[Script Info]
Title: 
[Events]
Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV, Effect, Text
Dialogue: 0,0:00:00.86,0:00:03.23,Default,,0000,0000,0000,,I'm going to talk about\Nthe strategizing brain.
Dialogue: 0,0:00:03.25,0:00:05.64,Default,,0000,0000,0000,,We're going to use an unusual\Ncombination of tools
Dialogue: 0,0:00:05.66,0:00:07.25,Default,,0000,0000,0000,,from game theory and neuroscience
Dialogue: 0,0:00:07.28,0:00:10.55,Default,,0000,0000,0000,,to understand how people interact socially\Nwhen value is on the line.
Dialogue: 0,0:00:10.57,0:00:14.04,Default,,0000,0000,0000,,So game theory is a branch of,\Noriginally, applied mathematics,
Dialogue: 0,0:00:14.07,0:00:17.50,Default,,0000,0000,0000,,used mostly in economics and political\Nscience, a little bit in biology,
Dialogue: 0,0:00:17.52,0:00:20.36,Default,,0000,0000,0000,,that gives us a mathematical\Ntaxonomy of social life,
Dialogue: 0,0:00:20.39,0:00:22.71,Default,,0000,0000,0000,,and it predicts what people\Nare likely to do
Dialogue: 0,0:00:22.73,0:00:24.05,Default,,0000,0000,0000,,and believe others will do
Dialogue: 0,0:00:24.07,0:00:27.04,Default,,0000,0000,0000,,in cases where everyone's actions\Naffect everyone else.
Dialogue: 0,0:00:27.07,0:00:30.75,Default,,0000,0000,0000,,That's a lot of things: competition,\Ncooperation, bargaining,
Dialogue: 0,0:00:30.78,0:00:33.13,Default,,0000,0000,0000,,games like hide-and-seek and poker.
Dialogue: 0,0:00:33.95,0:00:35.89,Default,,0000,0000,0000,,Here's a simple game to get us started.
Dialogue: 0,0:00:35.91,0:00:38.38,Default,,0000,0000,0000,,Everyone chooses a number\Nfrom zero to 100.
Dialogue: 0,0:00:38.40,0:00:40.86,Default,,0000,0000,0000,,We're going to compute\Nthe average of those numbers,
Dialogue: 0,0:00:40.88,0:00:44.92,Default,,0000,0000,0000,,and whoever's closest to two-thirds\Nof the average wins a fixed prize.
Dialogue: 0,0:00:44.95,0:00:47.69,Default,,0000,0000,0000,,So you want to be a little bit\Nbelow the average number
Dialogue: 0,0:00:47.72,0:00:48.87,Default,,0000,0000,0000,,but not too far below,
Dialogue: 0,0:00:48.89,0:00:52.51,Default,,0000,0000,0000,,and everyone else wants to be a little bit\Nbelow the average number as well.
Dialogue: 0,0:00:52.54,0:00:54.08,Default,,0000,0000,0000,,Think about what you might pick.
Dialogue: 0,0:00:54.11,0:00:55.27,Default,,0000,0000,0000,,As you're thinking,
Dialogue: 0,0:00:55.29,0:00:58.35,Default,,0000,0000,0000,,this is a toy model of something like\Nselling in the stock market
Dialogue: 0,0:00:58.37,0:00:59.99,Default,,0000,0000,0000,,during a rising market:
Dialogue: 0,0:01:00.01,0:01:03.12,Default,,0000,0000,0000,,You don't want to sell too early,\Nbecause you miss out on profits,
Dialogue: 0,0:01:03.15,0:01:06.20,Default,,0000,0000,0000,,but you don't want to wait too late,\Nto when everyone else sells,
Dialogue: 0,0:01:06.22,0:01:07.44,Default,,0000,0000,0000,,triggering a crash.
Dialogue: 0,0:01:07.47,0:01:11.05,Default,,0000,0000,0000,,You want to be a little bit ahead\Nof the competition, but not too far ahead.
Dialogue: 0,0:01:11.07,0:01:14.16,Default,,0000,0000,0000,,OK, here's two theories\Nabout how people might think about this,
Dialogue: 0,0:01:14.18,0:01:15.38,Default,,0000,0000,0000,,then we'll see some data.
Dialogue: 0,0:01:15.40,0:01:17.00,Default,,0000,0000,0000,,Some of these will sound familiar
Dialogue: 0,0:01:17.02,0:01:19.08,Default,,0000,0000,0000,,because you probably\Nare thinking that way.
Dialogue: 0,0:01:19.11,0:01:20.76,Default,,0000,0000,0000,,I'm using my brain theory to see.
Dialogue: 0,0:01:20.78,0:01:24.20,Default,,0000,0000,0000,,A lot of people say, "I really don't know\Nwhat people are going to pick,
Dialogue: 0,0:01:24.22,0:01:27.75,Default,,0000,0000,0000,,so I think the average will be 50" --\Nthey're not being strategic at all --
Dialogue: 0,0:01:27.77,0:01:29.84,Default,,0000,0000,0000,,and "I'll pick two-thirds\Nof 50, that's 33."
Dialogue: 0,0:01:29.86,0:01:31.01,Default,,0000,0000,0000,,That's a start.
Dialogue: 0,0:01:31.04,0:01:33.43,Default,,0000,0000,0000,,Other people, who are a little\Nmore sophisticated,
Dialogue: 0,0:01:33.45,0:01:34.69,Default,,0000,0000,0000,,using more working memory,
Dialogue: 0,0:01:34.72,0:01:36.34,Default,,0000,0000,0000,,say, "I think people will pick 33,
Dialogue: 0,0:01:36.37,0:01:38.58,Default,,0000,0000,0000,,because they're going\Nto pick a response to 50,
Dialogue: 0,0:01:38.60,0:01:40.90,Default,,0000,0000,0000,,and so I'll pick 22,\Nwhich is two-thirds of 33."
