TED演讲 | 拥有大脑的真正原因

I'm a neuroscientist. And in neuroscience, we have to deal with many difficult questions about the brain.

我是个神经生物学家。在神经生物学研究里，我们需要处理很多关于大脑的难题。

But I want to start with the easiest question and the question you really should have all asked yourselves at some point in your life,

不过我今天要从这个最简单的问题开始，这个问题是大家应该在以前问过自己的，

because it's a fundamental question if we want to understand brain function.

因为如果大家想了解大脑功能的话。

And that is, why do we and other animals have brains?

这个问题是很基本的，那就是，为什么我们和其他动物都有大脑呢？

Not all species on our planet have brains, so if we want to know what the brain is for, let's think about why we evolved one.

地球上不是所有物种都有大脑，所以要搞清大脑有什么用，那我们应该先考虑为什么我们进化出了大脑。

Now you may reason that we have one to perceive the world or to think, and that's completely wrong.

大家也许觉得我们有大脑是为了感知世界或者思考，其实是完全错误的。

If you think about this question for any length of time, it's blindingly obvious why we have a brain.

如果再好好思考一段时间的话，其实我们有大脑的原因很明显。

We have a brain for one reason and one reason only, and that's to produce adaptable and complex movements.

大脑的存在，有且仅有一个原因就是为了生成有适应性的、复杂的动作。

There is no other reason to have a brain.

没有其他原因了。

Think about it. Movement is the only way you have of affecting the world around you.

想想看。动作是你唯一对世界产生影响的方法。

That's not quite true. There's one other way, and that's through sweating.

当然也不尽然，另一个办法是通过排汗。

But apart from that, everything else goes through contractions of muscles.

不过除了那个以外，所有事情都是由肌肉的收缩来办到的。

So think about communication -- speech, gestures, writing, sign language -- they're all mediated through contractions of your muscles.

想想我们的沟通方式--语言、手势、写作、手语--都是由肌肉的收缩来完成的。

So it's really important to remember that sensory, memory and cognitive processes are all important,

所以重要的一点就是感知、记忆，认知过程也很关键，

but they're only important to either drive or suppress future movements.

但是它们的重要性都只是在于它们能够促使或者抑制动作的发生。

There can be no evolutionary advantage to laying down memories of childhood or perceiving the color of a rose if it doesn't affect the way you're going to move later in life.

如果不是能够影响到大家未来生活中的动作的话，能够记住小时候的回忆或者感知到玫瑰花的颜色，在进化这一点上没有任何特别的优势。

Now for those who don't believe this argument, we have trees and grass on our planet without the brain,

如果有谁不相信这个说法，我可以举例说不会运动的树和草都没有大脑，

but the clinching evidence is this animal here -- the humble sea squirt.

不过有一种动物提供了确凿的证据--不起眼的海鞘。

Rudimentary animal, has a nervous system, swims around in the ocean in its juvenile life.

海鞘是一种初等生物，有神经系统，生命初期在海中游动。

And at some point of its life, it implants on a rock.

然后在某一时刻海鞘会把自己移植到岩石上再也不移动。

And the first thing it does in implanting on that rock, which it never leaves, is to digest its own brain and nervous system for food.

海鞘在岩石上固定后的第一件事就是把自己的大脑和神经系统作为食物消化掉。

So once you don't need to move, you don't need the luxury of that brain.

也就是说，一旦不用再移动也就不需要大脑这样的奢侈品了。

And this animal is often taken as an analogy to what happens at universities when professors get tenure, but that's a different subject.

并且，这种动物常常被类比到大学里面，教授获得终身职位之后的情况，不过那是另一个话题了。

So I am a movement chauvinist. I believe movement is the most important function of the brain -- don't let anyone tell you that it's not true.

所以说，我是个运动沙文主义者，我认为运动是大脑最重要的功能，不要让别人告诉你这个观点不对。

Now if movement is so important, how well are we doing understanding how the brain controls movement?

那么，如果运动如此重要，我们已经能在多大程度上了解大脑如何控制动作了呢？

And the answer is we're doing extremely poorly; it's a very hard problem.

答案其实是非常少，因为这是相当困难的。

But we can look at how well we're doing by thinking about how well we're doing building machines which can do what humans can do.

不过我们已经制造出一些模拟人类动作的机器人，分析这些机器的动作水平就知道我们了解大脑控制动作的研究进展如何了。

Think about the game of chess. How well are we doing determining what piece to move where?

来看国际象棋这个游戏。我们让机器人决定如何走棋时候的表现如何呢？

If you pit Garry Kasparov here, when he's not in jail, against IBM's Deep Blue, well the answer is IBM's Deep Blue will occasionally win.

