When you're making a deal, what's going on in your brain?
-
0:01 - 0:03I'm going to talk about the strategizing brain.
-
0:03 - 0:05We're going to use an unusual combination of tools
-
0:05 - 0:07from game theory and neuroscience
-
0:07 - 0:10to understand how people interact socially when value is on the line.
-
0:10 - 0:14So game theory is a branch of, originally, applied mathematics,
-
0:14 - 0:17used mostly in economics and political science, a little bit in biology,
-
0:17 - 0:20that gives us a mathematical taxonomy of social life
-
0:20 - 0:22and it predicts what people are likely to do
-
0:22 - 0:24and believe others will do
-
0:24 - 0:27in cases where everyone's actions affect everyone else.
-
0:27 - 0:30That's a lot of things: competition, cooperation, bargaining,
-
0:30 - 0:34games like hide-and-seek, and poker.
-
0:34 - 0:36Here's a simple game to get us started.
-
0:36 - 0:38Everyone chooses a number from zero to 100,
-
0:38 - 0:41we're going to compute the average of those numbers,
-
0:41 - 0:45and whoever's closest to two-thirds of the average wins a fixed prize.
-
0:45 - 0:47So you want to be a little bit below the average number,
-
0:47 - 0:49but not too far below, and everyone else wants to be
-
0:49 - 0:51a little bit below the average number as well.
-
0:51 - 0:54Think about what you might pick.
-
0:54 - 0:57As you're thinking, this is a toy model of something like
-
0:57 - 1:00selling in the stock market during a rising market. Right?
-
1:00 - 1:02You don't want to sell too early, because you miss out on profits,
-
1:02 - 1:04but you don't want to wait too late
-
1:04 - 1:07to when everyone else sells, triggering a crash.
-
1:07 - 1:10You want to be a little bit ahead of the competition, but not too far ahead.
-
1:10 - 1:13Okay, here's two theories about how people might think about this,
-
1:13 - 1:15and then we'll see some data.
-
1:15 - 1:17Some of these will sound familiar because you probably are
-
1:17 - 1:21thinking that way. I'm using my brain theory to see.
-
1:21 - 1:24A lot of people say, "I really don't know what people are going to pick,
-
1:24 - 1:25so I think the average will be 50."
-
1:25 - 1:27They're not being really strategic at all.
-
1:27 - 1:31"And I'll pick two-thirds of 50. That's 33." That's a start.
-
1:31 - 1:33Other people who are a little more sophisticated,
-
1:33 - 1:34using more working memory,
-
1:34 - 1:38say, "I think people will pick 33 because they're going to pick a response to 50,
-
1:38 - 1:41and so I'll pick 22, which is two-thirds of 33."
-
1:41 - 1:43They're doing one extra step of thinking, two steps.
-
1:43 - 1:46That's better. And of course, in principle,
-
1:46 - 1:48you could do three, four or more,
-
1:48 - 1:50but it starts to get very difficult.
-
1:50 - 1:52Just like in language and other domains, we know that it's hard for people to parse
-
1:52 - 1:56very complex sentences with a kind of recursive structure.
-
1:56 - 1:58This is called a cognitive hierarchy theory, by the way.
-
1:58 - 2:00It's something that I've worked on and a few other people,
-
2:00 - 2:02and it indicates a kind of hierarchy along with
-
2:02 - 2:05some assumptions about how many people stop at different steps
-
2:05 - 2:07and how the steps of thinking are affected
-
2:07 - 2:10by lots of interesting variables and variant people, as we'll see in a minute.
-
2:10 - 2:14A very different theory, a much more popular one, and an older one,
-
2:14 - 2:17due largely to John Nash of "A Beautiful Mind" fame,
-
2:17 - 2:19is what's called equilibrium analysis.
-
2:19 - 2:22So if you've ever taken a game theory course at any level,
-
2:22 - 2:24you will have learned a little bit about this.
