 00:20:39.850
this test to a person
who had a long one?

00:20:39.850 --> 00:20:43.170
Because I'd have to read the
numbers, practice it, and I'd

00:20:43.170 --> 00:20:45.230
be desperately trying to get
eight my head, which I could

00:20:45.230 --> 00:20:46.530
barely do, maybe.

00:20:46.530 --> 00:20:49.400
Because I could tap eight back,
so if a person was--

00:20:49.400 --> 00:20:50.300
and then we'd do
the same thing.

00:20:50.300 --> 00:20:54.380
We'd say, if you can do eight,
we'll test you at nine and

00:20:54.380 --> 00:20:55.690
repeat them.

00:20:55.690 --> 00:20:57.410
And here's two findings,
and let's talk

00:20:57.410 --> 00:20:58.530
about HM for a moment.

00:20:58.530 --> 00:21:02.460
So one finding is now, in
patients now-- we're going to

00:21:02.460 --> 00:21:03.720
switch one thing a little bit.

00:21:03.720 --> 00:21:06.900
HM had the left and the right
hippocampus removed.

00:21:06.900 --> 00:21:08.960
In these patients, they
had either the left

00:21:08.960 --> 00:21:10.030
or the right removed.

00:21:10.030 --> 00:21:11.510
Either the left or the right.

00:21:11.510 --> 00:21:14.870
So here's what happens for
normal controls, for the

00:21:14.870 --> 00:21:16.360
numbers that are beyond
their span.

00:21:16.360 --> 00:21:18.780
Not too good, low performance,
but they benefit from

00:21:18.780 --> 00:21:20.090
repetition.

00:21:20.090 --> 00:21:23.050
Here's what happens for the
patients with left hippocampal

00:21:23.050 --> 00:21:25.290
removals for the
auditory-verbal

00:21:25.290 --> 00:21:26.540
things for the digits.

00:21:31.970 --> 00:21:35.350
So they're good for immediate
memory for this stuff, but

00:21:35.350 --> 00:21:38.190
they're terrible at developing
a long-term memory over here,

00:21:38.190 --> 00:21:39.680
if it's verbal.

00:21:39.680 --> 00:21:41.790
But the patients with
the right temporal

00:21:41.790 --> 00:21:44.320
ones are just fine.

00:21:44.320 --> 00:21:46.370
For spatial things, it's
exactly the opposite.

00:21:46.370 --> 00:21:48.870
The group that's really impaired
are the patients with

00:21:48.870 --> 00:21:52.060
a large right hippocampal
removal, and the people with

00:21:52.060 --> 00:21:54.360
the left hippocampal
do just fine.

00:21:54.360 --> 00:21:57.970
So we have two separations in
the brain for learning things

00:21:57.970 --> 00:21:59.980
in the medial temporal lobes.

00:21:59.980 --> 00:22:02.710
Neither the left nor the right
are involved in short-term

00:22:02.710 --> 00:22:06.180
memory, short-term memory
lasting only seconds.

00:22:06.180 --> 00:22:09.160
The left is important for verbal
long-term memories, and

00:22:09.160 --> 00:22:12.840
the right is important for
long-term spatial memories,

00:22:12.840 --> 00:22:15.390
which is very consistent with
what we know about left and

00:22:15.390 --> 00:22:19.770
right hemispheres for verbal
and spatial knowledge.

00:22:19.770 --> 00:22:25.300
So let me say a word more about
how we think about the

00:22:25.300 --> 00:22:28.040
organization of memory, and
we'll come back to this chart.

00:22:28.040 --> 00:22:30.040
So we're talking here, so far,
about what people call

00:22:30.040 --> 00:22:32.160
declarative or explicit
memory.

00:22:32.160 --> 00:22:34.160
You study some material,
you get tested on it.

00:22:34.160 --> 00:22:35.710
What did you study?

00:22:35.710 --> 00:22:37.430
And people sometimes make
a difference between

00:22:37.430 --> 00:22:38.930
two kinds of things.

00:22:38.930 --> 00:22:43.220
So episodic memory might
be if I ask you--

00:22:43.220 --> 00:22:45.360
OK, I need a volunteer
in your seat.

00:22:45.360 --> 00:22:46.760
It's really easy, this one.

00:22:54.580 --> 00:22:55.340
OK, thanks.

00:22:55.340 --> 00:22:56.830
I saw some people getting
ready to go.

00:22:56.830 --> 00:22:59.630
Well, OK.

00:22:59.630 --> 00:23:03.110
OK, what did you have for
breakfast this morning?

00:23:03.110 --> 00:23:05.170
And you don't have to answer
honestly if its embarrassing.

00:23:05.170 --> 00:23:06.740
You can just give
us a standard--

00:23:06.740 --> 00:23:08.460
(LAUGHING) I'm not
looking to make--

00:23:08.460 --> 00:23:09.580
yeah.

00:23:09.580 --> 00:23:10.020
AUDIENCE: Cereal.

00:23:10.020 --> 00:23:10.166
PROFESSOR: Cereal.

00:23:10.166 --> 00:23:10.945
How about you?

00:23:10.945 --> 00:23:11.280
AUDIENCE: What?

00:23:11.280 --> 00:23:11.890
AUDIENCE: Yogurt.

00:23:11.890 --> 00:23:12.060
PROFESSOR: Yogurt.

00:23:12.060 --> 00:23:12.680
Cereal and yogurt.

00:23:12.680 --> 00:23:13.790
Two excellent healthy answers.

00:23:13.790 --> 00:23:14.180
OK.

00:23:14.180 --> 00:23:17.930
So now I'm going to ask you--

00:23:17.930 --> 00:23:18.990
if I could ask you first.

00:23:18.990 --> 00:23:22.080
How many feet in a yard?

00:23:22.080 --> 00:23:23.060
AUDIENCE: Three.

00:23:23.060 --> 00:23:23.840
PROFESSOR: Excellent.

00:23:23.840 --> 00:23:24.710
OK.

00:23:24.710 --> 00:23:27.260
And now I'm going to ask
you, when and where

00:23:27.260 --> 00:23:29.870
did you learn that?

00:23:29.870 --> 00:23:30.610
OK?

00:23:30.610 --> 00:23:34.520
And you might go,
I don't know.

00:23:34.520 --> 00:23:35.500
I just learned it.

00:23:35.500 --> 00:23:36.360
Right?

00:23:36.360 --> 00:23:38.220
Or if I ask you, capital
of France.

00:23:38.220 --> 00:23:38.670
AUDIENCE: Paris.

00:23:38.670 --> 00:23:39.130
PROFESSOR: Paris.

00:23:39.130 --> 00:23:40.450
Excellent answer.

00:23:40.450 --> 00:23:44.540
When was the shocking moment
when this was revealed to you?

00:23:44.540 --> 00:23:46.600
When you said, I'll never
trust the world again--

00:23:46.600 --> 00:23:47.380
yes.

00:23:47.380 --> 00:23:48.250
Who knows, right?

00:23:48.250 --> 00:23:51.040
So what you're learning is,
episodic memory is memory for

00:23:51.040 --> 00:23:53.660
specific events, like what I did
last night, what I had for

00:23:53.660 --> 00:23:56.880
breakfast, specific
time and place.

00:23:56.880 --> 00:23:59.410
Semantic memory is what we could
call generic memory.

00:23:59.410 --> 00:24:02.610
We know lots of things about
the world, but we might not

00:24:02.610 --> 00:24:04.420
know when or where we learned.

00:24:04.420 --> 00:24:06.220
So that's a distinction
psychologists find useful to

00:24:06.220 --> 00:24:07.280
think about.

00:24:07.280 --> 00:24:08.170
So how about in HM?

00:24:08.170 --> 00:24:10.850
We said he can't remember
digits, or faces,

00:24:10.850 --> 00:24:11.810
or things like that.

00:24:11.810 --> 00:24:15.100
How about generic memory?

00:24:15.100 --> 00:24:19.310
So here's an experiment that
I was involved in, where we

00:24:19.310 --> 00:24:20.500
said-- because there
were some debates

00:24:20.500 --> 00:24:21.700
about this at the time.

00:24:21.700 --> 00:24:25.270
So we gave HM a multiple choice
test where he saw words

00:24:25.270 --> 00:24:29.240
or phrases that entered the
language after the onset of

00:24:29.240 --> 00:24:30.140
his amnesia.

00:24:30.140 --> 00:24:32.300
So we thought he wouldn't learn
them, but we didn't know

00:24:32.300 --> 00:24:34.450
for sure, from just everyday
experience.

00:24:34.450 --> 00:24:36.530
And by the way, I should tell
you, he watched tons of

00:24:36.530 --> 00:24:37.670
television.

00:24:37.670 --> 00:24:39.710
What his experience was like
watching television is an

00:24:39.710 --> 00:24:40.760
interesting question.

00:24:40.760 --> 00:24:42.230
You should be thinking now about
questions you want to

00:24:42.230 --> 00:24:44.510
ask about HM.

00:24:44.510 --> 00:24:46.180
In a moment, I'll come
back to that.

00:24:46.180 --> 00:24:48.430
But look at-- so, he has four
choices for the word

00:24:48.430 --> 00:24:51.290
amniocentesis, including the
correct one, but he picks "an

00:24:51.290 --> 00:24:53.110
infectious inflammatory
disease of the

00:24:53.110 --> 00:24:54.630
intestines." apartheid?

00:24:54.630 --> 00:24:57.455
In front of him is the correct
answer, two other answers, and

00:24:57.455 --> 00:25:00.010
he picks, "the separation of
young cows that have not yet

00:25:00.010 --> 00:25:00.870
given birth to calves."

00:25:00.870 --> 00:25:03.000
Boat people? "People who cater
bon voyage parties," which is

00:25:03.000 --> 00:25:04.600
completely the wrong--

00:25:04.600 --> 00:25:06.360
Brainwash, "a fluid that
surrounds and bathes the

00:25:06.360 --> 00:25:09.190
brain." Granola, "a portable
keyboard wind instrument."

00:25:09.190 --> 00:25:11.250
Software, "expensive clothing
made of a soft, twilled

00:25:11.250 --> 00:25:13.990
fabric."

00:25:13.990 --> 00:25:15.290
What's going on with him?