Dialogue: 0,0:01:40.92,0:01:43.49,Default,,0000,0000,0000,,They're doing one extra step\Nof thinking, two steps.
Dialogue: 0,0:01:43.91,0:01:45.12,Default,,0000,0000,0000,,That's better.
Dialogue: 0,0:01:45.14,0:01:47.87,Default,,0000,0000,0000,,Of course, in principle,\Nyou could do three, four or more,
Dialogue: 0,0:01:47.89,0:01:49.84,Default,,0000,0000,0000,,but it starts to get very difficult.
Dialogue: 0,0:01:49.86,0:01:51.79,Default,,0000,0000,0000,,Just like in language and other domains,
Dialogue: 0,0:01:51.82,0:01:54.88,Default,,0000,0000,0000,,we know that it's hard for people\Nto parse very complex sentences
Dialogue: 0,0:01:54.91,0:01:56.20,Default,,0000,0000,0000,,with a recursive structure.
Dialogue: 0,0:01:56.22,0:01:58.37,Default,,0000,0000,0000,,This is called the cognitive\Nhierarchy theory,
Dialogue: 0,0:01:58.39,0:02:00.64,Default,,0000,0000,0000,,something I've worked on\Nand a few other people,
Dialogue: 0,0:02:00.67,0:02:02.44,Default,,0000,0000,0000,,and it indicates a kind of hierarchy,
Dialogue: 0,0:02:02.46,0:02:05.90,Default,,0000,0000,0000,,along with some assumptions about\Nhow many people stop at different steps
Dialogue: 0,0:02:05.92,0:02:07.93,Default,,0000,0000,0000,,and how the steps of thinking are affected
Dialogue: 0,0:02:07.96,0:02:10.40,Default,,0000,0000,0000,,by lots of interesting variables\Nand variant people,
Dialogue: 0,0:02:10.42,0:02:11.62,Default,,0000,0000,0000,,as we'll see in a minute.
Dialogue: 0,0:02:11.65,0:02:14.78,Default,,0000,0000,0000,,A very different theory, a much\Nmore popular one and an older one,
Dialogue: 0,0:02:14.81,0:02:17.29,Default,,0000,0000,0000,,due largely to John Nash\Nof "A Beautiful Mind" fame,
Dialogue: 0,0:02:17.31,0:02:19.40,Default,,0000,0000,0000,,is what's called "equilibrium analysis."
Dialogue: 0,0:02:19.42,0:02:22.23,Default,,0000,0000,0000,,So if you've ever taken\Na game theory course at any level,
Dialogue: 0,0:02:22.26,0:02:24.04,Default,,0000,0000,0000,,you'll have learned a bit about this.
Dialogue: 0,0:02:24.06,0:02:25.89,Default,,0000,0000,0000,,An equilibrium is a mathematical state
Dialogue: 0,0:02:25.91,0:02:29.21,Default,,0000,0000,0000,,in which everybody has figured out\Nexactly what everyone else will do.
Dialogue: 0,0:02:29.24,0:02:30.58,Default,,0000,0000,0000,,It is a very useful concept,
Dialogue: 0,0:02:30.60,0:02:32.66,Default,,0000,0000,0000,,but behaviorally,\Nit may not exactly explain
Dialogue: 0,0:02:32.68,0:02:35.96,Default,,0000,0000,0000,,what people do the first time they play\Nthese types of economic games
Dialogue: 0,0:02:35.98,0:02:37.90,Default,,0000,0000,0000,,or in situations in the outside world.
Dialogue: 0,0:02:37.92,0:02:40.72,Default,,0000,0000,0000,,In this case, the equilibrium\Nmakes a very bold prediction,
Dialogue: 0,0:02:40.75,0:02:43.46,Default,,0000,0000,0000,,which is: everyone wants\Nto be below everyone else,
Dialogue: 0,0:02:43.48,0:02:45.18,Default,,0000,0000,0000,,therefore, they'll play zero.
Dialogue: 0,0:02:45.72,0:02:46.88,Default,,0000,0000,0000,,Let's see what happens.
Dialogue: 0,0:02:46.90,0:02:49.01,Default,,0000,0000,0000,,This experiment's been done\Nmany, many times.
Dialogue: 0,0:02:49.04,0:02:51.24,Default,,0000,0000,0000,,Some of the earliest ones\Nwere done in the '90s
Dialogue: 0,0:02:51.26,0:02:53.07,Default,,0000,0000,0000,,by me and Rosemarie Nagel and others.
Dialogue: 0,0:02:53.09,0:02:55.61,Default,,0000,0000,0000,,This is a beautiful data set\Nof 9,000 people
Dialogue: 0,0:02:55.64,0:02:58.86,Default,,0000,0000,0000,,who wrote in to three newspapers\Nand magazines that had a contest.
Dialogue: 0,0:02:58.88,0:03:00.92,Default,,0000,0000,0000,,The contest said, send in your numbers,
Dialogue: 0,0:03:00.95,0:03:04.28,Default,,0000,0000,0000,,and whoever is close to two-thirds\Nof the average will win a big prize.
Dialogue: 0,0:03:04.30,0:03:08.04,Default,,0000,0000,0000,,As you can see, there's so much data\Nhere, you can see the spikes very visibly.
Dialogue: 0,0:03:08.06,0:03:10.78,Default,,0000,0000,0000,,There's a spike at 33 --\Nthose are people doing one step.
Dialogue: 0,0:03:10.80,0:03:13.02,Default,,0000,0000,0000,,There is another spike visible at 22.
Dialogue: 0,0:03:13.04,0:03:16.06,Default,,0000,0000,0000,,Notice, by the way, most people\Npick numbers right around there;
Dialogue: 0,0:03:16.08,0:03:18.24,Default,,0000,0000,0000,,they don't necessarily\Npick exactly 33 and 22.