-
2:24 - 2:26An equilibrium is a mathematical state in which everybody
-
2:26 - 2:29has figured out exactly what everyone else will do.
-
2:29 - 2:31It is a very useful concept, but behaviorally,
-
2:31 - 2:33it may not exactly explain what people do
-
2:33 - 2:36the first time they play these types of economic games
-
2:36 - 2:38or in situations in the outside world.
-
2:38 - 2:40In this case, the equilibrium makes a very bold prediction,
-
2:40 - 2:43which is everyone wants to be below everyone else,
-
2:43 - 2:45therefore they'll play zero.
-
2:45 - 2:48Let's see what happens. This experiment's been done many, many times.
-
2:48 - 2:50Some of the earliest ones were done in the '90s
-
2:50 - 2:53by me and Rosemarie Nagel and others.
-
2:53 - 2:56This is a beautiful data set of 9,000 people who wrote in
-
2:56 - 2:59to three newspapers and magazines that had a contest.
-
2:59 - 3:01The contest said, send in your numbers
-
3:01 - 3:04and whoever is close to two-thirds of the average will win a big prize.
-
3:04 - 3:07And as you can see, there's so much data here, you can see the spikes very visibly.
-
3:07 - 3:10There's a spike at 33. Those are people doing one step.
-
3:10 - 3:13There is another spike visible at 22.
-
3:13 - 3:15And notice, by the way, that most people pick numbers right around there.
-
3:15 - 3:18They don't necessarily pick exactly 33 and 22.
-
3:18 - 3:20There's something a little bit noisy around it.
-
3:20 - 3:21But you can see those spikes, and they're there.
-
3:21 - 3:23There's another group of people who seem to have
-
3:23 - 3:25a firm grip on equilibrium analysis,
-
3:25 - 3:27because they're picking zero or one.
-
3:27 - 3:29But they lose, right?
-
3:29 - 3:33Because picking a number that low is actually a bad choice
-
3:33 - 3:35if other people aren't doing equilibrium analysis as well.
-
3:35 - 3:38So they're smart, but poor.
-
3:38 - 3:40(Laughter)
-
3:40 - 3:42Where are these things happening in the brain?
-
3:42 - 3:45One study by Coricelli and Nagel gives a really sharp, interesting answer.
-
3:45 - 3:47So they had people play this game
-
3:47 - 3:49while they were being scanned in an fMRI,
-
3:49 - 3:51and two conditions: in some trials,
-
3:51 - 3:53they're told you're playing another person
-
3:53 - 3:55who's playing right now and we're going to match up
-
3:55 - 3:57your behavior at the end and pay you if you win.
-
3:57 - 3:59In the other trials, they're told, you're playing a computer.
-
3:59 - 4:00They're just choosing randomly.
-
4:00 - 4:02So what you see here is a subtraction
-
4:02 - 4:05of areas in which there's more brain activity
-
4:05 - 4:08when you're playing people compared to playing the computer.
-
4:08 - 4:10And you see activity in some regions we've seen today,
-
4:10 - 4:13medial prefrontal cortex, dorsomedial, however, up here,
-
4:13 - 4:15ventromedial prefrontal cortex,
-
4:15 - 4:17anterior cingulate, an area that's involved
-
4:17 - 4:20in lots of types of conflict resolution, like if you're playing "Simon Says,"
-
4:20 - 4:24and also the right and left temporoparietal junction.
-
4:24 - 4:27And these are all areas which are fairly reliably known
-
4:27 - 4:29to be part of what's called a "theory of mind" circuit,
-
4:29 - 4:31or "mentalizing circuit."
-
4:31 - 4:34That is, it's a circuit that's used to imagine what other people might do.
-
4:34 - 4:36So these were some of the first studies to see this
-
4:36 - 4:38tied in to game theory.
-
4:38 - 4:41What happens with these one- and two-step types?
-
4:41 - 4:43So we classify people by what they picked,
-
4:43 - 4:44and then we look at the difference between
-
4:44 - 4:46playing humans versus playing computers,
-
4:46 - 4:48which brain areas are differentially active.