00:25:15.290 --> 00:25:18.980
Are these, like, crazy
weird answers?

00:25:18.980 --> 00:25:21.830
Or how would you-- what do we
think is going on with him?

00:25:21.830 --> 00:25:22.750
Why is he picking these?

00:25:22.750 --> 00:25:24.860
The other one is right
in front of him.

00:25:24.860 --> 00:25:27.950
What's he doing?

00:25:27.950 --> 00:25:29.380
AUDIENCE: Looking at
words he knows.

00:25:29.380 --> 00:25:30.130
PROFESSOR: Yeah.

00:25:30.130 --> 00:25:31.660
He's doing what teachers
teach you in school.

00:25:31.660 --> 00:25:33.420
Take apart the word into
its parts, right?

00:25:33.420 --> 00:25:37.680
And so he goes, software, soft
clothes that you wear.

00:25:37.680 --> 00:25:39.540
But what does that mean?

00:25:39.540 --> 00:25:42.850
He's missed, in every sense,
the advent of computers and

00:25:42.850 --> 00:25:44.140
software in the world.

00:25:44.140 --> 00:25:46.450
The idea, the concept,
the words--

00:25:46.450 --> 00:25:49.230
it's as if it didn't exist,
because in 1953, although

00:25:49.230 --> 00:25:52.050
software existed, it wasn't
a very popular concept.

00:25:52.050 --> 00:25:54.110
Most people didn't know about
the concept of that.

00:25:54.110 --> 00:25:55.600
In 1953--

00:25:55.600 --> 00:25:57.800
the way we looked at these,
these were not words that were

00:25:57.800 --> 00:26:02.690
in the dictionary
in the 1950s.

00:26:02.690 --> 00:26:05.085
So now, what would you like to
know about a man like HM?

00:26:08.055 --> 00:26:08.550
Yeah.

00:26:08.550 --> 00:26:11.025
AUDIENCE: So obviously he
couldn't remember something

00:26:11.025 --> 00:26:13.747
that happened two hours ago,
but if you're having a

00:26:13.747 --> 00:26:16.717
conversation with him, would he
not remember what you guys

00:26:16.717 --> 00:26:17.460
were talking about?

00:26:17.460 --> 00:26:20.430
Or could he remember the context
of the conversation?

00:26:20.430 --> 00:26:20.710
PROFESSOR: Right.

00:26:20.710 --> 00:26:22.380
So you're saying, he wouldn't
remember something two days

00:26:22.380 --> 00:26:23.300
ago, but you're talking
with him.

00:26:23.300 --> 00:26:23.870
What's that like?

00:26:23.870 --> 00:26:25.500
Right?

00:26:25.500 --> 00:26:26.650
So let me give you
one experimental

00:26:26.650 --> 00:26:27.410
thing and then an exper--

00:26:27.410 --> 00:26:29.310
so one experiment they did is
they left him with something

00:26:29.310 --> 00:26:32.650
like six numbers, well within
the span, and said, remember

00:26:32.650 --> 00:26:33.850
these numbers.

00:26:33.850 --> 00:26:37.210
And it could be, you know, 1,
2, 3, 4, 5, 6-- well, that

00:26:37.210 --> 00:26:38.910
wouldn't work so well.

00:26:38.910 --> 00:26:40.330
2, 4, 6, 8, 9.

00:26:40.330 --> 00:26:40.840
2, 4, 6, 8, 9.

00:26:40.840 --> 00:26:42.080
They leave the room
for 15 minutes.

00:26:42.080 --> 00:26:42.650
They come back in.

00:26:42.650 --> 00:26:44.020
He's an excellent subject.

00:26:44.020 --> 00:26:44.950
He's going, 2, 4, 6, 8, 9.

00:26:44.950 --> 00:26:46.030
2, 4, 6, 8, 9.

00:26:46.030 --> 00:26:46.200
What were the numbers?

00:26:46.200 --> 00:26:46.840
2, 4, 6, 8, 9.

00:26:46.840 --> 00:26:48.480
They go, that's excellent.

00:26:48.480 --> 00:26:49.880
Was there any trick you
did to do that?

00:26:49.880 --> 00:26:50.400
Or did you just--

00:26:50.400 --> 00:26:51.540
And he'd go, oh, oh, oh.

00:26:51.540 --> 00:26:53.670
What were the numbers?

00:26:53.670 --> 00:26:56.220
As soon as a thought is out
of the forefront of

00:26:56.220 --> 00:26:58.390
his mind, it's gone.

00:26:58.390 --> 00:26:59.590
So, yeah.

00:26:59.590 --> 00:27:03.950
He would not recognize us, and
we spent many hours with him.

00:27:03.950 --> 00:27:06.080
A striking thing, which was
amazing at first and then

00:27:06.080 --> 00:27:08.040
sometimes irksome if you worked
with him-- and he was a

00:27:08.040 --> 00:27:09.410
delightful guy--

00:27:09.410 --> 00:27:11.660
was that he would tell you
a story from his past.

00:27:11.660 --> 00:27:13.455
And he didn't have a huge array
of stories, because they

00:27:13.455 --> 00:27:16.180
were all from a long time ago.

00:27:16.180 --> 00:27:18.250
And he would tell you about
a gun collection he had.

00:27:18.250 --> 00:27:20.090
And he would tell he had three
guns, and they had this

00:27:20.090 --> 00:27:21.110
property and that property.

00:27:21.110 --> 00:27:24.180
He'd finish the story, and if
you just stayed quiet for a

00:27:24.180 --> 00:27:28.080
moment, he would forget that he
told you the story, but it

00:27:28.080 --> 00:27:30.130
would be vaguely on his
mind because he

00:27:30.130 --> 00:27:31.390
just told it to you.

00:27:31.390 --> 00:27:33.200
And he would say to you, hey,
did I tell you about my gun

00:27:33.200 --> 00:27:33.850
collection?

00:27:33.850 --> 00:27:36.730
And he would tell you, almost
verbatim, the same story.

00:27:36.730 --> 00:27:37.840
You'd just wait a moment.

00:27:37.840 --> 00:27:40.790
He forgot that he told you this
story, but it's slightly

00:27:40.790 --> 00:27:42.080
on his mind because you
just told you it.

00:27:42.080 --> 00:27:43.300
And he goes, hey, did
I tell you my--

00:27:43.300 --> 00:27:47.610
he could tell you this until you
could bear it no longer.

00:27:47.610 --> 00:27:50.140
So it was kind of amazing
how short his

00:27:50.140 --> 00:27:51.400
short-term memory was.

00:27:51.400 --> 00:27:52.740
Without this part of the brain,

00:27:52.740 --> 00:27:54.620
it's just a few seconds.

00:27:54.620 --> 00:27:57.200
His conversations were
OK, but I would

00:27:57.200 --> 00:27:58.240
call them pretty shallow.

00:27:58.240 --> 00:28:00.050
I don't know how to put it.

00:28:00.050 --> 00:28:02.410
You know, when you talk with
somebody and you're on the

00:28:02.410 --> 00:28:03.900
phone, and they're
doing something

00:28:03.900 --> 00:28:04.970
else, and you can tell--

00:28:04.970 --> 00:28:06.900
for him it was a little
bit like that.

00:28:06.900 --> 00:28:08.195
He was-- so, good social
convention.

00:28:08.195 --> 00:28:09.080
He would smile.

00:28:09.080 --> 00:28:10.170
He would chat with you.

00:28:10.170 --> 00:28:12.520
But if what you were talking
about depended on remembering

00:28:12.520 --> 00:28:15.750
something from three sentences
ago in any detail, gone.

00:28:15.750 --> 00:28:18.510
Just seconds, unless he's
practicing and practicing and

00:28:18.510 --> 00:28:19.960
practicing without doing
anything else.

00:28:19.960 --> 00:28:20.190
Yeah.

00:28:20.190 --> 00:28:22.640
AUDIENCE: So he did have
the capability to learn

00:28:22.640 --> 00:28:23.490
procedurally.

00:28:23.490 --> 00:28:23.850
PROFESSOR: Yes.

00:28:23.850 --> 00:28:24.550
We're going to come to that.

00:28:24.550 --> 00:28:26.510
So the question is, did he
ever learn-- he had other

00:28:26.510 --> 00:28:28.320
kinds of learning that
were amazingly

00:28:28.320 --> 00:28:29.600
completely normal in him.

00:28:29.600 --> 00:28:30.870
Procedural memory, you're
absolutely right.

00:28:30.870 --> 00:28:33.590
And we'll come to that just
in about 20 minutes, OK?

00:28:33.590 --> 00:28:34.480
Exactly right.

00:28:34.480 --> 00:28:36.700
And if I don't answer that well,
make me do that again.

00:28:36.700 --> 00:28:37.250
OK?

00:28:37.250 --> 00:28:37.800
But yeah.

00:28:37.800 --> 00:28:39.094
Yeah?

00:28:39.094 --> 00:28:40.942
AUDIENCE: Was he able to
remember that he couldn't

00:28:40.942 --> 00:28:41.870
remember things?

00:28:41.870 --> 00:28:42.130
PROFESSOR: Yeah.

00:28:42.130 --> 00:28:42.740
Really good question.

00:28:42.740 --> 00:28:44.990
Was he able to remember that he
couldn't remember things?

00:28:44.990 --> 00:28:46.480
Yes, he knew he had
a bad memory.

00:28:46.480 --> 00:28:49.310
So now the question is, in what
sense did he know he had

00:28:49.310 --> 00:28:50.530
a bad memory?

00:28:50.530 --> 00:28:55.060
Like, I know I can't draw very
well, amongst many other

00:28:55.060 --> 00:28:56.200
things I can't do very well.

00:28:56.200 --> 00:28:59.790
But if you talked to him,
he'd say, yeah, I

00:28:59.790 --> 00:29:00.990
don't have a good memory.

00:29:00.990 --> 00:29:02.800
But he's looking at you, and
he's going, like, I don't know

00:29:02.800 --> 00:29:05.580
where I am, who you are,
or what I'm doing.

00:29:05.580 --> 00:29:08.660
Everything is vague, for
many, many years.

00:29:08.660 --> 00:29:09.810
I must have bad memory.