如果大家趁国际象棋冠军加里·卡斯帕罗夫还没进监狱的时候，把他请来与IBM的深蓝对战，深蓝有时会赢。

And I think if IBM's Deep Blue played anyone in this room, it would win every time.

而且我觉得IBM的深蓝和在座的任何一位对战，应该每次都会赢。

That problem is solved. What about the problem of picking up a chess piece, dexterously manipulating it and putting it back down on the board?

所以这方面完全没有问题。但是如果让机器人灵巧地拿起棋子，再放回棋盘上去呢？

If you put a five year-old child's dexterity against the best robots of today, the answer is simple: the child wins easily. There's no competition at all.

如果大家让五岁小孩子与当今最厉害的机器人对决，结果很简单，小孩子会赢得轻而易举。完全没有悬念。

Now why is that top problem so easy and the bottom problem so hard?

那么为什么前面那个问题这么简单，后面这个问题就这么难呢？

One reason is a very smart five year-old could tell you the algorithm for that top problem -- look at all possible moves to the end of the game and choose the one that makes you win.

一个原因是，一个聪明点的五岁小孩子就已经能够告诉你上面那个问题的解决算法了--找出游戏结束之前所有可能的下法，选择赢面最大的一步来下。

So it's a very simple algorithm.

所以其实这是个很简单的算法。

Now of course there are other moves, but with vast computers we approximate and come close to the optimal solution.

当然也有其他的步法，不过用强大的计算机做近似计算就能很容易找到近似最优解。

When it comes to being dexterous, it's not even clear what the algorithm is you have to solve to be dexterous.

但在灵活性这个问题上，我们甚至连让机器人变灵活的算法都找不到。

And we'll see you have to both perceive and act on the world, which has a lot of problems.

可以看到，如果既感知世界，又作用于世界的话，其实是要面对很多问题的。

But let me show you cutting-edge robotics.

现在我来介绍一下最高端的机器人技术。

Now a lot of robotics is very impressive, but manipulation robotics is really just in the dark ages.

如今许多项机器人技术已经相当厉害，可是在机器人动作控制方面我们遇到了瓶颈，看不到曙光。

So this is the end of a Ph.D. project from one of the best robotics institutes.

这是在一个顶尖机器人控制学院里，一个博士研究项目的成果。

And the student has trained this robot to pour this water into a glass.

这位博士生训练这个机器人向这个玻璃杯里面倒水。

It's a hard problem because the water sloshes about, but it can do it.

这很难实现，因为水会洒出来，不过机器人能做到。

But it doesn't do it with anything like the agility of a human.

但是机器人无法达到任何接近于人类的灵活程度。

Now if you want this robot to do a different task, that's another three-year Ph.D. program.

并且如果想让机器人完成另一个任务的话，就得要另外一个历时三年的博士项目了。

There is no generalization at all from one task to another in robotics.

在机器人控制领域，从一项任务到另一项任务，根本没有共通性可言。

Now we can compare this to cutting-edge human performance.

现在我们再来与最尖端的人类性能做比较。

So what I'm going to show you is Emily Fox winning the world record for cup stacking.

我会播放艾米莉·福克斯叠杯子获得世界记录的场景。

Now the Americans in the audience will know all about cup stacking.

观众里有美国人的话一定知道叠杯子这个游戏。

It's a high school sport where you have 12 cups you have to stack and unstack against the clock in a prescribed order.

这是个高中里的娱乐项目，要求把12只杯子按照一定顺序以最快速度摞起来再展开回来。

And this is her getting the world record in real time.

下面就是她创下世界记录的实时情况。

And she's pretty happy. We have no idea what is going on inside her brain when she does that, and that's what we'd like to know.

看她多开心。我们不清楚她叠杯子的时候大脑里发生了什么，而且我们的确很想知道。

So in my group, what we try to do is reverse engineer how humans control movement.

所以我和我的团队，试图去逆向实现人类如何控制动作的过程。

And it sounds like an easy problem.

看起来问题很简单。

You send a command down, it causes muscles to contract.

你送出一个指令，令肌肉收缩。

Your arm or body moves, and you get sensory feedback from vision, from skin, from muscles and so on.

这样手臂或躯干就会运动，然后能够获得从视觉、皮肤和肌肉和其他地方获得感官回馈。

The trouble is these signals are not the beautiful signals you want them to be.

但问题是这些信号远没有想象中的那样完美。

So one thing that makes controlling movement difficult is, for example, sensory feedback is extremely noisy.

比如说，其中的难点之一就是感官回馈中的杂音非常大。

Now by noise, I do not mean sound.

我这里所说的“杂音”不是指声音。

We use it in the engineering and neuroscience sense meaning a random noise corrupting a signal.