-
4:48 - 4:50On the top you see the one-step players.
-
4:50 - 4:52There's almost no difference.
-
4:52 - 4:54The reason is, they're treating other people like a computer, and the brain is too.
-
4:54 - 4:59The bottom players, you see all the activity in dorsomedial PFC.
-
4:59 - 5:01So we know that those two-step players are doing something differently.
-
5:01 - 5:04Now if you were to step back and say, "What can we do with this information?"
-
5:04 - 5:06you might be able to look at brain activity and say,
-
5:06 - 5:07"This person's going to be a good poker player,"
-
5:07 - 5:09or, "This person's socially naive,"
-
5:09 - 5:10and we might also be able to study things
-
5:10 - 5:12like development of adolescent brains
-
5:12 - 5:15once we have an idea of where this circuitry exists.
-
5:15 - 5:18Okay. Get ready.
-
5:18 - 5:20I'm saving you some brain activity,
-
5:20 - 5:23because you don't need to use your hair detector cells.
-
5:23 - 5:26You should use those cells to think carefully about this game.
-
5:26 - 5:28This is a bargaining game.
-
5:28 - 5:30Two players who are being scanned using EEG electrodes
-
5:30 - 5:33are going to bargain over one to six dollars.
-
5:33 - 5:36If they can do it in 10 seconds, they're going to actually earn that money.
-
5:36 - 5:39If 10 seconds goes by and they haven't made a deal, they get nothing.
-
5:39 - 5:40That's kind of a mistake together.
-
5:40 - 5:43The twist is that one player, on the left,
-
5:43 - 5:46is informed about how much on each trial there is.
-
5:46 - 5:48They play lots of trials with different amounts each time.
-
5:48 - 5:50In this case, they know there's four dollars.
-
5:50 - 5:52The uninformed player doesn't know,
-
5:52 - 5:54but they know that the informed player knows.
-
5:54 - 5:56So the uninformed player's challenge is to say,
-
5:56 - 5:58"Is this guy really being fair
-
5:58 - 6:00or are they giving me a very low offer
-
6:00 - 6:03in order to get me to think that there's only one or two dollars available to split?"
-
6:03 - 6:06in which case they might reject it and not come to a deal.
-
6:06 - 6:09So there's some tension here between trying to get the most money
-
6:09 - 6:11but trying to goad the other player into giving you more.
-
6:11 - 6:14And the way they bargain is to point on a number line
-
6:14 - 6:16that goes from zero to six dollars,
-
6:16 - 6:19and they're bargaining over how much the uninformed player gets,
-
6:19 - 6:20and the informed player's going to get the rest.
-
6:20 - 6:23So this is like a management-labor negotiation
-
6:23 - 6:25in which the workers don't know how much profits
-
6:25 - 6:28the privately held company has, right,
-
6:28 - 6:30and they want to maybe hold out for more money,
-
6:30 - 6:32but the company might want to create the impression
-
6:32 - 6:35that there's very little to split: "I'm giving you the most that I can."
-
6:35 - 6:39First some behavior. So a bunch of the subject pairs, they play face to face.
-
6:39 - 6:41We have some other data where they play across computers.
-
6:41 - 6:43That's an interesting difference, as you might imagine.
-
6:43 - 6:45But a bunch of the face-to-face pairs
-
6:45 - 6:49agree to divide the money evenly every single time.
-
6:49 - 6:52Boring. It's just not interesting neurally.
-
6:52 - 6:54It's good for them. They make a lot of money.
-
6:54 - 6:57But we're interested in, can we say something about
-
6:57 - 7:00when disagreements occur versus don't occur?
-
7:00 - 7:02So this is the other group of subjects who often disagree.
-
7:02 - 7:05So they have a chance of -- they bicker and disagree
-
7:05 - 7:06and end up with less money.
-
7:06 - 7:10They might be eligible to be on "Real Housewives," the TV show.