Dialogue: 0,0:03:18.26,0:03:20.18,Default,,0000,0000,0000,,There's something a bit noisy around it.
Dialogue: 0,0:03:20.20,0:03:22.17,Default,,0000,0000,0000,,But you can see those spikes on that end.
Dialogue: 0,0:03:22.20,0:03:23.68,Default,,0000,0000,0000,,There's another group of people
Dialogue: 0,0:03:23.71,0:03:26.19,Default,,0000,0000,0000,,who seem to have a firm grip\Non equilibrium analysis,
Dialogue: 0,0:03:26.22,0:03:27.95,Default,,0000,0000,0000,,because they're picking zero or one.
Dialogue: 0,0:03:27.98,0:03:29.62,Default,,0000,0000,0000,,But they lose, right?
Dialogue: 0,0:03:29.65,0:03:33.03,Default,,0000,0000,0000,,Because picking a number that low\Nis actually a bad choice
Dialogue: 0,0:03:33.06,0:03:35.80,Default,,0000,0000,0000,,if other people aren't doing\Nequilibrium analysis as well.
Dialogue: 0,0:03:35.82,0:03:37.49,Default,,0000,0000,0000,,So they're smart, but poor.
Dialogue: 0,0:03:37.52,0:03:39.58,Default,,0000,0000,0000,,(Laughter)
Dialogue: 0,0:03:39.61,0:03:42.07,Default,,0000,0000,0000,,Where are these things\Nhappening in the brain?
Dialogue: 0,0:03:42.10,0:03:45.79,Default,,0000,0000,0000,,One study by Coricelli and Nagel\Ngives a really sharp, interesting answer.
Dialogue: 0,0:03:45.81,0:03:49.63,Default,,0000,0000,0000,,They had people play this game\Nwhile they were being scanned in an fMRI,
Dialogue: 0,0:03:49.65,0:03:50.81,Default,,0000,0000,0000,,and two conditions:
Dialogue: 0,0:03:50.83,0:03:52.22,Default,,0000,0000,0000,,in some trials, they're told,
Dialogue: 0,0:03:52.24,0:03:54.84,Default,,0000,0000,0000,,"You're playing another person\Nwho's playing right now.
Dialogue: 0,0:03:54.86,0:03:57.86,Default,,0000,0000,0000,,We'll match up your behavior\Nat the end and pay you if you win."
Dialogue: 0,0:03:57.89,0:04:00.62,Default,,0000,0000,0000,,In other trials, they're told,\N"You're playing a computer,
Dialogue: 0,0:04:00.64,0:04:02.16,Default,,0000,0000,0000,,they're just choosing randomly."
Dialogue: 0,0:04:02.19,0:04:04.35,Default,,0000,0000,0000,,So what you see here\Nis a subtraction of areas
Dialogue: 0,0:04:04.38,0:04:07.33,Default,,0000,0000,0000,,in which there's more brain activity\Nwhen you're playing people
Dialogue: 0,0:04:07.36,0:04:08.94,Default,,0000,0000,0000,,compared to playing the computer.
Dialogue: 0,0:04:08.96,0:04:11.50,Default,,0000,0000,0000,,And you see activity\Nin some regions we've seen today,
Dialogue: 0,0:04:11.52,0:04:13.77,Default,,0000,0000,0000,,medial prefrontal cortex,\Ndorsomedial, up here,
Dialogue: 0,0:04:13.79,0:04:16.18,Default,,0000,0000,0000,,ventromedial prefrontal cortex,\Nanterior cingulate,
Dialogue: 0,0:04:16.21,0:04:19.22,Default,,0000,0000,0000,,an area that's involved\Nin lots of types of conflict resolution,
Dialogue: 0,0:04:19.25,0:04:20.98,Default,,0000,0000,0000,,like if you're playing "Simon Says,"
Dialogue: 0,0:04:21.01,0:04:24.18,Default,,0000,0000,0000,,and also the right and left\Ntemporoparietal junction.
Dialogue: 0,0:04:24.20,0:04:27.08,Default,,0000,0000,0000,,And these are all areas\Nwhich are fairly reliably known to be
Dialogue: 0,0:04:27.10,0:04:29.36,Default,,0000,0000,0000,,part of what's called\Na "theory of mind" circuit
Dialogue: 0,0:04:29.38,0:04:30.90,Default,,0000,0000,0000,,or "mentalizing circuit."
Dialogue: 0,0:04:30.93,0:04:34.44,Default,,0000,0000,0000,,That is, it's a circuit that's used\Nto imagine what other people might do.
Dialogue: 0,0:04:34.46,0:04:38.37,Default,,0000,0000,0000,,These were some of the first studies\Nto see this tied in to game theory.
Dialogue: 0,0:04:38.78,0:04:41.02,Default,,0000,0000,0000,,What happens with these\None- and two-step types?
Dialogue: 0,0:04:41.05,0:04:43.30,Default,,0000,0000,0000,,So, we classify people\Nby what they picked,
Dialogue: 0,0:04:43.32,0:04:46.85,Default,,0000,0000,0000,,and then we look at the difference\Nbetween playing humans versus computers,
Dialogue: 0,0:04:46.88,0:04:48.94,Default,,0000,0000,0000,,which brain areas\Nare differentially active.
Dialogue: 0,0:04:48.97,0:04:50.93,Default,,0000,0000,0000,,On the top, you see the one-step players.
Dialogue: 0,0:04:50.96,0:04:52.34,Default,,0000,0000,0000,,There's almost no difference.
Dialogue: 0,0:04:52.37,0:04:55.24,Default,,0000,0000,0000,,The reason is, they're treating\Nother people like a computer,
Dialogue: 0,0:04:55.27,0:04:56.42,Default,,0000,0000,0000,,and the brain is too.
