Nova now how does the brain work

A few years back, he embarrassed President Jimmy Carter's Secret Service agents when he picked their pockets during a visit to Las Vegas.

Today, Apollo has agreed to share some of his secrets with me. First, he shows me a special motion he uses to distract his victims when he's picking their pockets. They're hoping he can help them solve a fundamental mystery: how does the brain decide what to pay attention to? We know where the visual centers are; we know where the auditory centers are. But we don't really have a very good idea about attention and awareness yet.

Back in their lab, they prepare test subjects to watch it. They fit them with this contraption, equipped with tiny cameras aimed at their subjects' eyes. And what we're analyzing is: where are the eyes at every given moment of time in comparison to what's being presented on the screen? What my eyes are doing right now is smooth pursuit. A smooth pursuit allows you to track a moving target. When our eyes see an object, the light reflected from its surface travels to the retina where it's transformed into neural signals.

These signals go to a part of the brain dedicated to vision. Here we start to form an image. But we don't pay attention to everything we look at. How does the brain control what we focus on and what we don't? In a new study, Susana and Stephen, working with Jose Manuel Alonso, found that when our eyes track something like Apollo's curved motion, there's more than one type of brain cell at work.

One type of cell detects motion, while the other suppresses the background. You see the light glint off. Why did I see something that didn't actually happen? Back at the lab, when volunteers watch the trick on a monitor, they're stumped, just as I am. So you see this motion that didn't actually take place. It's a survival mechanism. This time, it's the magicians who are asking the neuroscientists to explain a trick.

It's one of their favorites: where they make balls appear and disappear under plastic cups. Here's a little variation Teller came up with, where he takes the ball, places it in his hand and shows you underneath the cup, yet it still appears underneath the cup. You take that center ball, place it visibly in the center cup, these side balls, we put them away, we don't need them anymore.

We have three balls underneath here, and that's what we regroup: a giant ball underneath the center cup, one more giant ball either side, and, of course, for the finish, it's an American baseball, right there. He's so curious he's agreed to give an interview, provided we don't actually show him speaking. To him, it's obvious. The neuroscientists record him performing this move and show it to their volunteers. Then they show a different version, one that block's Teller's face.

That turns out to make a real difference. Like the curved hand motion in a coin trick, the magician's face commands your attention just enough to distract you from what's really going on. TELLER: If you come to a magic show with the intention of exercising your ability to discern fact from fancy and you fail, that's a fine piece of entertainment. So it's this wonderful playground where you can sort of relax and go, "Oh boy, it's really hard to understand the world. What's fascinating about our work is that we are a study of human nature, of human behavior, and we have certain information that has been passed down through generations that can be utilized in a way that interfaces with science.

And I'm really excited about the collaboration. And for some tasks, they've definitely got the advantage. In fact, they've beaten our top human chess champions. I mean go. Not go away, I mean make your move! JAKE WARD Correspondent : It's one of the most powerful ideas dominating science fiction: to build a machine with mental powers equal to or exceeding our own, to create artificial intelligence, And you?

We had better hope not, because, A. Things are getting smarter, and it's really hard for us to imagine how different it's going to be. JAKE WARD: It's now five years since this pack of driverless vehicles raced each other across the Mojave Desert, 13 years since a chess-playing computer defeated the world's top chess-playing human, and you probably interacted with an intelligent machine today.

You no longer have to go visit the neighborhood bookstore. Today, Amazon, super computers and sophisticated software can predict, with uncanny accuracy, what you like. In my case, it's a couple of history books and a lot of comic books. In fact, you can think of humans as natural language processing machines.

They can be implicit, they can be ambiguous, they use context. But a computer can't relate the words to human experience, they're really just symbols.

So, it's a real challenge for the computer to take lots of different pieces of language and try to figure out, "Gee, does this mean that? I have to test my voice to make sure I can be heard and that I am not speaking too loudly. For the past four years they've been designing this supercomputer to answer questions on, well, just about anything.

Practice Host It evolves; there's a huge variety of questions. So, it's kind of the benchmark, I think, for knowledge of a wide variety of topics and, also, being able to think fast. We're not going to do this. Too difficult to do; we'll embarrass ourselves. Did I expect to get fired? No, but What I more expected was, you know, you're going to lose credibility.

Practice Host : Cincinnati's daily paper has this name. Practice Host : Maria? Practice Contestant : What is hazlenuts? Practice Host : This trusted friend was the first non-dairy powdered creamer. Practice Host : What? Holy, no, that was wrong. We had a lot of problems. But we're learning a lot about how to make WATSON better at understanding language in different contexts and to understand in one context the word means one thing and in another it means another thing.

Practice Host : A garment worn by a child, perhaps aboard an operatic ship. Practice Host : Yes. How did you get that? Very nicely done. It can store an estimated 1,, gigabytes of information. By comparison, the most commonly used archive of the Internet is only four and a half times bigger. Instead, he's been stuffed with millions of documents: anthologies, dictionaries, encyclopedias. But no amount of information is ever enough, because amassing facts is only half the battle, since Jeopardy!

So it can be very challenging just for the computer to understand what it's being asked for, much less trying to answer it. Moreover, it was sort of my philosophical view that that's not the right way to approach A. And it turns out, from one document we were finding this association with a painter; from this other document about dragons, we were finding out that Saint George slays dragons.

Refusing to imprison this man for demonstrating during the s, De Gaulle said, "One does not arrest Voltaire. Jake did you come up with the right answer? Let's take a look on your board. Who is Jean Paul Sartre? That is correct. I think, at that level, it does. As soon as he receives a clue, they begin searching through millions of documents, independently gathering evidence, seeking out matches for the clue's words and phrases.

Once it makes that assumption, it takes each one and says, "Let me go and try to get evidence supporting this answer as the right one. I like popularity evidence. I like classification evidence. I want to buzz in. It's like a second language to me.

It's how it acts in the world doing interesting things and stuff that we want done, and it's a total interaction. Imagine a WATSON-like medical assistant that could associate a set of symptoms with diagnoses a doctor might never have heard of; or a WATSON that could gather evidence for a lawyer; a thinking machine that could, in short, augment us as a sort of second brain. We can learn from birds and build s; we can learn from brains and build useful artifacts. You've just tried to add this video to My List.

But first, we need you to sign in to PBS using one of the services below. You've just tried to add this show to My List. By creating an account, you acknowledge that PBS may share your information with our member stations and our respective service providers, and that you have read and understand the Privacy Policy and Terms of Use.

We can remove the first video in the list to add this one. We can remove the first show in the list to add this one. Clip: Season 5 Episode 3 52m 57s Video has closed captioning. Investigate the psychology of magic tricks, magnetic wands, artificial intelligence, and more. Problems Playing Video? Report a Problem Closed Captioning. Before you submit an error, please consult our Troubleshooting Guide.

Your report has been successfully submitted. Thank you for helping us improve PBS Video. Nova ScienceNow. See more at IMDbPro. Photos Add Image.

Top cast Edit. Alex Liu Self as Self. Storyline Edit. Add content advisory. Did you know Edit. Connections Features A Space Odyssey User reviews 1 Review.