00:29:09.810 --> 00:29:11.750
Like, you would too, like
a science fiction movie.

00:29:11.750 --> 00:29:13.500
You woke up, you didn't know who
you were, where you were,

00:29:13.500 --> 00:29:15.870
what you were doing, and you
can't remember anything except

00:29:15.870 --> 00:29:16.850
from the most distant past.

00:29:16.850 --> 00:29:18.800
You'd go, something is not
good with my memory.

00:29:18.800 --> 00:29:20.670
But you're smart, so you
would know that.

00:29:20.670 --> 00:29:23.030
What I don't know is if you woke
him up at 2 AM and you

00:29:23.030 --> 00:29:25.510
said, HM, how's your memory?

00:29:25.510 --> 00:29:27.940
I don't know if he'd go, well,
I'm a famous amnesic, of

00:29:27.940 --> 00:29:29.620
course it's terrible.

00:29:29.620 --> 00:29:32.240
I don't know if he knew it as
a fact, like we might know

00:29:32.240 --> 00:29:34.740
about what we're good and not
good at, or he just knew it by

00:29:34.740 --> 00:29:38.690
constantly being aware that he
didn't know anything he ought

00:29:38.690 --> 00:29:40.860
to know via memory.

00:29:40.860 --> 00:29:42.580
Does that make sense?

00:29:42.580 --> 00:29:43.870
Anything else about--?

00:29:43.870 --> 00:29:44.210
Yeah.

00:29:44.210 --> 00:29:47.870
AUDIENCE: How does one explain
scientific consent to--?

00:29:47.870 --> 00:29:48.290
PROFESSOR: Oh, this is good.

00:29:48.290 --> 00:29:49.310
It's a very interesting
question.

00:29:49.310 --> 00:29:51.750
Informed consent.

00:29:51.750 --> 00:29:55.800
So technically he was conserved
by the state once

00:29:55.800 --> 00:29:58.790
his family had passed away.

00:29:58.790 --> 00:30:00.380
But, you know, his intelligence
was fine, so you

00:30:00.380 --> 00:30:02.130
had to be a little bit careful
to do it right.

00:30:02.130 --> 00:30:03.760
But you could tell him, we're
going to do an experiment.

00:30:03.760 --> 00:30:04.860
We're going to test your
memory for words.

00:30:04.860 --> 00:30:05.730
Is that OK with you?

00:30:05.730 --> 00:30:06.370
And he'd say yes.

00:30:06.370 --> 00:30:08.320
Of course, if you had him read
a 12-page document, by the

00:30:08.320 --> 00:30:10.220
time he got to page 12-- but you
could talk him through it.

00:30:10.220 --> 00:30:11.895
His intelligence was fine.

00:30:11.895 --> 00:30:14.530
You could worry about the
edges, if there--

00:30:14.530 --> 00:30:18.920
so I think the investigators
felt an extra ethical thing,

00:30:18.920 --> 00:30:21.940
also, not to ask him to do
anything that a person

00:30:21.940 --> 00:30:22.880
might say no to.

00:30:22.880 --> 00:30:24.825
He was a wonderful participant,
but it was an

00:30:24.825 --> 00:30:27.846
extra responsibility for the
reason you're saying.

00:30:27.846 --> 00:30:31.550
AUDIENCE: Does that mean he
was happy all the time?

00:30:31.550 --> 00:30:32.630
PROFESSOR: Was he happy
all the time?

00:30:32.630 --> 00:30:33.620
Excellent question.

00:30:33.620 --> 00:30:36.650
He was a very sort of mellow--

00:30:36.650 --> 00:30:36.920
yeah.

00:30:36.920 --> 00:30:38.180
Well, you could think for a
moment whether having no

00:30:38.180 --> 00:30:39.480
memory would make you
happy or sad.

00:30:39.480 --> 00:30:41.100
So he couldn't hold a job.

00:30:41.100 --> 00:30:43.060
He couldn't sustain a human
relation because he would

00:30:43.060 --> 00:30:44.310
never know he'd met
you before.

00:30:47.220 --> 00:30:49.680
So, those are things we often
think of as pretty big in our

00:30:49.680 --> 00:30:51.790
happiness, the kind of the
work we do, in some broad

00:30:51.790 --> 00:30:54.250
sense, and the people
we relate to, right?

00:30:54.250 --> 00:30:56.640
And those were all
gone for him.

00:30:56.640 --> 00:30:58.410
On the other hand, he didn't
exactly realize that.

00:30:58.410 --> 00:30:59.420
He didn't realize
he didn't work.

00:30:59.420 --> 00:31:01.460
He didn't realize he didn't
have relations he might be

00:31:01.460 --> 00:31:04.440
expected to have, because he'd
forget that about as fast as

00:31:04.440 --> 00:31:06.840
he could think about it.

00:31:06.840 --> 00:31:09.180
So I--

00:31:09.180 --> 00:31:12.770
he was kind of a very, I would
say, mild, happy person.

00:31:12.770 --> 00:31:15.350
So there's three possibilities
that crossed my head.

00:31:15.350 --> 00:31:17.020
And scientifically, we can
never figure them out.

00:31:17.020 --> 00:31:21.320
One of them is, maybe just
he was that way.

00:31:21.320 --> 00:31:23.610
I mean some people are just
mellow and happy, right?

00:31:23.610 --> 00:31:26.200
They don't have--
without amnesia.

00:31:26.200 --> 00:31:29.860
Second possibility is-- he had
removed, also, the amygdala.

00:31:29.860 --> 00:31:31.930
We'll talk about that
later in the course.

00:31:31.930 --> 00:31:34.070
That's a structure that's pretty
important for some

00:31:34.070 --> 00:31:35.990
aspects of emotion
and feeling.

00:31:35.990 --> 00:31:37.950
So that could have
been relevant.

00:31:37.950 --> 00:31:39.760
The third one is, but this is
kind of getting to your

00:31:39.760 --> 00:31:41.700
question, which was just--

00:31:41.700 --> 00:31:42.930
intuitively, I feel like this.

00:31:42.930 --> 00:31:45.590
Again, this is not science
but is just a feeling.

00:31:45.590 --> 00:31:50.030
That what makes us mostly
sort of happy or sad?

00:31:50.030 --> 00:31:53.840
For me, it's mostly relatively
recent things that went well

00:31:53.840 --> 00:31:57.400
or didn't go well, or things
I'm sad about recently, and

00:31:57.400 --> 00:31:59.500
things on my horizon.

00:31:59.500 --> 00:32:03.030
Things I'm worried about, or
things I'm looking forward to.

00:32:03.030 --> 00:32:05.030
And you guys might be different,
but I don't get

00:32:05.030 --> 00:32:07.490
that worked up these days
about second grade.

00:32:07.490 --> 00:32:10.170
At the time I'm sure it was
pretty emotional about it.

00:32:10.170 --> 00:32:13.820
So if you think about him, he
doesn't remember anything

00:32:13.820 --> 00:32:15.560
recent that was good or bad.

00:32:15.560 --> 00:32:20.180
He doesn't remember anything
coming up that's interesting,

00:32:20.180 --> 00:32:23.020
or threatening, or risky,
or unpleasant.

00:32:23.020 --> 00:32:24.800
So he lives in this
constant now.

00:32:24.800 --> 00:32:26.670
You know, people tell you,
live in the moment.

00:32:26.670 --> 00:32:29.820
Nobody could have lived more
in the moment than he did.

00:32:29.820 --> 00:32:32.530
It's impossible to live more
in the moment when the

00:32:32.530 --> 00:32:34.050
previous moment you can
consult is from

00:32:34.050 --> 00:32:36.140
30, 40 years ago.

00:32:36.140 --> 00:32:39.650
So, yeah, he was pretty happy,
and a pretty pleasant guy.

00:32:39.650 --> 00:32:41.400
Essentially, I don't think most
of us want that life,

00:32:41.400 --> 00:32:43.180
although he's probably happier
than many people who have all

00:32:43.180 --> 00:32:45.220
their memories.

00:32:45.220 --> 00:32:48.110
One thing he's the opposite
of, just to remind you--

00:32:48.110 --> 00:32:49.730
we talked about it now, but just
to make it clear-- it's

00:32:49.730 --> 00:32:51.460
the opposite of television
amnesia.

00:32:51.460 --> 00:32:56.010
What happens in soap operas and
murder movies or comedy

00:32:56.010 --> 00:32:57.460
shows when they get bonked
on the head?

00:33:00.310 --> 00:33:01.840
Television amnesia.

00:33:01.840 --> 00:33:05.850
They forget who they are and
where they came from, right?

00:33:05.850 --> 00:33:07.820
And then they marry their worst
enemy or something like

00:33:07.820 --> 00:33:09.800
this, to make the story
go forward.

00:33:09.800 --> 00:33:10.800
That's television amnesia.

00:33:10.800 --> 00:33:12.650
But after that, they're
kind of fine, right?

00:33:12.650 --> 00:33:14.380
They're kind of operating
but they just

00:33:14.380 --> 00:33:15.560
forgot where they come--

00:33:15.560 --> 00:33:19.350
that we never see after
brain injuries.

00:33:19.350 --> 00:33:24.520
We never see a person who
forgets who they are but gets

00:33:24.520 --> 00:33:25.260
around fine in the world.

00:33:25.260 --> 00:33:26.210
HM is the opposite.

00:33:26.210 --> 00:33:29.560
He remembers stuff from
before, but he

00:33:29.560 --> 00:33:30.440
can't learn new things.

00:33:30.440 --> 00:33:31.690
Opposite of TV amnesia.

00:33:33.950 --> 00:33:35.280
Any other questions about HM?

00:33:35.280 --> 00:33:36.698
Yeah.

00:33:36.698 --> 00:33:38.634
AUDIENCE: As he got older,
could he recognize

00:33:38.634 --> 00:33:41.540
people from his past?

00:33:41.540 --> 00:33:43.230
AUDIENCE: Yes, and let me talk
a little bit about that.

00:33:43.230 --> 00:33:45.060
If I don't answer that,
let me know.

00:33:45.060 --> 00:33:46.080
So let me talk about this.