在工程学和神经学里面，我们用“杂音”来表示干扰正常信号的随机杂乱信号。

So the old days before digital radio when you were tuning in your radio and you heard "crrcckkk" on the station you wanted to hear, that was the noise.

所以以前还没有电子收音机的时候，如果调那种老式收音机，电台里面有时候有“喀喇喀喇”的声音，那个就是杂音。

But more generally, this noise is something that corrupts the signal.

推广来说，杂音就是相对于正常信号的干扰信号。

So for example, if you put your hand under a table and try to localize it with your other hand, you can be off by several centimeters due to the noise in sensory feedback.

比如说，如果把一只手放在桌子底下，然后在桌子上面用另一只手去对准，最后位置可能相差好几厘米，这就是因为感官回馈里面的杂音在起作用。

Similarly, when you put motor output on movement output, it's extremely noisy.

同样，运动神经输出的肌肉动作和实际输出之间也是有很多杂音的。

Forget about trying to hit the bull's eye in darts, just aim for the same spot over and over again.

且不谈扔飞镖的时候瞄准靶心去扔，只看重复瞄准同一点的时候发生什么情况。

You have a huge spread due to movement variability.

由于每次动作都有差异，最后瞄准的结果会形成一片散点。

And more than that, the outside world, or task, is both ambiguous and variable.

更何况外界环境和要执行的任务常常模糊和变化着的。

The teapot could be full, it could be empty.

看这个茶壶，可能是满的，也可能是空的。

It changes over time. So we work in a whole sensory movement task soup of noise.

每次都不一样。所以我们其实是随时处在一大堆感官动作杂音环绕之中做动作的。

Now this noise is so great that society places a huge premium on those of us who can reduce the consequences of noise.

这种杂音相当厉害，以至于我们社会会给那些能有效减少杂音带来的后果的人巨额奖赏。

So if you're lucky enough to be able to knock a small white ball into a hole several hundred yards away using a long metal stick,

所以在座哪位能做到像老虎伍兹那样，用一根长金属杆把一个小白球打进几百米开外的洞里，

our society will be willing to reward you with hundreds of millions of dollars.

我们的社会愿意奖励你百万千万的钱。

Now what I want to convince you of is the brain also goes through a lot of effort to reduce the negative consequences of this sort of noise and variability.

好，我接下来想说明的是其实我们的大脑为了减少噪音和变化性的负面影响，也做了很多工作。

And to do that, I'm going to tell you about a framework which is very popular in statistics and machine learning of the last 50 years called Bayesian decision theory.

为此，我来介绍一个在过去50年里统计学和机器学习方面都很常用到的架构，叫做贝叶斯决策论。

And it's more recently a unifying way to think about how the brain deals with uncertainty.

近来这个理论常被用来从整体上理解大脑如何处理这种不确定性。

And the fundamental idea is you want to make inferences and then take actions.

基本思路是先做推断，然后做出动作。

So let's think about the inference. You want to generate beliefs about the world.

我们先来看推断。在推断中，我们要建立对于当前情景的“信念”。

So what are beliefs? Beliefs could be: where are my arms in space?

那么什么是信念？信念可以是我的胳膊在空间里的什么位置。

Am I looking at a cat or a fox?

也可以是我在看一只猫还是一只狐狸？

But we're going to represent beliefs with probabilities.

不过我们要把信念用概率来表示。

So we're going to represent a belief with a number between zero and one -- zero meaning I don't believe it at all, one means I'm absolutely certain.

所以在这里我们把所谓的信念表示成0到1之间的一个数--称之为置信值，0表示完全不相信，1表示完全确信。

And numbers in between give you the gray levels of uncertainty.

到1之间就表示不同灰度的不确定程度。

And the key idea to Bayesian inference is you have two sources of information from which to make your inference.

然后注意，贝叶斯推断的重点是我们靠两个信息源来做出推断。

You have data, and data in neuroscience is sensory input.

第一我们有数据，在神经科学里这个数据就是感官输入进来的内容。

So I have sensory input, which I can take in to make beliefs.

所以感官输入是其中一个信息源，用于生成我刚才说的“置信值 ”。

But there's another source of information, and that's effectively prior knowledge.

不过还有第二个信息源，事实上也就是先前的知识。

You accumulate knowledge throughout your life in memories.

因为我们在一生中在过去的记忆中积累知识。

And the point about Bayesian decision theory is it gives you the mathematics of the optimal way to combine your prior knowledge with your sensory evidence to generate new beliefs.

好了，贝叶斯决策论的重点就在于这个理论提供了一种计算方法，能找到最优的办法来整合，知识积累和感官输入这两种信息源以生成新的置信值。

And I've put the formula up there. I'm not going to explain what that formula is, but it's very beautiful.