Dialogue: 0,0:04:56.44,0:04:59.47,Default,,0000,0000,0000,,The bottom players, you see\Nall the activity in dorsomedial PFC.
Dialogue: 0,0:04:59.49,0:05:02.50,Default,,0000,0000,0000,,So we know the two-step players\Nare doing something differently.
Dialogue: 0,0:05:02.52,0:05:04.52,Default,,0000,0000,0000,,Now, what can we do with this information?
Dialogue: 0,0:05:04.55,0:05:06.99,Default,,0000,0000,0000,,You might be able to look\Nat brain activity and say,
Dialogue: 0,0:05:07.01,0:05:10.65,Default,,0000,0000,0000,,"This person will be a good poker player,"\Nor "This person's socially naive."
Dialogue: 0,0:05:10.68,0:05:14.21,Default,,0000,0000,0000,,We might also be able to study things\Nlike development of adolescent brains
Dialogue: 0,0:05:14.24,0:05:16.67,Default,,0000,0000,0000,,once we have an idea\Nof where this circuitry exists.
Dialogue: 0,0:05:16.70,0:05:17.85,Default,,0000,0000,0000,,OK. Get ready.
Dialogue: 0,0:05:17.87,0:05:19.97,Default,,0000,0000,0000,,I'm saving you some brain activity,
Dialogue: 0,0:05:19.100,0:05:22.74,Default,,0000,0000,0000,,because you don't need to use\Nyour hair detector cells.
Dialogue: 0,0:05:22.76,0:05:26.02,Default,,0000,0000,0000,,You should use those cells\Nto think carefully about this game.
Dialogue: 0,0:05:26.04,0:05:27.56,Default,,0000,0000,0000,,This is a bargaining game.
Dialogue: 0,0:05:27.58,0:05:30.60,Default,,0000,0000,0000,,Two players who are being\Nscanned using EEG electrodes
Dialogue: 0,0:05:30.62,0:05:33.40,Default,,0000,0000,0000,,are going to bargain\Nover one to six dollars.
Dialogue: 0,0:05:33.42,0:05:36.11,Default,,0000,0000,0000,,If they can do it in 10 seconds,\Nthey'll earn that money.
Dialogue: 0,0:05:36.13,0:05:39.28,Default,,0000,0000,0000,,If 10 seconds go by and they haven't\Nmade a deal, they get nothing.
Dialogue: 0,0:05:39.30,0:05:40.93,Default,,0000,0000,0000,,That's kind of a mistake together.
Dialogue: 0,0:05:40.95,0:05:43.57,Default,,0000,0000,0000,,The twist is that one player, on the left,
Dialogue: 0,0:05:43.59,0:05:45.94,Default,,0000,0000,0000,,is informed about how much\Non each trial there is.
Dialogue: 0,0:05:45.96,0:05:48.68,Default,,0000,0000,0000,,They play lots of trials\Nwith different amounts each time.
Dialogue: 0,0:05:48.71,0:05:50.80,Default,,0000,0000,0000,,In this case, they know\Nthere's four dollars.
Dialogue: 0,0:05:50.83,0:05:54.43,Default,,0000,0000,0000,,The uninformed player doesn't know,\Nbut they know the informed player knows.
Dialogue: 0,0:05:54.45,0:05:56.65,Default,,0000,0000,0000,,So the uninformed player's\Nchallenge is to say,
Dialogue: 0,0:05:56.67,0:05:57.82,Default,,0000,0000,0000,,"Is this guy being fair,
Dialogue: 0,0:05:57.85,0:05:59.78,Default,,0000,0000,0000,,or are they giving me a very low offer
Dialogue: 0,0:05:59.80,0:06:03.57,Default,,0000,0000,0000,,in order to get me to think there's only\None or two dollars available to split?"
Dialogue: 0,0:06:03.60,0:06:06.32,Default,,0000,0000,0000,,in which case they might reject it\Nand not come to a deal.
Dialogue: 0,0:06:06.34,0:06:09.39,Default,,0000,0000,0000,,So there's some tension here\Nbetween trying to get the most money
Dialogue: 0,0:06:09.42,0:06:12.09,Default,,0000,0000,0000,,but trying to goad the other player\Ninto giving you more.
Dialogue: 0,0:06:12.12,0:06:14.59,Default,,0000,0000,0000,,And the way they bargain\Nis to point on a number line
Dialogue: 0,0:06:14.62,0:06:16.30,Default,,0000,0000,0000,,that goes from zero to six dollars.
Dialogue: 0,0:06:16.32,0:06:19.15,Default,,0000,0000,0000,,They're bargaining over how much\Nthe uninformed player gets,
Dialogue: 0,0:06:19.18,0:06:21.18,Default,,0000,0000,0000,,and the informed player will get the rest.
Dialogue: 0,0:06:21.20,0:06:23.35,Default,,0000,0000,0000,,So this is like\Na management-labor negotiation
Dialogue: 0,0:06:23.37,0:06:25.10,Default,,0000,0000,0000,,in which the workers don't know
Dialogue: 0,0:06:25.12,0:06:28.33,Default,,0000,0000,0000,,how much profits\Nthe privately held company has,
Dialogue: 0,0:06:28.36,0:06:30.79,Default,,0000,0000,0000,,and they want to maybe\Nhold out for more money,
Dialogue: 0,0:06:30.81,0:06:33.21,Default,,0000,0000,0000,,but the company might want\Nto create the impression
Dialogue: 0,0:06:33.23,0:06:36.19,Default,,0000,0000,0000,,that there's very little to split:\N"I'm giving the most I can."
Dialogue: 0,0:06:36.22,0:06:39.62,Default,,0000,0000,0000,,First, some behavior: a bunch\Nof the subject pairs play face-to-face.