00:33:46.080 --> 00:33:48.160
This sort of touches on this,
loss of already-known

00:33:48.160 --> 00:33:49.080
information.

00:33:49.080 --> 00:33:51.310
And if I don't answer you
well, just put your

00:33:51.310 --> 00:33:53.550
hand up again, OK?

00:33:53.550 --> 00:33:56.780
So it's easy to test learning
new things, because you can

00:33:56.780 --> 00:33:58.410
give somebody something in the
laboratory and you see

00:33:58.410 --> 00:33:59.420
if they learn it.

00:33:59.420 --> 00:34:00.790
Knowing things they should
have known from

00:34:00.790 --> 00:34:02.090
before is much harder.

00:34:02.090 --> 00:34:05.430
Different people know different
things, in the past.

00:34:05.430 --> 00:34:06.660
So here's how they'd test it.

00:34:06.660 --> 00:34:10.630
And here's a thing that people
have discovered, which is

00:34:10.630 --> 00:34:14.989
people are pretty
unknowledgeable, mostly, about

00:34:14.989 --> 00:34:17.550
faces of people who
were famous at the

00:34:17.550 --> 00:34:18.659
time they were famous.

00:34:18.659 --> 00:34:20.929
But if you're, like, 20, you
don't know people who were

00:34:20.929 --> 00:34:21.929
famous before you were born.

00:34:21.929 --> 00:34:22.600
Mostly.

00:34:22.600 --> 00:34:24.690
There's some exceptions.

00:34:24.690 --> 00:34:28.210
So they'd take faces from
different decades, and they'd

00:34:28.210 --> 00:34:30.510
show you them and they'd ask
you, do you know who this is?

00:34:30.510 --> 00:34:32.800
So do you know who this is?

00:34:32.800 --> 00:34:34.050
This is Lindbergh.

00:34:36.370 --> 00:34:37.620
Frank Sinatra.

00:34:39.699 --> 00:34:40.330
Bob Hope.

00:34:40.330 --> 00:34:42.679
Does that name even
ring a bell?

00:34:42.679 --> 00:34:43.639
I know.

00:34:43.639 --> 00:34:44.750
This is totally--

00:34:44.750 --> 00:34:49.330
however old you are is totally
how you do on this test.

00:34:49.330 --> 00:34:51.002
Lyndon Baines Johnson, president
of the United States

00:34:51.002 --> 00:34:53.780
after Kennedy.

00:34:53.780 --> 00:34:56.989
Douglas MacArthur,
the general.

00:34:56.989 --> 00:34:58.090
This, you might.

00:34:58.090 --> 00:34:59.340
Richard Nixon.

00:35:02.910 --> 00:35:06.590
McCarthy, as in McCarthyism.

00:35:06.590 --> 00:35:08.610
Golda Meir.

00:35:08.610 --> 00:35:10.275
Elvis.

00:35:10.275 --> 00:35:12.920
OK now.

00:35:12.920 --> 00:35:13.790
If you--

00:35:13.790 --> 00:35:15.270
pretend you became amnesic.

00:35:15.270 --> 00:35:16.210
You became amnesic.

00:35:16.210 --> 00:35:19.320
Pretend you're old enough to
become amnesic in 1990.

00:35:19.320 --> 00:35:20.520
This is if you're feeling--

00:35:20.520 --> 00:35:24.490
would you know this
face, in 1990?

00:35:24.490 --> 00:35:25.360
Let's say 1980.

00:35:25.360 --> 00:35:26.430
Let's make it simple.

00:35:26.430 --> 00:35:27.170
1980?

00:35:27.170 --> 00:35:28.280
No.

00:35:28.280 --> 00:35:29.956
Would you know this face?

00:35:29.956 --> 00:35:30.850
No.

00:35:30.850 --> 00:35:31.160
OK.

00:35:31.160 --> 00:35:32.630
But you know them now.

00:35:32.630 --> 00:35:37.260
So that tells you that you can
put faces to decades, roughly.

00:35:37.260 --> 00:35:41.000
You know, 1980, only Barack
Obama's family knew Barack

00:35:41.000 --> 00:35:41.760
Obama's face, right?

00:35:41.760 --> 00:35:43.010
And now we all do.

00:35:45.410 --> 00:35:47.000
Temporally limited retrograde
amnesia.

00:35:47.000 --> 00:35:49.400
So this is how well you do--
higher is better--

00:35:49.400 --> 00:35:50.380
in decades.

00:35:50.380 --> 00:35:52.110
In 1960, here's HM.

00:35:52.110 --> 00:35:55.160
So here's faces that were famous
in the '20s, '30s, and

00:35:55.160 --> 00:35:57.800
'40s, before he became amnesic,

00:35:57.800 --> 00:36:00.120
and he's pretty normal.

00:36:00.120 --> 00:36:01.670
These are public faces.

00:36:01.670 --> 00:36:03.610
And then he's terrible
for the '50s and

00:36:03.610 --> 00:36:04.840
'60s, after his amnesia.

00:36:04.840 --> 00:36:09.390
So his memory for faces from
the past is pretty normal.

00:36:09.390 --> 00:36:11.820
His memory for faces from the
onset of his amnesia forward

00:36:11.820 --> 00:36:12.720
is terrible.

00:36:12.720 --> 00:36:13.430
Is that OK?

00:36:13.430 --> 00:36:14.725
Does that answer your
question reasonably?

00:36:17.490 --> 00:36:19.640
But here's another really
weird thing that people

00:36:19.640 --> 00:36:23.260
observed in many cases, which is
kind of hard to understand,

00:36:23.260 --> 00:36:24.800
but it's been observed
in many cases.

00:36:24.800 --> 00:36:26.470
So here's another patient.

00:36:26.470 --> 00:36:28.590
He was a bus driver
in Washington, DC.

00:36:28.590 --> 00:36:30.850
He moved to work in a drugstore
in Boston, then a

00:36:30.850 --> 00:36:32.480
mattress factory in Boston.

00:36:32.480 --> 00:36:36.460
Then he was hospitalized at the
Boston VA in November 1965

00:36:36.460 --> 00:36:38.685
with a huge hematoma on the
right temporal parietal-- a

00:36:38.685 --> 00:36:40.010
big bleeding.

00:36:40.010 --> 00:36:41.970
And he was pretty aphasic
and stuporous.

00:36:41.970 --> 00:36:43.750
He couldn't answer questions.

00:36:43.750 --> 00:36:47.370
In a month, digit span, his
short-term memory returns.

00:36:47.370 --> 00:36:52.050
He states the date as
approximately something like

00:36:52.050 --> 00:36:54.580
September 1965.

00:36:54.580 --> 00:36:56.130
He has a severe anterograde
amnesia.

00:36:56.130 --> 00:36:58.770
For example, he fails to learned
the names of the

00:36:58.770 --> 00:37:00.550
nurses he sees every day.

00:37:00.550 --> 00:37:03.120
But when asked-- this is
the striking thing.

00:37:03.120 --> 00:37:04.760
When asked, where do you live?

00:37:04.760 --> 00:37:08.150
He says, I live in Washington,
with certainty, as if he lost

00:37:08.150 --> 00:37:10.600
all these memories
from before.

00:37:10.600 --> 00:37:13.310
By March, he starts to learn
the nurses' names.

00:37:13.310 --> 00:37:15.640
He's recovering his
memory capacity.

00:37:15.640 --> 00:37:18.330
He couldn't remember that he'd
moved to Boston, but he

00:37:18.330 --> 00:37:20.840
doesn't fight the idea that
he lives in Boston.

00:37:20.840 --> 00:37:22.870
Then he remembers that he
worked at a drugstore.

00:37:22.870 --> 00:37:24.740
Then he remembers that he worked
in a mattress factory.

00:37:24.740 --> 00:37:26.840
And by the time that he's
discharged, it's only the last

00:37:26.840 --> 00:37:29.750
24 hours before he went into the
hospital that he's lost.

00:37:29.750 --> 00:37:30.730
So two things.

00:37:30.730 --> 00:37:33.640
There's a coupling between the
anterograde amnesia and the

00:37:33.640 --> 00:37:34.600
retrograde amnesia.

00:37:34.600 --> 00:37:36.930
He can't learn new things,
but he lost old things.

00:37:36.930 --> 00:37:40.550
And as the anterograde amnesia
resolves, the retrograde

00:37:40.550 --> 00:37:43.460
amnesia resolves, in some sense,
in temporal order.

00:37:43.460 --> 00:37:47.840
It's really odd, but it's been
observed many times.

00:37:47.840 --> 00:37:51.480
So there is some link between
what the hippocampus does and

00:37:51.480 --> 00:37:54.730
memory for the past.

00:37:54.730 --> 00:37:58.010
Now in some ways, it's really
hard to study these things in

00:37:58.010 --> 00:37:59.670
detail in a patient like HM.

00:37:59.670 --> 00:38:01.530
But some other patients give
us another perspective.

00:38:01.530 --> 00:38:04.330
Although every one of these
examples is not quite perfect.

00:38:04.330 --> 00:38:07.180
Here's patients who undergo
treatment for depression with

00:38:07.180 --> 00:38:08.470
electroconvulsive therapy.

00:38:08.470 --> 00:38:09.870
And appropriately--

00:38:09.870 --> 00:38:13.510
in movies like One Flew Over
the Cuckoo's Nest, wrongly

00:38:13.510 --> 00:38:16.340
used electroconvulsive
therapy is wrong.

00:38:16.340 --> 00:38:19.310
For some patients with severe
depression who fail to respond

00:38:19.310 --> 00:38:22.210
to medications and who have
suicidal ideation--

00:38:22.210 --> 00:38:24.060
literally attempt to take their
lives or talk a lot

00:38:24.060 --> 00:38:27.320
about it in a believable way
to family and physician--

00:38:27.320 --> 00:38:29.740
electroconvulsive therapy
can be very helpful.

00:38:29.740 --> 00:38:32.710
It's a hard choice to make,
and a complicated one.

00:38:32.710 --> 00:38:33.330
Here's what happens.

00:38:33.330 --> 00:38:35.650
Patients typically go-- and it
varies-- but they go in for

00:38:35.650 --> 00:38:38.040
the electroconvulsive therapy.