我现在把公式放在这里。我不详细解释这个公式了，但是这个公式非常漂亮。

And it has real beauty and real explanatory power.

不仅有和谐的内在美，还有实实在在的说服力。

And what it really says, and what you want to estimate, is the probability of different beliefs given your sensory input.

这个公式真正的用途，也就是我们要估测的结果，是给出我们感官输入的情况下不同置信值出现的概率。

So let me give you an intuitive example.

现在我举一个直观的例子。

Imagine you're learning to play tennis, when you play tennis and you want to decide where the ball is going to bounce as it comes over the net towards you.

想象你现在在网球场练网球，当这个网球越过球网飞过来的时候，你要决定这个球落在哪里。

There are two sources of information Bayes' rule tells you.

根据贝叶斯的理论你现在有两个信息源。

There's sensory evidence -- you can use visual information auditory information, and that might tell you it's going to land in that red spot.

一个是感官输入--你的视觉和听觉收到的信息，告诉你球应该会落在图上的红点处。

But you know that your senses are not perfect,

但是，你也知道自己的感官并不完美，

and therefore there's some variability of where it's going to land shown by that cloud of red, representing numbers between 0.5 and maybe 0.1.

所以球的落地点可能会有误差--在图上用这块红色区域来表示这个区域的概率在0.5和大概0.1附近浮动。

That information is available in the current shot, but there's another source of information not available on the current shot,

以上这些信息是实时获得的，但是另一个获得信息的渠道，不是实时获得的，

but only available by repeated experience in the game of tennis,

而是来自只有通过反复练习网球才得来的经验，

and that's that the ball doesn't bounce with equal probability over the court during the match.

经验告诉你，这个网球在比赛过程中，不会以均等的概率落在球场里。

If you're playing against a very good opponent,

如果你的对手水平很高，

they may distribute it in that green area, which is the prior distribution, making it hard for you to return.

可能会让球的落点分布在图上的绿色区域，也就是所谓的“先验分布”，这些位置很难接到球。

Now both these sources of information carry important information.

那么现在，两个信息源都包含重要的信息。

And what Bayes' rule says is that I should multiply the numbers on the red by the numbers on the green to get the numbers of the yellow, which have the ellipses, and that's my belief.

根据贝叶斯的理论，我们应该把红色区域和绿色区域的数据相乘得到椭圆形的黄色区域，这就是我们的置信值。

So it's the optimal way of combining information.

并且这是整合信息的最佳方案。

I wouldn't tell you all this if it wasn't that a few years ago,

几年前我们考察过，人们学习新动作新技巧的时候，

we showed this is exactly what people do when they learn new movement skills.

真的在遵循这样的模式，所以我刚才才举了这个例子。

And what it means is we really are Bayesian inference machines.

这说明我们实际上都是天生的贝叶斯推断器。

As we go around, we learn about statistics of the world and lay that down,

在成长的过程中，我们不但学会了并记下了生活中的统计数据，

but we also learn about how noisy our own sensory apparatus is, and then combine those in a real Bayesian way.

也掌握了我们自己感官的杂音可能有多大，然后我们用贝叶斯法来整合处理这些数据。

Now a key part to the Bayesian is this part of the formula.

那么贝叶斯公式里很关键的是这一项--预测。

And what this part really says is I have to predict the probability of different sensory feedbacks given my beliefs.

这一项的意思是我们需要预测在原有置信值条件下，不同感官回馈的概率。

So that really means I have to make predictions of the future.

也就是说我们要对未来的可能做出预测。

And I want to convince you the brain does make predictions of the sensory feedback it's going to get.

我现在要说服各位相信一点，那就是大脑的确在对未来可能的感官回馈做出预测。

And moreover, it profoundly changes your perceptions by what you do.

并且你本身做了什么动作在很大程度上影响了你感知到的东西。

And to do that, I'll tell you about how the brain deals with sensory input.

为了说明这点，我来解释我们的大脑如何处理感官输入。

So you send a command out, you get sensory feedback back, and that transformation is governed by the physics of your body and your sensory apparatus.

我们先送出一个指令，然后得到一个感官回馈输入回来，之间的转换过程是由身体和感觉器官的物理过程完成的。

But you can imagine looking inside the brain. And here's inside the brain.

不过我们能想象大脑内部发生了什么。图上就是大脑内部。

You might have a little predictor, a neural simulator, of the physics of your body and your senses.

里面可能有一个预测结构，也就是神经组成的模拟器来模拟身体和感觉器官的物理过程。

So as you send a movement command down, you tap a copy of that off and run it into your neural simulator to anticipate the sensory consequences of your actions.