Dialogue: 0,0:06:39.65,0:06:42.09,Default,,0000,0000,0000,,We have other data\Nwhere they play across computers.
Dialogue: 0,0:06:42.11,0:06:44.68,Default,,0000,0000,0000,,That's an interesting difference,\Nas you might imagine.
Dialogue: 0,0:06:44.71,0:06:46.48,Default,,0000,0000,0000,,But a bunch of the face-to-face pairs
Dialogue: 0,0:06:46.51,0:06:49.23,Default,,0000,0000,0000,,agree to divide the money\Nevenly every single time.
Dialogue: 0,0:06:49.26,0:06:51.92,Default,,0000,0000,0000,,Boring. It's just not\Ninteresting neurally.
Dialogue: 0,0:06:52.31,0:06:54.53,Default,,0000,0000,0000,,It's good for them --\Nthey make a lot of money.
Dialogue: 0,0:06:54.56,0:06:56.10,Default,,0000,0000,0000,,But we're interested in:
Dialogue: 0,0:06:56.12,0:06:59.87,Default,,0000,0000,0000,,Can we say something about when\Ndisagreements occur versus don't occur?
Dialogue: 0,0:06:59.90,0:07:02.66,Default,,0000,0000,0000,,So this is the other group\Nof subjects, who often disagree.
Dialogue: 0,0:07:02.68,0:07:06.16,Default,,0000,0000,0000,,They bicker and disagree\Nand end up with less money.
Dialogue: 0,0:07:06.18,0:07:09.15,Default,,0000,0000,0000,,They might be eligible to be\Non "Real Housewives," the TV show.
Dialogue: 0,0:07:09.17,0:07:10.26,Default,,0000,0000,0000,,(Laughter)
Dialogue: 0,0:07:10.28,0:07:11.97,Default,,0000,0000,0000,,You see on the left,
Dialogue: 0,0:07:11.99,0:07:14.62,Default,,0000,0000,0000,,when the amount to divide\Nis one, two or three dollars,
Dialogue: 0,0:07:14.65,0:07:16.27,Default,,0000,0000,0000,,they disagree about half the time;
Dialogue: 0,0:07:16.29,0:07:18.64,Default,,0000,0000,0000,,when it's four, five, six,\Nthey agree quite often.
Dialogue: 0,0:07:18.67,0:07:20.87,Default,,0000,0000,0000,,This turns out to be\Nsomething that's predicted
Dialogue: 0,0:07:20.89,0:07:22.85,Default,,0000,0000,0000,,by a very complicated type of game theory
Dialogue: 0,0:07:22.88,0:07:25.98,Default,,0000,0000,0000,,you should come to graduate school\Nat CalTech and learn about.
Dialogue: 0,0:07:26.01,0:07:28.40,Default,,0000,0000,0000,,It's a little too complicated\Nto explain right now,
Dialogue: 0,0:07:28.42,0:07:31.06,Default,,0000,0000,0000,,but the theory tells you\Nthat this shape should occur.
Dialogue: 0,0:07:31.09,0:07:33.15,Default,,0000,0000,0000,,Your intuition might tell you that, too.
Dialogue: 0,0:07:33.17,0:07:36.04,Default,,0000,0000,0000,,Now I'm going to show you\Nthe results from the EEG recording.
Dialogue: 0,0:07:36.06,0:07:37.22,Default,,0000,0000,0000,,Very complicated.
Dialogue: 0,0:07:37.24,0:07:39.63,Default,,0000,0000,0000,,The right brain schematic\Nis the uninformed person,
Dialogue: 0,0:07:39.66,0:07:41.06,Default,,0000,0000,0000,,and the left is the informed.
Dialogue: 0,0:07:41.08,0:07:43.82,Default,,0000,0000,0000,,Remember that we scanned\Nboth brains at the same time,
Dialogue: 0,0:07:43.85,0:07:46.12,Default,,0000,0000,0000,,so we can ask about time-synced activity
Dialogue: 0,0:07:46.14,0:07:49.16,Default,,0000,0000,0000,,in similar or different\Nareas simultaneously,
Dialogue: 0,0:07:49.18,0:07:51.45,Default,,0000,0000,0000,,just like if you wanted\Nto study a conversation,
Dialogue: 0,0:07:51.47,0:07:54.05,Default,,0000,0000,0000,,and you were scanning two people\Ntalking to each other.
Dialogue: 0,0:07:54.07,0:07:56.33,Default,,0000,0000,0000,,You'd expect common\Nactivity in language regions
Dialogue: 0,0:07:56.36,0:07:58.32,Default,,0000,0000,0000,,when they're listening and communicating.
Dialogue: 0,0:07:58.34,0:08:02.17,Default,,0000,0000,0000,,So the arrows connect regions\Nthat are active at the same time.
Dialogue: 0,0:08:02.20,0:08:03.52,Default,,0000,0000,0000,,The direction of the arrows
Dialogue: 0,0:08:03.54,0:08:06.31,Default,,0000,0000,0000,,flows from the region\Nthat's active first in time,
Dialogue: 0,0:08:06.33,0:08:10.13,Default,,0000,0000,0000,,and the arrowhead goes\Nto the region that's active later.
Dialogue: 0,0:08:10.15,0:08:12.20,Default,,0000,0000,0000,,So in this case, if you look carefully,
Dialogue: 0,0:08:12.22,0:08:14.24,Default,,0000,0000,0000,,most of the arrows\Nflow from right to left.
Dialogue: 0,0:08:14.27,0:08:17.55,Default,,0000,0000,0000,,That is, it looks\Nas if the uninformed brain activity
Dialogue: 0,0:08:17.58,0:08:19.19,Default,,0000,0000,0000,,is happening first,
Dialogue: 0,0:08:19.21,0:08:23.06,Default,,0000,0000,0000,,and then it's followed\Nby activity in the informed brain.