-
7:10 - 7:12You see on the left,
-
7:12 - 7:15when the amount to divide is one, two or three dollars,
-
7:15 - 7:16they disagree about half the time,
-
7:16 - 7:18and when the amount is four, five, six, they agree quite often.
-
7:18 - 7:20This turns out to be something that's predicted
-
7:20 - 7:22by a very complicated type of game theory
-
7:22 - 7:25you should come to graduate school at CalTech and learn about.
-
7:25 - 7:27It's a little too complicated to explain right now,
-
7:27 - 7:31but the theory tells you that this shape kind of should occur.
-
7:31 - 7:33Your intuition might tell you that too.
-
7:33 - 7:35Now I'm going to show you the results from the EEG recording.
-
7:35 - 7:38Very complicated. The right brain schematic
-
7:38 - 7:41is the uninformed person, and the left is the informed.
-
7:41 - 7:43Remember that we scanned both brains at the same time,
-
7:43 - 7:46so we can ask about time-synced activity
-
7:46 - 7:49in similar or different areas simultaneously,
-
7:49 - 7:51just like if you wanted to study a conversation
-
7:51 - 7:53and you were scanning two people talking to each other
-
7:53 - 7:55and you'd expect common activity in language regions
-
7:55 - 7:58when they're actually kind of listening and communicating.
-
7:58 - 8:02So the arrows connect regions that are active at the same time,
-
8:02 - 8:04and the direction of the arrows flows
-
8:04 - 8:06from the region that's active first in time,
-
8:06 - 8:10and the arrowhead goes to the region that's active later.
-
8:10 - 8:12So in this case, if you look carefully,
-
8:12 - 8:14most of the arrows flow from right to left.
-
8:14 - 8:17That is, it looks as if the uninformed brain activity
-
8:17 - 8:19is happening first,
-
8:19 - 8:23and then it's followed by activity in the informed brain.
-
8:23 - 8:26And by the way, these were trials where their deals were made.
-
8:26 - 8:28This is from the first two seconds.
-
8:28 - 8:30We haven't finished analyzing this data,
-
8:30 - 8:32so we're still peeking in, but the hope is
-
8:32 - 8:35that we can say something in the first couple of seconds
-
8:35 - 8:36about whether they'll make a deal or not,
-
8:36 - 8:38which could be very useful in thinking about avoiding litigation
-
8:38 - 8:40and ugly divorces and things like that.
-
8:40 - 8:43Those are all cases in which a lot of value is lost
-
8:43 - 8:46by delay and strikes.
-
8:46 - 8:48Here's the case where the disagreements occur.
-
8:48 - 8:50You can see it looks different than the one before.
-
8:50 - 8:53There's a lot more arrows.
-
8:53 - 8:54That means that the brains are synced up
-
8:54 - 8:57more closely in terms of simultaneous activity,
-
8:57 - 8:59and the arrows flow clearly from left to right.
-
8:59 - 9:01That is, the informed brain seems to be deciding,
-
9:01 - 9:03"We're probably not going to make a deal here."
-
9:03 - 9:06And then later there's activity in the uninformed brain.
-
9:06 - 9:09Next I'm going to introduce you to some relatives.
-
9:09 - 9:11They're hairy, smelly, fast and strong.
-
9:11 - 9:14You might be thinking back to your last Thanksgiving.
-
9:14 - 9:17Maybe if you had a chimpanzee with you.
-
9:17 - 9:21Charles Darwin and I and you broke off from the family tree
-
9:21 - 9:23from chimpanzees about five million years ago.
-
9:23 - 9:25They're still our closest genetic kin.
-
9:25 - 9:27We share 98.8 percent of the genes.
-
9:27 - 9:29We share more genes with them than zebras do with horses.
-
9:29 - 9:31And we're also their closest cousin.
-
9:31 - 9:34They have more genetic relation to us than to gorillas.
-
9:34 - 9:37So how humans and chimpanzees behave differently
-
9:37 - 9:39might tell us a lot about brain evolution.