当一个动作指令发出来的时候，大脑复制一份这个指令，然后在这个神经模拟器上运行，来预测出这个动作带来的感官回馈结果。

So as I shake this ketchup bottle, I get some true sensory feedback as the function of time in the bottom row.

所以我往下磕这个番茄酱瓶子的时候，在下面那行里我得到真实的感官回馈，是个关于时间的函数。

And if I've got a good predictor, it predicts the same thing.

同时大脑里的预测结构如果准确的话，也预测出了同样的结果。

Well why would I bother doing that?

好了，那我为什么非要多此一举来预测呢？

I'm going to get the same feedback anyway. Well there's good reasons.

毕竟最终我总能得到同样的回馈。不过这样确实是有原因的。

Imagine, as I shake the ketchup bottle, someone very kindly comes up to me and taps it on the back for me.

想象我磕这个番茄酱瓶子的时候，台下有位热心观众过来帮我拍了下瓶底。

Now I get an extra source of sensory information due to that external act.

于是我多了一个感官信息的来源因为我多接受了这个外在的动作。

So I get two sources.

所以感官回馈现在有两个来源。

I get you tapping on it, and I get me shaking it, but from my senses' point of view, that is combined together into one source of information.

一个是你拍瓶底的动作，另一个是我磕瓶子的动作，但是从我的感觉来讲，这两个是合起来作为一个信息源的。

Now there's good reason to believe that you would want to be able to distinguish external events from internal events.

可是我们有理由相信，我们希望能够把外来事件和内在的事件区分开。

Because external events are actually much more behaviorally relevant than feeling everything that's going on inside my body.

因为实际上相对于我身体上发生的内在事件，外来事件在行为上才更有分析价值。

So one way to reconstruct that is to compare the prediction -- which is only based on your movement commands -- with the reality.

所以区分开来的一种办法就是把我的预测结果--因为这个结果只基于我的动作指令--和真实的感官反馈做比较。

Any discrepancy should hopefully be external.

希望两者的差值应该就是外力的结果。

So as I go around the world, I'm making predictions of what I should get, subtracting them off.

所以我和外界接触的时候，我预测出我自己的动作可能得到什么回馈，减掉这些。

Everything left over is external to me.

剩下的其他部分就是外界对我的产生的作用。

What evidence is there for this?

那有什么证据支持这点吗？

Well there's one very clear example where a sensation generated by myself feels very different then if generated by another person.

我们认为其中一个非常明了的例子就是自身产生的动作带来的感觉与他人做同样动作带来的感觉是十分不同的。

And so we decided the most obvious place to start was with tickling.

我们认为最直接的着手点就是挠痒痒这件事。

It's been known for a long time, you can't tickle yourself as well as other people can.

众所周知，人们挠自己的时候总是不如别人挠自己的时候痒。

But it hasn't really been shown, it's because you have a neural simulator, simulating your own body and subtracting off that sense.

但是没有被证明为什么，其实是因为每个人自己有一个神经模拟器模拟出自己动作带来的效果。

So we can bring the experiments of the 21st century by applying robotic technologies to this problem.

然后在感觉里自动减去这个效果，所以我们可以用21世纪的实验条件把机器人技术应用到这个问题上。

And in effect, what we have is some sort of stick in one hand attached to a robot,

具体办法是，我们让试验者一只手握着一根连接在机器人上的小棍，

and they're going to move that back and forward.

由手带动一起前后移动。

And then we're going to track that with a computer and use it to control another robot, which is going to tickle their palm with another stick.

这时候我们用计算机追踪这个移动模式，然后用同样的模式带动另一个机器人，去用另一根小棍挠试验者的另一只手的手心。

And then we're going to ask them to rate a bunch of things including ticklishness.

之后，我们会让试验者给一系列指标打分，也包括痒痒的程度。

I'll show you just one part of our study.

在这里我展示我们实验的其中一部分。

And here I've taken away the robots, but basically people move with their right arm sinusoidally back and forward.

图上没有画那两个机器人，基本上人们的右臂是以类正弦的方式前后移动的。

And we replay that to the other hand with a time delay.

这时候我们把这个移动加上某个时间延迟，在另一只手上重放。

Either no time delay, in which case light would just tickle your palm, or with a time delay of two-tenths of three-tenths of a second.

延迟可以是0，就像直接挠手心一样，也可以是0.1秒、0.2秒、0.3秒这样的延迟。

So the important point here is the right hand always does the same things -- sinusoidal movement.

这里的重点是试验者的右手一直做同样的类正弦动作。

The left hand always is the same and puts sinusoidal tickle.

左手一直在被同样的类正弦动作挠着。

All we're playing with is a tempo causality.