Dialogue: 0,0:08:23.09,0:08:26.54,Default,,0000,0000,0000,,And by the way, these are trials\Nwhere their deals were made.
Dialogue: 0,0:08:26.56,0:08:28.32,Default,,0000,0000,0000,,This is from the first two seconds.
Dialogue: 0,0:08:28.34,0:08:31.50,Default,,0000,0000,0000,,We haven't finished analyzing this data,\Nso we're still peeking in,
Dialogue: 0,0:08:31.52,0:08:34.93,Default,,0000,0000,0000,,but the hope is that we can say something\Nin the first couple of seconds
Dialogue: 0,0:08:34.96,0:08:36.92,Default,,0000,0000,0000,,about whether they'll make a deal or not,
Dialogue: 0,0:08:36.94,0:08:39.95,Default,,0000,0000,0000,,which could be very useful in thinking\Nabout avoiding litigation
Dialogue: 0,0:08:39.97,0:08:41.84,Default,,0000,0000,0000,,and ugly divorces and things like that.
Dialogue: 0,0:08:41.86,0:08:45.94,Default,,0000,0000,0000,,Those are all cases in which a lot\Nof value is lost by delay and strikes.
Dialogue: 0,0:08:46.63,0:08:48.79,Default,,0000,0000,0000,,Here's the case where\Nthe disagreements occur.
Dialogue: 0,0:08:48.82,0:08:51.21,Default,,0000,0000,0000,,You can see it looks different\Nthan the one before.
Dialogue: 0,0:08:51.24,0:08:52.58,Default,,0000,0000,0000,,There's a lot more arrows.
Dialogue: 0,0:08:52.60,0:08:55.25,Default,,0000,0000,0000,,That means that the brains\Nare synced up more closely
Dialogue: 0,0:08:55.28,0:08:56.90,Default,,0000,0000,0000,,in terms of simultaneous activity,
Dialogue: 0,0:08:56.92,0:08:59.12,Default,,0000,0000,0000,,and the arrows flow clearly\Nfrom left to right.
Dialogue: 0,0:08:59.15,0:09:01.44,Default,,0000,0000,0000,,That is, the informed brain\Nseems to be deciding,
Dialogue: 0,0:09:01.46,0:09:03.65,Default,,0000,0000,0000,,"We're probably not going\Nto make a deal here."
Dialogue: 0,0:09:03.68,0:09:06.42,Default,,0000,0000,0000,,And then later, there's activity\Nin the uninformed brain.
Dialogue: 0,0:09:06.80,0:09:09.20,Default,,0000,0000,0000,,Next, I'm going to introduce you\Nto some relatives.
Dialogue: 0,0:09:09.23,0:09:11.39,Default,,0000,0000,0000,,They're hairy, smelly, fast and strong.
Dialogue: 0,0:09:11.41,0:09:13.91,Default,,0000,0000,0000,,You might be thinking back\Nto your last Thanksgiving.
Dialogue: 0,0:09:13.93,0:09:14.95,Default,,0000,0000,0000,,(Laughter)
Dialogue: 0,0:09:14.97,0:09:17.45,Default,,0000,0000,0000,,Maybe, if you had a chimpanzee with you.
Dialogue: 0,0:09:17.47,0:09:21.48,Default,,0000,0000,0000,,Charles Darwin and I and you broke\Noff from the family tree from chimpanzees
Dialogue: 0,0:09:21.50,0:09:22.90,Default,,0000,0000,0000,,about five million years ago.
Dialogue: 0,0:09:22.92,0:09:24.74,Default,,0000,0000,0000,,They're still our closest genetic kin.
Dialogue: 0,0:09:24.76,0:09:26.48,Default,,0000,0000,0000,,We share 98.8 percent of the genes.
Dialogue: 0,0:09:26.50,0:09:29.46,Default,,0000,0000,0000,,We share more genes with them\Nthan zebras do with horses.
Dialogue: 0,0:09:29.49,0:09:31.40,Default,,0000,0000,0000,,And we're also their closest cousin.
Dialogue: 0,0:09:31.42,0:09:34.04,Default,,0000,0000,0000,,They have more genetic relation\Nto us than to gorillas.
Dialogue: 0,0:09:34.07,0:09:36.80,Default,,0000,0000,0000,,So, how humans and chimpanzees\Nbehave differently
Dialogue: 0,0:09:36.83,0:09:38.92,Default,,0000,0000,0000,,might tell us a lot about brain evolution.
Dialogue: 0,0:09:39.33,0:09:41.63,Default,,0000,0000,0000,,This is an amazing memory test
Dialogue: 0,0:09:41.65,0:09:44.44,Default,,0000,0000,0000,,from [Kyoto], Japan,\Nthe Primate Research Institute,
Dialogue: 0,0:09:44.47,0:09:46.47,Default,,0000,0000,0000,,where they've done a lot of this research.
Dialogue: 0,0:09:46.49,0:09:49.32,Default,,0000,0000,0000,,This goes back a ways.\NThey're interested in working memory.
Dialogue: 0,0:09:49.34,0:09:51.06,Default,,0000,0000,0000,,The chimp will see, watch carefully,
Dialogue: 0,0:09:51.08,0:09:54.66,Default,,0000,0000,0000,,they'll see 200 milliseconds' exposure --\Nthat's fast, eight movie frames --
Dialogue: 0,0:09:54.69,0:09:56.67,Default,,0000,0000,0000,,of numbers one, two, three, four, five.
Dialogue: 0,0:09:56.69,0:09:58.94,Default,,0000,0000,0000,,Then they disappear\Nand are replaced by squares,
Dialogue: 0,0:09:58.96,0:10:00.59,Default,,0000,0000,0000,,and they have to press the squares
Dialogue: 0,0:10:00.61,0:10:02.81,Default,,0000,0000,0000,,that correspond to the numbers\Nfrom low to high
Dialogue: 0,0:10:02.83,0:10:04.14,Default,,0000,0000,0000,,to get an apple reward.