00:38:38.040 --> 00:38:39.790
And nowadays, it's done,
typically, only in the right

00:38:39.790 --> 00:38:40.270
hemisphere.

00:38:40.270 --> 00:38:42.350
In this study, done some
years ago around

00:38:42.350 --> 00:38:44.670
1972, it was bilateral.

00:38:44.670 --> 00:38:46.270
They put electrodes on both
sides of the brain.

00:38:46.270 --> 00:38:50.800
They passed current and they
induced a seizure.

00:38:50.800 --> 00:38:53.590
And they do that something
like every second day for

00:38:53.590 --> 00:38:54.520
about 10 days.

00:38:54.520 --> 00:38:57.460
It's a very aggressive form
of treatment, certainly.

00:38:57.460 --> 00:39:00.460
But for some patients, the
depression just lifts, and

00:39:00.460 --> 00:39:02.040
they're no longer talking about
taking their lives.

00:39:02.040 --> 00:39:06.140
So for some people it seems to
work when other things don't.

00:39:06.140 --> 00:39:07.700
During the course of the
treatment, these patients

00:39:07.700 --> 00:39:09.700
develop and anterograde
amnesia.

00:39:09.700 --> 00:39:11.790
They become HM-like.

00:39:11.790 --> 00:39:13.990
While they're in the hospital,
they become HM-like.

00:39:13.990 --> 00:39:15.600
They can't learn new things.

00:39:15.600 --> 00:39:16.560
They forget the nurses.

00:39:16.560 --> 00:39:18.290
You can test them formally.

00:39:18.290 --> 00:39:20.710
And so Larry Squire did the
following experiment with

00:39:20.710 --> 00:39:21.910
these kinds of patients.

00:39:21.910 --> 00:39:24.800
So we don't know that it's the
hippocampus that's the

00:39:24.800 --> 00:39:26.410
critical thing in these
patients, but they

00:39:26.410 --> 00:39:28.740
look a lot like HM.

00:39:28.740 --> 00:39:30.850
And we know that the
hippocampus has

00:39:30.850 --> 00:39:32.530
a low seizure threshold.

00:39:32.530 --> 00:39:36.160
That's why it occurs so often
as the locus of epilepsy.

00:39:36.160 --> 00:39:38.950
So what they did is they
took television shows.

00:39:38.950 --> 00:39:41.020
And this is another world than
you can possibly imagine.

00:39:41.020 --> 00:39:43.590
I grew up in a world where there
were basically three

00:39:43.590 --> 00:39:44.710
television stations.

00:39:44.710 --> 00:39:46.670
And you're feeling sad for
me already, right?

00:39:46.670 --> 00:39:48.690
And so everybody knew
what was on those

00:39:48.690 --> 00:39:51.160
three television stations.

00:39:51.160 --> 00:39:52.990
But they took shows that weren't
very successful, that

00:39:52.990 --> 00:39:54.900
were on for one season only.

00:39:54.900 --> 00:39:57.320
So they weren't super popular,
but a lot of people saw them.

00:39:57.320 --> 00:39:59.250
And there were only
three television

00:39:59.250 --> 00:40:01.270
networks and stations.

00:40:01.270 --> 00:40:04.290
So now what they did is they
tested them on these events or

00:40:04.290 --> 00:40:07.410
other things like that before
they had ECT-- so this is

00:40:07.410 --> 00:40:08.780
their knowledge before
they had ECT.

00:40:08.780 --> 00:40:12.930
Of course they remember recent
things, 1971, better than

00:40:12.930 --> 00:40:14.840
stuff from years ago.

00:40:14.840 --> 00:40:16.060
We all do.

00:40:16.060 --> 00:40:19.000
But the striking thing is that
when these patients got ECT,

00:40:19.000 --> 00:40:22.610
during the course of their
amnesia they only lost memory

00:40:22.610 --> 00:40:24.940
for TV shows from the
last two years.

00:40:24.940 --> 00:40:26.010
They didn't lose
those memories.

00:40:26.010 --> 00:40:27.130
These are the very
same patients.

00:40:27.130 --> 00:40:30.780
The very same patients before
they got ECT and while they're

00:40:30.780 --> 00:40:35.090
getting ECT, they lose memory
only for the last two years of

00:40:35.090 --> 00:40:36.690
TV shows and not all
the other ones.

00:40:36.690 --> 00:40:39.850
Or the same thing you can do
with famous public events.

00:40:39.850 --> 00:40:41.440
So this is what people call a

00:40:41.440 --> 00:40:43.425
temporally-limited retrograde amnesia.

00:40:43.425 --> 00:40:46.640
It goes back for some years, the
knowledge you already had.

00:40:46.640 --> 00:40:48.550
And here's the amazing thing.

00:40:48.550 --> 00:40:50.560
When these patients go home,
their memory largely comes

00:40:50.560 --> 00:40:53.410
back and so does their knowledge
of the TV shows.

00:40:53.410 --> 00:40:55.030
It's as if it's in
their brain.

00:40:55.030 --> 00:40:57.580
They don't have access to it,
for the last two years of

00:40:57.580 --> 00:40:58.540
information.

00:40:58.540 --> 00:41:03.290
But it's still sitting there and
they regain that access.

00:41:03.290 --> 00:41:05.400
Two more examples on this.

00:41:05.400 --> 00:41:07.110
So here's an example--

00:41:07.110 --> 00:41:08.620
because with humans
it's always tough.

00:41:08.620 --> 00:41:10.120
So they do the same thing
with monkeys.

00:41:10.120 --> 00:41:13.180
With monkeys, they could create
a surgical lesion like

00:41:13.180 --> 00:41:15.230
HM and train them on material.

00:41:15.230 --> 00:41:17.070
They knew exactly what
they learned.

00:41:17.070 --> 00:41:19.940
Here they had the surgery
either 2, 4, 8 or

00:41:19.940 --> 00:41:21.940
12, or 8, 16 weeks.

00:41:21.940 --> 00:41:24.630
And you can see the monkeys with
the hippocampal damage

00:41:24.630 --> 00:41:28.090
were performing poorly if they
learned the material 2 or 4

00:41:28.090 --> 00:41:33.400
weeks ago but not 8
or 16 weeks ago.

00:41:33.400 --> 00:41:36.540
Again, a temporally-limited
retrograde amnesia.

00:41:36.540 --> 00:41:39.000
So we're getting the idea that
the hippocampus is important,

00:41:39.000 --> 00:41:41.350
not only for forming
new memories--

00:41:41.350 --> 00:41:42.980
and this is kind of wild.

00:41:42.980 --> 00:41:45.590
It's still important for you to
remember stuff from a week

00:41:45.590 --> 00:41:48.280
ago, a month ago, and
a couple years ago.

00:41:48.280 --> 00:41:52.060
And it takes years for your
memory to become independent,

00:41:52.060 --> 00:41:54.510
or what people call
consolidated, so that it no

00:41:54.510 --> 00:41:58.280
longer depends upon
the hippocampus.

00:41:58.280 --> 00:42:00.960
So the hippocampus both is
required to form a new memory,

00:42:00.960 --> 00:42:03.070
and it seems like it's necessary
to remember that

00:42:03.070 --> 00:42:06.810
memory from months to years,
before that memory becomes

00:42:06.810 --> 00:42:08.380
independent of the
hippocampus.

00:42:08.380 --> 00:42:12.770
So right now in you, it's as
if you had a neurochemistry

00:42:12.770 --> 00:42:17.960
team finishing your memories
from the end of eighth grade.

00:42:17.960 --> 00:42:18.600
We're done.

00:42:18.600 --> 00:42:20.970
We're moving on to
ninth grade!

00:42:20.970 --> 00:42:24.340
Because it takes years for your
hippocampus to no longer

00:42:24.340 --> 00:42:27.480
be required to get those
memories that you

00:42:27.480 --> 00:42:29.010
acquired years ago.

00:42:29.010 --> 00:42:31.070
And the same thing-- they did
these famous faces, like the

00:42:31.070 --> 00:42:32.100
ones I showed you before?

00:42:32.100 --> 00:42:33.960
So these had to be somewhat
older people.

00:42:33.960 --> 00:42:37.640
Here's activation in the 1990s,
if they were looking at

00:42:37.640 --> 00:42:39.880
famous faces and the experiment
was done in the

00:42:39.880 --> 00:42:40.670
late 1990s.

00:42:40.670 --> 00:42:44.110
So recent faces, hippocampus
turned on, but not for '80s,

00:42:44.110 --> 00:42:46.040
'70s, or '60s, or
'50s, or '40s.

00:42:46.040 --> 00:42:48.810
Again, they said the hippocampus
is required to

00:42:48.810 --> 00:42:51.260
retrieve memories that's in
the rest of your brain for

00:42:51.260 --> 00:42:54.400
some time period, but then after
much time passes, it's

00:42:54.400 --> 00:42:55.670
no longer required.

00:42:55.670 --> 00:42:57.630
So this just summarizes
those things.

00:42:57.630 --> 00:42:59.560
So now, Tyler, if we
could do the film.

00:42:59.560 --> 00:43:01.170
We're going to how you
an amnesic patient.

00:43:01.170 --> 00:43:04.460
Most amnesic patients
are like this man.

00:43:04.460 --> 00:43:06.570
Very intelligent, very verbal.

00:43:06.570 --> 00:43:10.280
Not dramatic, but with terribly
impaired memory.

00:43:21.100 --> 00:43:24.510
So last part of the talk is this
idea of memory systems

00:43:24.510 --> 00:43:27.250
and procedural memory, the idea
that you're a symphony of

00:43:27.250 --> 00:43:28.880
the neural systems that learn
different things.

00:43:28.880 --> 00:43:31.620
So your brain is like the
university, right?

00:43:31.620 --> 00:43:33.800
Department of Chemistry,
department of Biology,

00:43:33.800 --> 00:43:34.570
department of Brain and

00:43:34.570 --> 00:43:36.130
Cognitive Science, or Economics.

00:43:36.130 --> 00:43:39.770
In you are multiple very
specialized learning circuits

00:43:39.770 --> 00:43:41.530
that are good for learning
different things.

00:43:41.530 --> 00:43:43.640
So this idea of a memory system,
a particular part of

00:43:43.640 --> 00:43:47.690
the brain that has a particular
learning process.