唯一我们改变的就是因与果之间的步调。

And as we go from naught to 0.1 second, it becomes more ticklish.

当我们把延迟从0加到0.1秒的时候，感觉上越来越痒。

As you go from 0.1 to 0.2, it becomes more ticklish at the end.

再从0.1到0.2秒，越到后来越痒。

And by 0.2 of a second, it's equivalently ticklish to the robot that just tickled you without you doing anything.

到了0.2秒的时候发痒的程度已经等同于自己什么也不做，机器人直接来挠的效果了。

So whatever is responsible for this cancellation is extremely tightly coupled with tempo causality.

因此不管是什么造成了自身动作的抵消，一定与因果之间的步调大有联系。

And based on this illustration, we really convinced ourselves in the field that the brain's making precise predictions and subtracting them off from the sensations.

受这个实验的启发，我们终于相信了那个论点，也就是大脑能够做出准确预测并将预测结果从实体感觉里减去。

Now I have to admit, these are the worst studies my lab has ever run.

现在我要承认，这个是我实验室里进行过的最糟糕的实验。

Because the tickle sensation on the palm comes and goes, you need large numbers of subjects with these stars making them significant.

因为痒痒的感觉就像昙花一现，所以需要大量的试验者，还只能靠星级来给痒痒的程度打分。

So we were looking for a much more objective way to assess this phenomena.

所以我们需要一个更客观的办法来评估这个现象。

And in the intervening years I had two daughters.

其间几年中，我有了两个女儿。

And one thing you notice about children in backseats of cars on long journeys,

于是关于长途旅行里小孩子坐在后座上，你注意到一件事，

they get into fights -- which started with one of them doing something to the other, the other retaliating.

那就是他们会互相掐架--开始于其中一个对另一个动了什么手脚，另一个就要还手。

It quickly escalates. And children tend to get into fights which escalate in terms of force.

然后冲突很快就会升级。而小孩子们之间的冲突升级一般都体现在用力大小上。

Now when I screamed at my children to stop, sometimes they would both say to me the other person hit them harder.

于是我嚷嚷着叫我的孩子们住手的时候，有的时候她们俩都会告诉我，对方下手比自己更重。

Now I happen to know my children don't lie, so I thought, as a neuroscientist,

我恰好知道我的孩子们不会说假话，所以作为一个神经学家，

and it was important how I could explain how they were telling inconsistent truths.

我想我如何解释两人自相矛盾的真话应该是很重要的吧。

And we hypothesize based on the tickling study that when one child hits another, they generate the movement command.

于是我们根据挠痒痒的实验做出了假设，就是当一个孩子打另一个的时候，会生成动作命令。

They predict the sensory consequences and subtract it off.

他们预测出了感知到的结果，然后减掉了。

So they actually think they've hit the person less hard than they have -- rather like the tickling.

实际上他们以为他们下手比真正要轻，和挠痒的道理类似。

Whereas the passive recipient doesn't make the prediction, feels the full blow.

然而那个被打的孩子没有做预测，感觉到了全部的打击。

So if they retaliate with the same force, the first person will think it's been escalated.

所以他们用同样这个打击的力气还手的话，第一个人会觉得力气变大，冲突升级了。

So we decided to test this in the lab. Now we don't work with children, we don't work with hitting, but the concept is identical.

所以我们决定在实验室里做这个测试，好了我们不用小孩子，也不用打的，不过概念是相同的。

We bring in two adults. We tell them they're going to play a game.

我们找来两个成人。告诉他们来做个小游戏。

And so here's player one and player two sitting opposite to each other.

这里显示了玩家一和玩家二面对面坐着。

And the game is very simple. We started with a motor with a little lever, a little force transfuser.

游戏很简单。开始的时候这里有个电机，连着小杠杆，作为传力器。

And we use this motor to apply force down to player one's fingers for three seconds and then it stops.

小电机向下对玩家一的手指施力持续3秒钟后停止。

And that player's been told, remember the experience of that force and use your other finger to apply the same force down to the other subject's finger through a force transfuser -- and they do that.

我们告诉这名玩家一，记住这个力的感觉，再用另一个手指以同样的力，通过传力器下压玩家二的手指--玩家一这么做了。

And player two's been told, remember the experience of that force.

然后玩家二被告知记住这个力的感觉。

Use your other hand to apply the force back down.

用另一只手把同样的力施加回来。

And so they take it in turns to apply the force they've just experienced back and forward.

这两个玩家轮流来回施加方才所受的力。

But critically, they're briefed about the rules of the game in separate rooms.

不过关键的一点是这两个玩家是在不同房间里被介绍游戏规则的。

So they don't know the rules the other person's playing by.