Dialogue: 0,0:10:04.16,0:10:05.66,Default,,0000,0000,0000,,Let's see how they can do it.
Dialogue: 0,0:10:16.48,0:10:17.64,Default,,0000,0000,0000,,This is a young chimp.
Dialogue: 0,0:10:17.66,0:10:20.58,Default,,0000,0000,0000,,The young ones are better\Nthan the old ones, just like humans.
Dialogue: 0,0:10:20.60,0:10:21.61,Default,,0000,0000,0000,,(Laughter)
Dialogue: 0,0:10:21.63,0:10:23.11,Default,,0000,0000,0000,,And they're highly experienced,
Dialogue: 0,0:10:23.13,0:10:25.46,Default,,0000,0000,0000,,they've done this thousands of times.
Dialogue: 0,0:10:25.48,0:10:28.37,Default,,0000,0000,0000,,Obviously there's a big training\Neffect, as you can imagine.
Dialogue: 0,0:10:28.39,0:10:29.40,Default,,0000,0000,0000,,(Laughter)
Dialogue: 0,0:10:29.43,0:10:31.57,Default,,0000,0000,0000,,You can see they're very\Nblasé and effortless.
Dialogue: 0,0:10:31.60,0:10:34.81,Default,,0000,0000,0000,,Not only can they do it very well,\Nthey do it in a sort of lazy way.
Dialogue: 0,0:10:34.83,0:10:35.84,Default,,0000,0000,0000,,(Laughter)
Dialogue: 0,0:10:35.86,0:10:37.62,Default,,0000,0000,0000,,Who thinks you could beat the chimps?
Dialogue: 0,0:10:37.65,0:10:38.71,Default,,0000,0000,0000,,(Laughter)
Dialogue: 0,0:10:38.73,0:10:40.27,Default,,0000,0000,0000,,Wrong. (Laughter)
Dialogue: 0,0:10:40.29,0:10:42.88,Default,,0000,0000,0000,,We can try. We'll try. Maybe we'll try.
Dialogue: 0,0:10:42.90,0:10:46.89,Default,,0000,0000,0000,,OK, so the next part of the study\NI'm going to go quickly through
Dialogue: 0,0:10:46.92,0:10:49.89,Default,,0000,0000,0000,,is based on an idea of Tetsuro Matsuzawa.
Dialogue: 0,0:10:49.92,0:10:53.04,Default,,0000,0000,0000,,He had a bold idea he called\Nthe "cognitive trade-off hypothesis."
Dialogue: 0,0:10:53.06,0:10:56.54,Default,,0000,0000,0000,,We know chimps are faster and stronger;\Nthey're also obsessed with status.
Dialogue: 0,0:10:56.57,0:10:59.25,Default,,0000,0000,0000,,His thought was, maybe\Nthey've preserved brain activities
Dialogue: 0,0:10:59.27,0:11:00.88,Default,,0000,0000,0000,,and practice them in development
Dialogue: 0,0:11:00.90,0:11:04.71,Default,,0000,0000,0000,,that are really, really important to them\Nto negotiate status and to win,
Dialogue: 0,0:11:04.74,0:11:07.73,Default,,0000,0000,0000,,which is something like strategic\Nthinking during competition.
Dialogue: 0,0:11:07.75,0:11:09.29,Default,,0000,0000,0000,,So we're going to check that out
Dialogue: 0,0:11:09.31,0:11:11.94,Default,,0000,0000,0000,,by having the chimps actually play a game
Dialogue: 0,0:11:11.96,0:11:14.48,Default,,0000,0000,0000,,by touching two touch screens.
Dialogue: 0,0:11:14.50,0:11:17.56,Default,,0000,0000,0000,,The chimps are interacting\Nwith each other through the computers.
Dialogue: 0,0:11:17.58,0:11:18.93,Default,,0000,0000,0000,,They'll press left or right.
Dialogue: 0,0:11:18.96,0:11:22.43,Default,,0000,0000,0000,,One chimp is called a matcher;\Nthey win if they press left-left,
Dialogue: 0,0:11:22.46,0:11:25.60,Default,,0000,0000,0000,,like a seeker finding someone\Nin hide-and-seek, or right-right.
Dialogue: 0,0:11:25.63,0:11:27.23,Default,,0000,0000,0000,,The mismatcher wants to mismatch;
Dialogue: 0,0:11:27.26,0:11:29.96,Default,,0000,0000,0000,,they want to press\Nthe opposite screen of the chimp.
Dialogue: 0,0:11:29.98,0:11:32.45,Default,,0000,0000,0000,,And the rewards are apple cube rewards.
Dialogue: 0,0:11:32.48,0:11:34.80,Default,,0000,0000,0000,,So here's how game theorists\Nlook at these data.
Dialogue: 0,0:11:34.83,0:11:36.85,Default,,0000,0000,0000,,This is a graph of the percentage of times
Dialogue: 0,0:11:36.87,0:11:39.08,Default,,0000,0000,0000,,the matcher picked right on the x-axis
Dialogue: 0,0:11:39.10,0:11:41.26,Default,,0000,0000,0000,,and the percentage of times\Nthey picked right
Dialogue: 0,0:11:41.28,0:11:43.48,Default,,0000,0000,0000,,by the mismatcher on the y-axis.
Dialogue: 0,0:11:43.51,0:11:46.84,Default,,0000,0000,0000,,So a point here is the behavior\Nby a pair of players,
Dialogue: 0,0:11:46.86,0:11:49.06,Default,,0000,0000,0000,,one trying to match,\None trying to mismatch.