00:43:47.690 --> 00:43:50.180
And a key concept in that is,
so far we've been talking

00:43:50.180 --> 00:43:52.340
about what people would
normally call memory--

00:43:52.340 --> 00:43:53.710
explicit or direct testing.

00:43:53.710 --> 00:43:55.530
You know, what did you study?

00:43:55.530 --> 00:43:56.200
What year is it?

00:43:56.200 --> 00:43:58.060
You're asked directly
to do something.

00:43:58.060 --> 00:44:01.310
But another way to show memory
and learning is through

00:44:01.310 --> 00:44:05.220
changes in performance
with practice.

00:44:05.220 --> 00:44:07.460
So people will sometimes call
the first kind declarative

00:44:07.460 --> 00:44:10.200
memory, where you directly have
conscious memory about

00:44:10.200 --> 00:44:11.190
facts and episodes.

00:44:11.190 --> 00:44:13.260
You know that you know
something, or

00:44:13.260 --> 00:44:14.890
that something occurred.

00:44:14.890 --> 00:44:16.680
And then the other kind, we'll
talk about now, procedural

00:44:16.680 --> 00:44:19.350
memory, accessible only through
performance and

00:44:19.350 --> 00:44:21.420
knowing how.

00:44:21.420 --> 00:44:23.480
Now with these kinds of skill
learning or procedural memory,

00:44:23.480 --> 00:44:25.900
we see how people get better
when they practice.

00:44:25.900 --> 00:44:28.830
Getting better is a change in
your behavior that's learned

00:44:28.830 --> 00:44:30.085
and a change in your brain
that performs.

00:44:30.085 --> 00:44:32.010
This is for your performance.

00:44:32.010 --> 00:44:34.710
So here's the experiment that
was the original one Brenda

00:44:34.710 --> 00:44:36.970
Milner did with HM,
where you have to

00:44:36.970 --> 00:44:38.670
trace a star in a mirror.

00:44:38.670 --> 00:44:41.230
You see your hand in a mirror,
but you don't have direct view

00:44:41.230 --> 00:44:42.000
of your hand.

00:44:42.000 --> 00:44:44.125
Have any of you done something
like this for some odds and

00:44:44.125 --> 00:44:45.060
ends reason?

00:44:45.060 --> 00:44:45.440
Anybody?

00:44:45.440 --> 00:44:47.480
Sometimes--

00:44:47.480 --> 00:44:49.350
do I see a hand?

00:44:49.350 --> 00:44:49.750
No.

00:44:49.750 --> 00:44:50.750
OK.

00:44:50.750 --> 00:44:51.940
It's surprisingly hard.

00:44:51.940 --> 00:44:53.830
The horizontals and verticals
are pretty easy.

00:44:53.830 --> 00:44:55.970
The diagonals, as you
move your hand, you

00:44:55.970 --> 00:44:56.920
go, oh, it's no problem.

00:44:56.920 --> 00:44:58.730
I just reverse everything.

00:44:58.730 --> 00:45:01.540
And you see your hand just
drift the wrong way.

00:45:01.540 --> 00:45:03.740
And you go, OK, no,
no, no, no.

00:45:03.740 --> 00:45:06.350
I just have to go 45
degrees difference.

00:45:06.350 --> 00:45:07.080
I can do it.

00:45:07.080 --> 00:45:09.340
And you're controlling your
hand, but then you move again

00:45:09.340 --> 00:45:10.420
and it still goes
the wrong way.

00:45:10.420 --> 00:45:12.970
And so at first, you make
lots of mistakes.

00:45:12.970 --> 00:45:16.550
We understand this because you
have such automaticity between

00:45:16.550 --> 00:45:19.030
what you see and how you move
all the time, that the mirror

00:45:19.030 --> 00:45:21.860
reverse, you have to overcome
that automatic relationship.

00:45:21.860 --> 00:45:24.410
So at first you make a lot of
mistakes and move slowly.

00:45:24.410 --> 00:45:26.000
You practice, and
you get better.

00:45:26.000 --> 00:45:28.800
Here's the remarkable thing
with-- so here's HM.

00:45:28.800 --> 00:45:32.700
Each of these dots is the next
time he does it consecutively.

00:45:32.700 --> 00:45:35.580
He gets better from doing
it over and over again.

00:45:35.580 --> 00:45:36.390
That's not a shock.

00:45:36.390 --> 00:45:36.850
Here's the shock.

00:45:36.850 --> 00:45:39.565
He comes the next day, he keeps
the learning he had and

00:45:39.565 --> 00:45:40.150
he gets better.

00:45:40.150 --> 00:45:42.510
He comes in day three,
he keeps learning.

00:45:42.510 --> 00:45:46.180
Now for him, a day is like
infinity, for memory.

00:45:46.180 --> 00:45:48.270
When he comes in the room, he
says, what do I do with this?

00:45:48.270 --> 00:45:49.800
You'd go, well, we did
it the last two days.

00:45:49.800 --> 00:45:50.230
Don't you know?

00:45:50.230 --> 00:45:52.050
He said, no, I have no idea.

00:45:52.050 --> 00:45:53.050
Have you done this before?

00:45:53.050 --> 00:45:53.540
No.

00:45:53.540 --> 00:45:55.310
You can give him multiple choice
of different things he

00:45:55.310 --> 00:45:55.770
might have done.

00:45:55.770 --> 00:45:57.050
He'll pick the wrong thing.

00:45:57.050 --> 00:45:59.600
When you come in on day three
and you're this good, you're

00:45:59.600 --> 00:46:00.280
pretty proud.

00:46:00.280 --> 00:46:01.220
In fact, vain.

00:46:01.220 --> 00:46:02.180
You're going, like,
I can do this.

00:46:02.180 --> 00:46:02.970
I'm awesome.

00:46:02.970 --> 00:46:04.560
Just let me at it.

00:46:04.560 --> 00:46:06.110
Because you know that
you did it.

00:46:06.110 --> 00:46:07.890
You know how good you are
through practice.

00:46:07.890 --> 00:46:11.180
He has no idea, but his learning
is entirely normal.

00:46:11.180 --> 00:46:13.290
Not just better than
you'd think?

00:46:13.290 --> 00:46:15.240
Just as good as you.

00:46:15.240 --> 00:46:18.610
So a memory system in his brain,
different than the

00:46:18.610 --> 00:46:22.600
hippocampus, has kept that
memory, uses that memory.

00:46:22.600 --> 00:46:24.670
And the amazing thing is that
this means that that's also

00:46:24.670 --> 00:46:27.090
true in you, as far as we
understand these things.

00:46:27.090 --> 00:46:30.570
When you learn a physical or
mental skill, a lot of it is

00:46:30.570 --> 00:46:34.690
learned in a way that has
nothing to do with your memory

00:46:34.690 --> 00:46:35.800
that you learned the skill.

00:46:35.800 --> 00:46:37.560
It's all by doing.

00:46:37.560 --> 00:46:39.470
And it doesn't depend
on the hippocampus.

00:46:39.470 --> 00:46:42.730
And some years later, I did
this again a little bit.

00:46:42.730 --> 00:46:46.390
Here's his performance first,
second, third day, a week

00:46:46.390 --> 00:46:48.530
later, two weeks later,
and a year later.

00:46:48.530 --> 00:46:52.580
Complete retention for a man who
can forget within moments.

00:46:52.580 --> 00:46:54.820
A year later, a skill.

00:46:54.820 --> 00:46:57.770
He forgets that he did it
ever within moments.

00:46:57.770 --> 00:47:01.215
So another kind of task that
people do-- same idea.

00:47:01.215 --> 00:47:03.710
A motor skill you can lose, it's
called rotary pursuit.

00:47:03.710 --> 00:47:06.080
There's a disc like this that
revolves around a platter that

00:47:06.080 --> 00:47:07.600
turns pretty fast.

00:47:07.600 --> 00:47:10.360
Your job is to maintain
contact between--

00:47:10.360 --> 00:47:12.100
this was sort of sad before
video games, right?

00:47:12.100 --> 00:47:16.490
This is about as interesting
as 1940s games got.

00:47:16.490 --> 00:47:19.250
It was whizzed around, and it's
fast enough that it's not

00:47:19.250 --> 00:47:19.750
easy to do.

00:47:19.750 --> 00:47:21.550
And at first your hand goes off
it, but with practice you

00:47:21.550 --> 00:47:25.060
get really good at revolving
with a rotating disc.

00:47:25.060 --> 00:47:29.330
And HM and patients like him
learned that very well.

00:47:29.330 --> 00:47:30.460
Another motor skill.

00:47:30.460 --> 00:47:32.210
So who's bad at learning this?

00:47:32.210 --> 00:47:34.620
Which part of the brain is your
instrument for learning

00:47:34.620 --> 00:47:36.350
skills through practice?

00:47:36.350 --> 00:47:38.280
So our best evidence comes
from patients with a very

00:47:38.280 --> 00:47:40.090
difficult disease called
Huntington's disease.

00:47:40.090 --> 00:47:42.030
It's a genetic rare disease.

00:47:42.030 --> 00:47:44.190
Its onset is in the
30s or 40s.

00:47:44.190 --> 00:47:46.760
These patients get severe motor
problems and movement

00:47:46.760 --> 00:47:49.380
problems, and then, over
time, cognitive ones

00:47:49.380 --> 00:47:52.120
and psychiatric ones.

00:47:52.120 --> 00:47:54.470
There's no treatment for it.

00:47:54.470 --> 00:47:57.240
So this is the caudate withering
away in a patient

00:47:57.240 --> 00:47:59.140
with Huntington's disease
compared to a healthy

00:47:59.140 --> 00:48:00.670
person's, post mortem.

00:48:00.670 --> 00:48:05.240
Or if you look at an MRI, here's
your basal ganglia.

00:48:05.240 --> 00:48:07.850
Here's the withered away basal
ganglia in the patients with

00:48:07.850 --> 00:48:09.310
Huntington's disease.

00:48:09.310 --> 00:48:11.290
And there's a brief--

00:48:11.290 --> 00:48:14.920
can we do, Tyler-- there's a
brief video of a patient with

00:48:14.920 --> 00:48:16.280
early-stage Huntington's
disease.