所以他们互相不知道对方遵循的规则。

And what we've measured is the force as a function of terms.

而我们测量的是力的大小关于回合次数的函数。

And if we look at what we start with, a quarter of a Newton there, a number of turns, perfect would be that red line.

我们来看图，开始的时候力是四分之一牛顿，横轴是回合的次数，完美的结果应该是这条红线。

And what we see in all pairs of subjects is this -- a 70 percent escalation in force on each go.

但是我们在所有实验配对中看到的都是这个--每一次施力中大小上扬70%。

So it really suggests, when you're doing this -- based on this study and others we've done -- that the brain is canceling the sensory consequences and underestimating the force it's producing.

所以确实表明，我们在运动的时候--根据这个实验和我们所做的其他实验--大脑一直在抵消感官带来的结果。

So it re-shows the brain makes predictions and fundamentally changes the precepts.

而低估自己产生的力，所以再次说明大脑会做出预测，并且在根本上改变感知到的内容。

So we've made inferences, we've done predictions, now we have to generate actions.

到现在为止，我们做了推断，也做了预测，现在我们要生成动作了。

And what Bayes' rule says is, given my beliefs, the action should in some sense be optimal.

贝叶斯法则说的是，基于置信值动作应该在某种意义上是最优的。

But we've got a problem. Tasks are symbolic -- I want to drink, I want to dance -- but the movement system has to contract 600 muscles in a particular sequence.

但在这里出现问题了。动作任务是符号性质的--我要喝水，我要跳舞--但是运动系统需要按特定顺序收缩600块肌肉。

And there's a big gap between the task and the movement system.

况且有一个巨大的鸿沟横在动作任务和运动系统之间。

So it could be bridged in infinitely many different ways.

意思是我们有无穷多种方式越过这个鸿沟。

So think about just a point to point movement.

来考虑仅仅是点到点的移动。

I could choose these two paths out of an infinite number of paths.

我可以在无穷条可能的路径中选出这两条。

Having chosen a particular path, I can hold my hand on that path as infinitely many different joint configurations.

选定一条特定路径之后，我可以把手放在这条路径上，但是又有无穷多种关节构型。

And I can hold my arm in a particular joint configuration either very stiff or very relaxed.

另外我把的胳膊固定在某一种构型之后，我可以紧张，也可以松弛。

So I have a huge amount of choice to make.

所以我有非常多的选择余地。

Now it turns out, we are extremely stereotypical. We all move the same way pretty much.

不过结果是，我们都特别典型。都基本上用同一种方式移动。

And so it turns out we're so stereotypical, our brains have got dedicated neural circuitry to decode this stereotyping.

我们典型到我们的大脑已经开辟出特定的神经回路，来解码这类典型。

So if I take some dots and set them in motion with biological motion, your brain's circuitry would understand instantly what's going on.

因此如果我拿一些点，然后让它们按生物动作形式来运动，大家的大脑回路会立即知道怎么回事。

Now this is a bunch of dots moving.

好了，这仅仅是一堆点在动。

You will know what this person is doing, whether happy, sad, old, young -- a huge amount of information.

我们就知道这个人在做什么，是喜还是悲，是老还是少--很大量的信息。

If these dots were cars going on a racing circuit, you would have absolutely no idea what's going on.

如果这些点是赛车在赛道上绕圈，我们对发生了什么就完全没有概念了。

So why is it that we move the particular ways we do?

那我们为什么用这种特定的模式移动呢？

Well let's think about what really happens. Maybe we don't all quite move the same way.

来想想真实情况可能是什么。假如我们并不都用同一种模式移动。

Maybe there's variation in the population.

假如人群中有差异。

And maybe those who move better than others have got more chance of getting their children into the next generation.

而且假如那些比其他人移动得更好的个体有更多机会让他们的子代进入下一代的群体里面。

So in evolutionary scales, movements get better.

那么在进化级别上，动作越来越优化。

Perhaps in life, movements get better through learning.

另外，也许在一生当中，学习会让动作更优化。

So what is it about a movement which is good or bad?

那么，是什么决定了动作的好与坏呢？

Imagine I want to intercept this ball.

来想象我要截下这个球。

Here are two possible paths to that ball.

这里有两种路径来做到。

Well if I choose the left-hand path, I can work out the forces required in one of my muscles as a function of time.

假如我选了左手边这条路径，我可以算出某一块肌肉需要施多大力，用关于时间的函数表示。

But there's noise added to this.

但是还有杂音附加在上面。

So what I actually get, based on this lovely, smooth, desired force, is a very noisy version.