-
9:39 - 9:42So this is an amazing memory test
-
9:42 - 9:44from Nagoya, Japan, Primate Research Institute,
-
9:44 - 9:46where they've done a lot of this research.
-
9:46 - 9:49This goes back quite a ways. They're interested in working memory.
-
9:49 - 9:50The chimp is going to see, watch carefully,
-
9:50 - 9:53they're going to see 200 milliseconds' exposure
-
9:53 - 9:55— that's fast, that's eight movie frames —
-
9:55 - 9:57of numbers one, two, three, four, five.
-
9:57 - 9:59Then they disappear and they're replaced by squares,
-
9:59 - 10:00and they have to press the squares
-
10:00 - 10:03that correspond to the numbers from low to high
-
10:03 - 10:04to get an apple reward.
-
10:04 - 10:09Let's see how they can do it.
-
10:16 - 10:18This is a young chimp. The young ones
-
10:18 - 10:21are better than the old ones, just like humans.
-
10:21 - 10:22And they're highly experienced, so they've done this
-
10:22 - 10:24thousands and thousands of time.
-
10:24 - 10:27Obviously there's a big training effect, as you can imagine.
-
10:28 - 10:29(Laughter)
-
10:29 - 10:31You can see they're very blasé and kind of effortless.
-
10:31 - 10:35Not only can they do it very well, they do it in a sort of lazy way.
-
10:35 - 10:39Right? Who thinks you could beat the chimps?
-
10:39 - 10:40Wrong. (Laughter)
-
10:40 - 10:43We can try. We'll try. Maybe we'll try.
-
10:43 - 10:45Okay, so the next part of this study
-
10:45 - 10:47I'm going to go quickly through
-
10:47 - 10:49is based on an idea of Tetsuro Matsuzawa.
-
10:49 - 10:53He had a bold idea that -- what he called the cognitive trade-off hypothesis.
-
10:53 - 10:54We know chimps are faster and stronger.
-
10:54 - 10:55They're also very obsessed with status.
-
10:55 - 10:58His thought was, maybe they've preserved brain activities
-
10:58 - 11:01and they practice them in development
-
11:01 - 11:02that are really, really important to them
-
11:02 - 11:05to negotiate status and to win,
-
11:05 - 11:08which is something like strategic thinking during competition.
-
11:08 - 11:09So we're going to check that out
-
11:09 - 11:12by having the chimps actually play a game
-
11:12 - 11:14by touching two touch screens.
-
11:14 - 11:17The chimps are actually interacting with each other through the computers.
-
11:17 - 11:18They're going to press left or right.
-
11:18 - 11:20One chimp is called a matcher.
-
11:20 - 11:22They win if they press left, left,
-
11:22 - 11:26like a seeker finding someone in hide-and-seek, or right, right.
-
11:26 - 11:27The mismatcher wants to mismatch.
-
11:27 - 11:30They want to press the opposite screen of the chimp.
-
11:30 - 11:32And the rewards are apple cube rewards.
-
11:32 - 11:35So here's how game theorists look at these data.
-
11:35 - 11:37This is a graph of the percentage of times
-
11:37 - 11:39the matcher picked right on the x-axis,
-
11:39 - 11:41and the percentage of times they predicted right
-
11:41 - 11:44by the mismatcher on the y-axis.
-
11:44 - 11:47So a point here is the behavior by a pair of players,
-
11:47 - 11:49one trying to match, one trying to mismatch.
-
11:49 - 11:52The NE square in the middle -- actually NE, CH and QRE --
-
11:52 - 11:55those are three different theories of Nash equilibrium, and others,
-
11:55 - 11:57tells you what the theory predicts,
-
11:57 - 11:59which is that they should match 50-50,
-
11:59 - 12:02because if you play left too much, for example,
-
12:02 - 12:04I can exploit that if I'm the mismatcher by then playing right.
-
12:04 - 12:07And as you can see, the chimps, each chimp is one triangle,
-
12:07 - 12:11are circled around, hovering around that prediction.