00:48:26.740 --> 00:48:30.190
So it's not the most severe
problem they have, but when

00:48:30.190 --> 00:48:32.490
they do one of these motor
kinds of tasks--

00:48:32.490 --> 00:48:35.460
here's normal typical people
learning a motor task, getting

00:48:35.460 --> 00:48:36.020
better on it.

00:48:36.020 --> 00:48:37.640
It's good to be high
on this graph.

00:48:37.640 --> 00:48:39.285
Here's the patients with
Huntington's disease showing

00:48:39.285 --> 00:48:40.060
no learning at all.

00:48:40.060 --> 00:48:41.460
And this is a test
that an amnesic

00:48:41.460 --> 00:48:43.220
patient would learn normally.

00:48:43.220 --> 00:48:45.160
So very convincing-- and you
could say, well, if they have

00:48:45.160 --> 00:48:47.620
a motor problem, they're not
going to get a motor skill.

00:48:47.620 --> 00:48:51.700
So one thing they do is they
make the platter turn very

00:48:51.700 --> 00:48:54.580
slowly, so at the beginning
they're doing just as well.

00:48:54.580 --> 00:48:56.440
And they still show no
learning at all.

00:48:56.440 --> 00:48:59.330
So there's a lot of other
evidence sparked by this that

00:48:59.330 --> 00:49:02.320
just shows that the basal
ganglia is an instrument

00:49:02.320 --> 00:49:06.000
that's essential for us to
learn physical or motor

00:49:06.000 --> 00:49:08.220
skills, perceptual skills,
and a variety of

00:49:08.220 --> 00:49:09.740
cognitive skills as well.

00:49:09.740 --> 00:49:11.820
Things where we practice,
practice, practice to become

00:49:11.820 --> 00:49:12.670
excellent--

00:49:12.670 --> 00:49:16.080
basal ganglia is doing a lot
of that work for you.

00:49:16.080 --> 00:49:18.270
The last kind of learning I'll
talk to you about, and the

00:49:18.270 --> 00:49:21.920
last disease, is a sort of odder
form of learning called

00:49:21.920 --> 00:49:23.760
repetition priming.

00:49:23.760 --> 00:49:26.520
It's a change in performance
because of

00:49:26.520 --> 00:49:28.160
something you did recently.

00:49:28.160 --> 00:49:30.610
And if I give you a concrete
example, it'll feel better.

00:49:30.610 --> 00:49:34.230
So here's one of
many examples.

00:49:34.230 --> 00:49:37.050
Imagine you study a list of
words like stamp, landmark,

00:49:37.050 --> 00:49:37.760
speak, clock.

00:49:37.760 --> 00:49:39.030
You study a list of
words like these.

00:49:39.030 --> 00:49:41.390
I can give you an explicit or
declarative memory test.

00:49:41.390 --> 00:49:42.630
What words did you see?

00:49:42.630 --> 00:49:43.430
You'd recall them.

00:49:43.430 --> 00:49:45.960
Or which word did you see, that
we'd do multiple choice.

00:49:45.960 --> 00:49:48.420
That's regular memory that
we know depends on the

00:49:48.420 --> 00:49:49.710
hippocampus.

00:49:49.710 --> 00:49:52.370
I can also give you weird test
like this, which says, tell me

00:49:52.370 --> 00:49:55.210
the first word you think of
that starts with STA.

00:49:55.210 --> 00:49:57.000
Now if you have good memory,
you'll go, do you want me to

00:49:57.000 --> 00:49:58.270
give you stamp?

00:49:58.270 --> 00:49:59.050
Is that the point?

00:49:59.050 --> 00:49:59.940
And you go, no, no, no.

00:49:59.940 --> 00:50:01.380
And they go, oh, it's
a Freudian thing.

00:50:01.380 --> 00:50:03.480
You want to know some weird
things I'm gonna think of.

00:50:03.480 --> 00:50:04.340
And you go, no, no, no.

00:50:04.340 --> 00:50:06.630
Just tell me the first word
you think of, OK?

00:50:06.630 --> 00:50:08.850
And you can give many answers
that are all perfectly fine.

00:50:08.850 --> 00:50:11.950
You can give stall, stand,
staccato, star, many STA

00:50:11.950 --> 00:50:14.070
words, and all of
them are fine.

00:50:14.070 --> 00:50:15.090
That's all you have to do.

00:50:15.090 --> 00:50:18.360
But if you saw stamp, one
particular completion of this,

00:50:18.360 --> 00:50:22.240
about 10 minutes ago, you're
biased or primed to come up

00:50:22.240 --> 00:50:22.790
with that word.

00:50:22.790 --> 00:50:25.760
Not all the time, but much
more than by chance.

00:50:25.760 --> 00:50:29.840
A recent experience biases you
to behave in a certain way.

00:50:29.840 --> 00:50:30.990
And here's the remarkable
thing.

00:50:30.990 --> 00:50:35.230
When you do this experiment with
HM, if you just had the

00:50:35.230 --> 00:50:37.150
list of words and you
wait about 20

00:50:37.150 --> 00:50:38.660
seconds, here's your recall.

00:50:38.660 --> 00:50:40.970
Now you're not surprised that
recall from a healthy person

00:50:40.970 --> 00:50:42.190
is about seven words, right?

00:50:42.190 --> 00:50:43.280
You tried that yourself.

00:50:43.280 --> 00:50:45.350
HM, sadly, has no score
because he always

00:50:45.350 --> 00:50:46.840
goes, like, what list?

00:50:46.840 --> 00:50:48.050
So no recall.

00:50:48.050 --> 00:50:51.670
Multiple choice recognition,
where there are three choices,

00:50:51.670 --> 00:50:55.000
here's typical people,
HM about chance.

00:50:55.000 --> 00:50:57.270
The probability above chance--

00:50:57.270 --> 00:51:00.100
that you give stamp as
the completion to

00:51:00.100 --> 00:51:02.160
STA if you saw stamp--

00:51:02.160 --> 00:51:04.550
if I showed you star, probably
you'd give star.

00:51:04.550 --> 00:51:07.240
The specific word you
saw 10 minutes ago?

00:51:07.240 --> 00:51:10.660
Perfectly the same in HM as
normal control subjects.

00:51:10.660 --> 00:51:14.550
He was just as influenced by
that word he saw 10 minutes

00:51:14.550 --> 00:51:18.310
ago as you are, to give
you that answer.

00:51:18.310 --> 00:51:20.710
So this got people very excited,
because they said,

00:51:20.710 --> 00:51:23.370
now we can experimentally and
scientifically study something

00:51:23.370 --> 00:51:27.640
that Freud talked about as a
very vague idea, the cognitive

00:51:27.640 --> 00:51:28.350
unconscious.

00:51:28.350 --> 00:51:31.760
He's not conscious that he saw
a list of words 10 minutes

00:51:31.760 --> 00:51:34.700
ago, but it's making him
choose to behave in

00:51:34.700 --> 00:51:36.870
a certain way now.

00:51:36.870 --> 00:51:38.700
Now you have both things
in you, as far as

00:51:38.700 --> 00:51:39.670
we understand it.

00:51:39.670 --> 00:51:42.130
You can both remember you saw
the word list, but if you

00:51:42.130 --> 00:51:44.340
switch your thinking a little
bit, your behavior a little

00:51:44.340 --> 00:51:48.210
bit, the unconscious system gets
into the driver's seat.

00:51:48.210 --> 00:51:52.070
Because the people with good
memory were giving stamp often

00:51:52.070 --> 00:51:54.770
but no more or less than HM.

00:51:54.770 --> 00:51:55.950
So unconscious.

00:51:55.950 --> 00:51:59.930
So what part of the brain does
this kind of memory?

00:51:59.930 --> 00:52:02.860
And our insight to that comes
from Alzheimer's disease.

00:52:02.860 --> 00:52:08.390
So HM's kind of amnesia
is extremely rare.

00:52:08.390 --> 00:52:09.630
Huntington's is quite rare.

00:52:09.630 --> 00:52:12.200
Alzheimer's is tragically
common.

00:52:12.200 --> 00:52:15.510
In fact, they went to East
Boston, the area near Logan,

00:52:15.510 --> 00:52:17.780
and knocked on doors
of people's homes.

00:52:17.780 --> 00:52:18.850
Not people in nursing homes.

00:52:18.850 --> 00:52:19.670
At home.

00:52:19.670 --> 00:52:20.570
And they estimated--

00:52:20.570 --> 00:52:22.620
and there's debates about
this back and forth.

00:52:22.620 --> 00:52:28.000
But about half the people over
85 who answered the door

00:52:28.000 --> 00:52:32.340
qualify for diagnosis of
Alzheimer's disease.

00:52:32.340 --> 00:52:34.480
About half over 85.

00:52:34.480 --> 00:52:35.360
There's debates about this.

00:52:35.360 --> 00:52:37.200
There's debates about whether
all of us would get it if we

00:52:37.200 --> 00:52:39.350
lived long enough.

00:52:39.350 --> 00:52:43.790
But as American society gets
older, as us baby boomers all

00:52:43.790 --> 00:52:47.920
move into our '60s, '70s, and
'80s, there's going to be a

00:52:47.920 --> 00:52:50.800
heck of a lot of people around
with Alzheimer's disease.

00:52:50.800 --> 00:52:53.080
And some of you have probably
experienced this, with

00:52:53.080 --> 00:52:54.610
grandparents if nothing else.

00:52:54.610 --> 00:52:57.350
It's a very devastating disorder
for the patient, for

00:52:57.350 --> 00:52:58.325
the spouse or family.

00:52:58.325 --> 00:53:00.480
It's a very challenging
disorder.

00:53:00.480 --> 00:53:01.960
So what is Alzheimer's
disease?

00:53:01.960 --> 00:53:03.885
Well behaviorally,
it's an insidious

00:53:03.885 --> 00:53:05.180
and progressive dementia.

00:53:05.180 --> 00:53:09.070
Dementia means, unlike HM, it's
many losses of abilities.

00:53:09.070 --> 00:53:09.790
Unlike HM.

00:53:09.790 --> 00:53:11.470
HM had his surgery
one day, boom.