所以根据这条美好、光滑、期望中的曲线，我事实上得到的是个杂音很大的版本。

So if I pick the same command through many times, I will get a different noisy version each time, because noise changes each time.

那如果我多次发布同样的动作指令，每次会得到不同的含杂音的版本，因为每次杂音不一样。

So what I can show you here is how the variability of the movement will evolve if I choose that way.

所以我现在在这儿展示的只是这个动作的可变性有多大，前提是如果我选这种方式的话。

If I choose a different way of moving -- on the right for example -- then I'll have a different command, different noise, playing through a noisy system, very complicated.

如果我选另一种方式--比如右边这个--那我就会有不同的指令，不同的杂音，还要透过杂音系统，总之非常复杂。

All we can be sure of is the variability will be different.

我们唯一能确定的是，两个方式的可变性会不同。

If I move in this particular way, I end up with a smaller variability across many movements.

要是我用这种方式运动的话，最后得到的可变性在很多种动作里面是最小的。

So if I have to choose between those two, I would choose the right one because it's less variable.

所以要是我在这两者里面选其一，我就会选右边那个，因为可变性更小。

And the fundamental idea is you want to plan your movements so as to minimize the negative consequence of the noise.

话说回来，我们的基本思路是要找到一种动作方式，能够把杂音带来的负面影响降到最低。

And one intuition to get is actually the amount of noise or variability I show here gets bigger as the force gets bigger.

并且我们凭直觉就知道我在这里展示的杂音，或者说可变性随力的大小的增加而增加。

So you want to avoid big forces as one principle.

所以我们的原则之一就是避免施加太大的力。

So we've shown that using this, we can explain a huge amount of data -- that exactly people are going about their lives planning movements so as to minimize negative consequences of noise.

用这个原则，我们就能解释很多得到的数据--那就是人们在生活中有意安排动作方式来把杂音带来的负面影响降到最低。

So I hope I've convinced you the brain is there and evolved to control movement.

现在，我想我已经让大家相信大脑之所以存在是为了控制动作。

And it's an intellectual challenge to understand how we do that.

而且了解怎样做到这点是需要费一番脑筋的。

But it's also relevant for disease and rehabilitation.

不过同样具有相关价值的是在疾病和康复方面。

There are many diseases which effect movement.

有很多病症有对动作产生影响。

And hopefully if we understand how we control movement, we can apply that to robotic technology.

所以如果有朝一日掌握了人类控制动作的机制，我们就可以用到机器人技术中去。

And finally, I want to remind you, when you see animals do what look like very simple tasks,

最后我要提醒大家，当你们看见动物做些看似非常简单的动作的时候，

the actual complexity of what is going on inside their brain is really quite dramatic.

它们运行在大脑内部的真正复杂度其实是相当令人叹为观止的。

Thank you very much.

谢谢大家。

Quick question for you, Dan. So you're a movement -- (DW: Chauvinist.) -- chauvinist.

我问一个简短的问题。所以你是一个运动--（沃：沙文主义者。）--沙文主义者。

Does that mean that you think that the other things we think our brains are about -- the dreaming, the yearning,

是不是意味着你认为我们觉得让大脑有存在意义的其他事情--那些梦想、渴望、

the falling in love and all these things -- are a kind of side show, an accident?

恋爱等等一切东西都是些余兴节目或者意外收获呢？

No, no, actually I think they're all important to drive the right movement behavior to get reproduction in the end.

不，不，其实我认为这些也很重要，它们能促使合适的动作行为，以达到繁殖的最终目的。

So I think people who study sensation or memory without realizing why you're laying down memories of childhood.

所以我觉得有些人在研究感觉或者记忆，却没意识到为什么我们要记下童年的回忆。

The fact that we forget most of our childhood, for example, is probably fine, because it doesn't effect our movements later in life.

我们会忘记大部分幼年的记忆这个事实，比如说其实应该无关紧要，因为那些不会影响我们长大以后的动作。

You only need to store things which are really going to effect movement.

我们只需要存储能对动作产生真正影响的事情就行了。

So you think that people thinking about the brain, and consciousness generally, could get real insight by saying, where does movement play in this game?

那你是否认为人们研究大脑，以及更笼统的意识，如果要获得真正的领悟，就要问动作在这里起了什么作用？

So people have found out for example that studying vision in the absence of realizing why you have vision is a mistake.

人们已经发现，比如说在不明确视觉何以存在的时候，就去研究视觉是个失误。

You have to study vision with the realization of how the movement system is going to use vision.

研究视觉的时候，我们必须先意识到运动系统要怎样用这个视觉。

And it uses it very differently once you think about it that way.

这样想的时候用法就不一样了。

Well that was quite fascinating. Thank you very much indeed.

哦，确很有趣。非常感谢你。