Dialogue: 0,0:11:49.08,0:11:52.40,Default,,0000,0000,0000,,The NE square in the middle --\Nactually, NE, CH and QRE --
Dialogue: 0,0:11:52.42,0:11:55.55,Default,,0000,0000,0000,,those are three different theories\Nof Nash equilibrium and others,
Dialogue: 0,0:11:55.57,0:11:57.25,Default,,0000,0000,0000,,tells you what the theory predicts,
Dialogue: 0,0:11:57.28,0:11:59.40,Default,,0000,0000,0000,,which is that they should match 50-50,
Dialogue: 0,0:11:59.43,0:12:01.85,Default,,0000,0000,0000,,because if you play left\Ntoo much, for example,
Dialogue: 0,0:12:01.88,0:12:04.84,Default,,0000,0000,0000,,I can exploit that if I'm the mismatcher\Nby then playing right.
Dialogue: 0,0:12:04.87,0:12:07.84,Default,,0000,0000,0000,,And as you can see, the chimps --\Neach chimp is one triangle --
Dialogue: 0,0:12:07.86,0:12:10.52,Default,,0000,0000,0000,,are circled around,\Nhovering around that prediction.
Dialogue: 0,0:12:11.20,0:12:12.91,Default,,0000,0000,0000,,Now we move the payoffs.
Dialogue: 0,0:12:12.94,0:12:16.42,Default,,0000,0000,0000,,We're going to make the left-left payoff\Nfor the matcher a little higher.
Dialogue: 0,0:12:16.45,0:12:17.94,Default,,0000,0000,0000,,Now they get three apple cubes.
Dialogue: 0,0:12:17.96,0:12:21.24,Default,,0000,0000,0000,,Game theoretically, that should\Nmake the mismatcher's behavior shift:
Dialogue: 0,0:12:21.26,0:12:24.77,Default,,0000,0000,0000,,the mismatcher will think, "Oh, this guy's\Ngoing to go for the big reward,
Dialogue: 0,0:12:24.80,0:12:27.32,Default,,0000,0000,0000,,so I'll go to the right,\Nmake sure he doesn't get it."
Dialogue: 0,0:12:27.35,0:12:29.38,Default,,0000,0000,0000,,And as you can see,\Ntheir behavior moves up
Dialogue: 0,0:12:29.40,0:12:32.10,Default,,0000,0000,0000,,in the direction of this change\Nin the Nash equilibrium.
Dialogue: 0,0:12:32.12,0:12:34.37,Default,,0000,0000,0000,,Finally, we changed\Nthe payoffs one more time.
Dialogue: 0,0:12:34.39,0:12:35.64,Default,,0000,0000,0000,,Now it's four apple cubes,
Dialogue: 0,0:12:35.66,0:12:38.50,Default,,0000,0000,0000,,and their behavior again moves\Ntowards the Nash equilibrium.
Dialogue: 0,0:12:38.52,0:12:41.19,Default,,0000,0000,0000,,It's sprinkled around,\Nbut if you average the chimps out,
Dialogue: 0,0:12:41.22,0:12:42.79,Default,,0000,0000,0000,,they're really close, within .01.
Dialogue: 0,0:12:42.82,0:12:45.44,Default,,0000,0000,0000,,They're actually closer\Nthan any species we've observed.
Dialogue: 0,0:12:45.47,0:12:48.57,Default,,0000,0000,0000,,What about humans? You think\Nyou're smarter than a chimpanzee?
Dialogue: 0,0:12:49.35,0:12:52.65,Default,,0000,0000,0000,,Here's two human groups in green and blue.
Dialogue: 0,0:12:52.68,0:12:55.97,Default,,0000,0000,0000,,They're closer to 50-50; they're not\Nresponding to payoffs as closely.
Dialogue: 0,0:12:55.99,0:12:58.29,Default,,0000,0000,0000,,And also if you study\Ntheir learning in the game,
Dialogue: 0,0:12:58.31,0:13:00.41,Default,,0000,0000,0000,,they aren't as sensitive\Nto previous rewards.
Dialogue: 0,0:13:00.44,0:13:04.02,Default,,0000,0000,0000,,The chimps play better than the humans,\Nin terms of adhering to game theory.
Dialogue: 0,0:13:04.05,0:13:07.25,Default,,0000,0000,0000,,And these are two different groups\Nof humans, from Japan and Africa;
Dialogue: 0,0:13:07.27,0:13:08.61,Default,,0000,0000,0000,,they replicate quite nicely.
Dialogue: 0,0:13:08.64,0:13:11.18,Default,,0000,0000,0000,,None of them are close\Nto where the chimps are.
Dialogue: 0,0:13:11.67,0:13:12.96,Default,,0000,0000,0000,,So, some things we learned:
Dialogue: 0,0:13:12.99,0:13:16.73,Default,,0000,0000,0000,,people seem to do a limited amount of\Nstrategic thinking using theory of mind.
Dialogue: 0,0:13:16.76,0:13:18.85,Default,,0000,0000,0000,,We have preliminary\Nevidence from bargaining
Dialogue: 0,0:13:18.88,0:13:21.79,Default,,0000,0000,0000,,that early warning signs in the brain\Nmight be used to predict
Dialogue: 0,0:13:21.82,0:13:24.45,Default,,0000,0000,0000,,whether there'll be a bad\Ndisagreement that costs money,
Dialogue: 0,0:13:24.47,0:13:26.71,Default,,0000,0000,0000,,and chimps are "better"\Ncompetitors than humans,
Dialogue: 0,0:13:26.73,0:13:27.98,Default,,0000,0000,0000,,as judged by game theory.
Dialogue: 0,0:13:27.100,0:13:29.15,Default,,0000,0000,0000,,Thank you.
Dialogue: 0,0:13:29.17,0:13:32.29,Default,,0000,0000,0000,,(Applause)