00:53:11.470 --> 00:53:12.510
That was it.

00:53:12.510 --> 00:53:14.430
Alzheimer's starts very
slowly and gets worse

00:53:14.430 --> 00:53:15.740
and worse over time.

00:53:15.740 --> 00:53:18.840
The biggest problem early on
is memory, the most common

00:53:18.840 --> 00:53:21.430
problem but over time,
language, thinking,

00:53:21.430 --> 00:53:24.060
concentration, spatial thinking,
sometimes mood and

00:53:24.060 --> 00:53:28.740
personality, all of these
get altered over time.

00:53:28.740 --> 00:53:31.430
In the brain when you look at
the changes, you see that the

00:53:31.430 --> 00:53:34.300
changes occur especially in the
hippocampal region-- so

00:53:34.300 --> 00:53:36.700
that's why the memory
problem is like HM--

00:53:36.700 --> 00:53:40.270
and also in some other parts
of the cortex, but not so

00:53:40.270 --> 00:53:42.960
much, for example, in basal
ganglia, that we just spoke

00:53:42.960 --> 00:53:46.000
about as important for
procedural memory.

00:53:46.000 --> 00:53:49.640
Here's a control brain and
equal-aged Alzheimer's brain.

00:53:49.640 --> 00:53:52.820
The reason why these gyri, this
whole side widened is

00:53:52.820 --> 00:53:57.740
because of widespread neuronal
shrinkage or death.

00:53:57.740 --> 00:54:00.540
And what's happening inside the
brain, when you look at

00:54:00.540 --> 00:54:03.480
post mortem, there's sort of
neurofibrillary tangles and

00:54:03.480 --> 00:54:05.740
plaques that are associated
with the diseases.

00:54:05.740 --> 00:54:07.800
The tangles are something
about the

00:54:07.800 --> 00:54:09.410
cells that have died.

00:54:09.410 --> 00:54:12.130
The plaques are thought to be
maybe part of the process that

00:54:12.130 --> 00:54:14.590
leads to the disorder, but
there's a debate about

00:54:14.590 --> 00:54:16.840
that to this day.

00:54:16.840 --> 00:54:20.100
And if you look, on post mortem,
where the tangles are,

00:54:20.100 --> 00:54:23.050
the tangles up here, you can see
they're most dense in the

00:54:23.050 --> 00:54:26.600
hippocampus or near the
hippocampus, like HM.

00:54:26.600 --> 00:54:29.030
And so that's why we think so
many patients with Alzheimer's

00:54:29.030 --> 00:54:31.830
have an HM-like memory
problem and then more

00:54:31.830 --> 00:54:33.710
problems after that.

00:54:33.710 --> 00:54:36.680
So here's a healthy person
top of the brain.

00:54:36.680 --> 00:54:40.330
Here's the left and right
hippocampus in an 81-year-old.

00:54:40.330 --> 00:54:43.090
And here in an 80-year-old
Alzheimer's patient, the

00:54:43.090 --> 00:54:45.905
hippocampus has withered away,
both sides of the brain,

00:54:45.905 --> 00:54:48.030
bilaterally.

00:54:48.030 --> 00:54:51.020
So one more brain perspective
on this.

00:54:51.020 --> 00:54:53.030
This is a SPECT scan, that just
shows you where there's

00:54:53.030 --> 00:54:54.920
blood flow and metabolism.

00:54:54.920 --> 00:54:56.480
Top of the brain, bottom
of the brain.

00:54:56.480 --> 00:54:59.500
Here's the insides of the
temporal lobes in a healthy

00:54:59.500 --> 00:55:01.290
elderly adult.

00:55:01.290 --> 00:55:02.320
Look how much worse--

00:55:02.320 --> 00:55:03.100
this is HM.

00:55:03.100 --> 00:55:03.950
He looks pretty good, right?

00:55:03.950 --> 00:55:06.145
You have to look really hard to
see that the hippocampus is

00:55:06.145 --> 00:55:08.180
missing because it's such
a small structure.

00:55:08.180 --> 00:55:10.830
And here's an early-stage
Alzheimer's patient.

00:55:10.830 --> 00:55:13.500
Lots of the tissue is there, but
a lot of-- you know, it's

00:55:13.500 --> 00:55:17.550
not functioning because there's
so much damage.

00:55:17.550 --> 00:55:20.080
Here's the medial temporal lobes
but also cortical areas

00:55:20.080 --> 00:55:21.330
that are compromised.

00:55:23.160 --> 00:55:23.690
OK, again.

00:55:23.690 --> 00:55:26.770
Not their biggest problem, but
in terms of thinking about

00:55:26.770 --> 00:55:28.020
different parts of the
brain and what parts

00:55:28.020 --> 00:55:30.430
of memory they do--

00:55:30.430 --> 00:55:32.720
So these are amnesic patients,
including HM, Alzheimer's

00:55:32.720 --> 00:55:34.250
patients, and green
in controls.

00:55:34.250 --> 00:55:36.340
So recall of the list, what
words were on the list,

00:55:36.340 --> 00:55:39.110
explicit memory bad in the
two patient groups.

00:55:39.110 --> 00:55:42.190
Recognition pretty poor.

00:55:42.190 --> 00:55:44.240
Word stem completion, the kind
of thing that we said is

00:55:44.240 --> 00:55:47.440
completely normal in
HM, that's impaired

00:55:47.440 --> 00:55:48.480
in Alzheimer's disease.

00:55:48.480 --> 00:55:50.730
So we think that kind of
priming depends on the

00:55:50.730 --> 00:55:54.880
neocortex, the part of the brain
that's also injured in

00:55:54.880 --> 00:55:55.490
Alzheimer's.

00:55:55.490 --> 00:55:58.600
Because we know it doesn't
depend on the hippocampus,

00:55:58.600 --> 00:55:59.220
because the patients with

00:55:59.220 --> 00:56:01.900
hippocampus-only damage are fine.

00:56:01.900 --> 00:56:05.120
So we end up with a picture
like this of something on

00:56:05.120 --> 00:56:05.830
memory systems.

00:56:05.830 --> 00:56:08.680
For explicit or declarative
memory, from patients with

00:56:08.680 --> 00:56:11.330
amnesia, we know it depends on
the medial temporal lobe.

00:56:11.330 --> 00:56:13.790
Left, verbal; right, spatial.

00:56:13.790 --> 00:56:16.510
For skill learning, anything you
practice many times to get

00:56:16.510 --> 00:56:19.880
better at, we know from
Huntington's disease, that

00:56:19.880 --> 00:56:21.950
depends on the basal ganglia.

00:56:21.950 --> 00:56:25.080
For kinds of priming, recent
experience altering something

00:56:25.080 --> 00:56:27.880
about brain organization,
Alzheimer's disease, we know

00:56:27.880 --> 00:56:31.320
from that that it's likely
mediated by the neocortex.

00:56:31.320 --> 00:56:32.740
These are from patient studies,
but the imaging

00:56:32.740 --> 00:56:34.730
studies are very well aligned
with these things.

00:56:34.730 --> 00:56:38.050
So we think it's true not only
of patients, but of you, in

00:56:38.050 --> 00:56:40.550
terms of the instruments
of your brain.

00:56:40.550 --> 00:56:42.935
So now we have not only
episodic memory

00:56:42.935 --> 00:56:43.820
and semantic memory.

00:56:43.820 --> 00:56:44.910
We talked about procedural
memory.

00:56:44.910 --> 00:56:46.040
We talked about priming.

00:56:46.040 --> 00:56:47.520
I didn't talk about conditioning
today, but we've

00:56:47.520 --> 00:56:50.240
talked about that before.

00:56:50.240 --> 00:56:53.290
So I have two more slides
to add on conceptually.

00:56:53.290 --> 00:56:56.160
So in what sense was Karl Ashley
right or wrong when he

00:56:56.160 --> 00:56:57.450
said memory's all
over the brain?

00:56:57.450 --> 00:56:59.240
Here's how we think
about it now.

00:56:59.240 --> 00:57:02.230
We think he was right that all
of your brain is learning.

00:57:02.230 --> 00:57:04.670
Practically all of your
brain is plastic.

00:57:04.670 --> 00:57:06.090
It changes with experience.

00:57:06.090 --> 00:57:07.960
But different parts
of your brain are

00:57:07.960 --> 00:57:08.880
learning different things.

00:57:08.880 --> 00:57:10.130
The basal ganglia is
learning how to

00:57:10.130 --> 00:57:11.680
be skilled at something.

00:57:11.680 --> 00:57:14.410
The neocortex, how to gain
knowledge and shift your

00:57:14.410 --> 00:57:17.100
representation of knowledge,
like in priming.

00:57:17.100 --> 00:57:19.510
The medial temporal lobe on the
left is learning verbal

00:57:19.510 --> 00:57:21.590
facts, and the right
is nonverbal

00:57:21.590 --> 00:57:22.980
facts or spatial facts.

00:57:22.980 --> 00:57:25.400
So all of your brain is
learning, but it's learning

00:57:25.400 --> 00:57:26.430
different things.

00:57:26.430 --> 00:57:28.895
And in that sense we end up with
this metaphor that-- you

00:57:28.895 --> 00:57:31.430
know, just like there's
a string section and a

00:57:31.430 --> 00:57:34.320
percussion section or something
like that, in you,

00:57:34.320 --> 00:57:36.290
there's all these different
instruments that are highly

00:57:36.290 --> 00:57:39.470
tuned in the way they're
organized to learn all the

00:57:39.470 --> 00:57:40.610
different things you learned.

00:57:40.610 --> 00:57:43.650
They're all pretty useful
instruments to have.

00:57:43.650 --> 00:57:45.870
And without these patients and
the imaging studies, we

00:57:45.870 --> 00:57:46.760
wouldn't have known that.

00:57:46.760 --> 00:57:47.950
You would be like the Martian.

00:57:47.950 --> 00:57:49.440
One instrument?

00:57:49.440 --> 00:57:50.300
A thousand instruments?

00:57:50.300 --> 00:57:52.990
Well, we know there's multiple
ones, and we know pretty much

00:57:52.990 --> 00:57:54.560
which structures are
essential for them.

00:57:54.560 --> 00:57:55.810
Thanks.