-
12:11 - 12:13Now we move the payoffs.
-
12:13 - 12:16We're actually going to make the left, left payoff for the matcher a little bit higher.
-
12:16 - 12:18Now they get three apple cubes.
-
12:18 - 12:20Game theoretically, that should actually make the mismatcher's behavior shift,
-
12:20 - 12:22because what happens is, the mismatcher will think,
-
12:22 - 12:24oh, this guy's going to go for the big reward,
-
12:24 - 12:27and so I'm going to go to the right, make sure he doesn't get it.
-
12:27 - 12:29And as you can see, their behavior moves up
-
12:29 - 12:32in the direction of this change in the Nash equilibrium.
-
12:32 - 12:34Finally, we changed the payoffs one more time.
-
12:34 - 12:36Now it's four apple cubes,
-
12:36 - 12:38and their behavior again moves towards the Nash equilibrium.
-
12:38 - 12:40It's sprinkled around, but if you average the chimps out,
-
12:40 - 12:42they're really, really close, within .01.
-
12:42 - 12:45They're actually closer than any species we've observed.
-
12:45 - 12:48What about humans? You think you're smarter than a chimpanzee?
-
12:48 - 12:52Here's two human groups in green and blue.
-
12:52 - 12:56They're closer to 50-50. They're not responding to payoffs as closely,
-
12:56 - 12:57and also if you study their learning in the game,
-
12:57 - 12:59they aren't as sensitive to previous rewards.
-
12:59 - 13:00The chimps are playing better than the humans,
-
13:00 - 13:03better in the sense of adhering to game theory.
-
13:03 - 13:04And these are two different groups of humans
-
13:04 - 13:08from Japan and Africa. They replicate quite nicely.
-
13:08 - 13:11None of them are close to where the chimps are.
-
13:11 - 13:13So here are some things we learned today.
-
13:13 - 13:14People seem to do a limited amount of strategic thinking
-
13:14 - 13:16using theory of mind.
-
13:16 - 13:18We have some preliminary evidence from bargaining
-
13:18 - 13:21that early warning signs in the brain might be used to predict
-
13:21 - 13:23whether there will be a bad disagreement that costs money,
-
13:23 - 13:25and chimps are better competitors than humans,
-
13:25 - 13:27as judged by game theory.
-
13:27 - 13:29Thank you.
-
13:29 - 13:33(Applause)
- Title:
- When you're making a deal, what's going on in your brain?
- Speaker:
- Colin Camerer
- Description:
-
When two people are trying to make a deal -- whether they’re competing or cooperating -- what’s really going on inside their brains? Behavioral economist Colin Camerer shows research that reveals just how little we’re able to predict what others are thinking. And he presents an unexpected study that shows chimpanzees might just be better at it than we are. (Filmed at TEDxCalTech.)
- Video Language:
- English
- Team:
- closed TED
- Project:
- TEDTalks
- Duration:
- 13:49
Camille Martínez commented on English subtitles for When you're making a deal, what's going on in your brain? | ||
Camille Martínez edited English subtitles for When you're making a deal, what's going on in your brain? | ||
Camille Martínez edited English subtitles for When you're making a deal, what's going on in your brain? | ||
Camille Martínez edited English subtitles for When you're making a deal, what's going on in your brain? | ||
Joanna Pietrulewicz edited English subtitles for When you're making a deal, what's going on in your brain? | ||
Joanna Pietrulewicz edited English subtitles for When you're making a deal, what's going on in your brain? | ||
Joanna Pietrulewicz edited English subtitles for When you're making a deal, what's going on in your brain? | ||
Brian Greene edited English subtitles for When you're making a deal, what's going on in your brain? |
Camille Martínez
Hello,
The English transcript was updated on 5/3/20. In addition to other edits, please note the following change:
09:42 from Nagoya, Japan, Primate Research Institute ---> from [Kyoto], Japan, the Primate Research Institute
Thank you!