Mary Lou Jepsen: Could future devices read images from our brains?

譯者: Thunder Song 審譯者: Regina Chu

I had brain surgery 18 years ago,

18 年前我動了大腦手術，

and since that time, brain science has become

從那個時候開始，大腦科學變成了

a personal passion of mine.

我的一個個人愛好。

I'm actually an engineer.

我其實是個工程師。

And first let me say, I recently joined

首先我想說，我最近加入了

Google's Moonshot group,

谷歌的「科技登月」計畫， （Google X Moonshots）

where I had a division,

在那裏我有個部門，

the display division in Google X,

谷歌實驗室的顯示器部門，

and the brain science work I'm speaking about today

今天我要說的大腦科學，

is work I did before I joined Google

是我在加入谷歌之前的工作，

and on the side outside of Google.

以及不涉及谷歌的方面的事情。

So that said, there's a stigma

所以有這麼一個汙名，

when you have brain surgery.

就是當妳進行了大腦手術，

Are you still smart or not?

妳是不是依舊聰明？

And if not, can you make yourself smart again?

如果不是，妳能再次變得聰明嗎？

After my neurosurgery,

在我的神經外科手術後，

part of my brain was missing,

我大腦的一部份遺失了，

and I had to deal with that.

而且我需要解決這個問題。

It wasn't the grey matter, but it was the gooey part dead center

那一部份並不是灰質， 而是那個黏糊糊的死點，

that makes key hormones and neurotransmitters.

生產關鍵的荷爾蒙 和神經傳導物質的部分。

Immediately after my surgery,

我的手術一結束，

I had to decide what amounts of each of over

我需要決定

每天攝入多少那些強效的化學藥品。

a dozen powerful chemicals to take each day,

because if I just took nothing,

因為如果我什麽都不做，

I would die within hours.

那我會在幾個小時以後死去

Every day now for 18 years -- every single day --

18 年來的每一天，真的是每一天，

I've had to try to decide the combinations

我不得不去決定如何組合 並且混合各種化學藥品，

and mixtures of chemicals,

and try to get them, to stay alive.

同時我也要獲取它們， 以便我能維持生命。

There have been several close calls.

曾經有幾次我已經幾乎面對死亡。

But luckily, I'm an experimentalist at heart,

但幸運地是， 我的本質是一個實驗者，

so I decided I would experiment

所以我決定我要通過實驗

to try to find more optimal dosages

去嘗試找出最佳的劑量，

because there really isn't a clear road map

因為這之前並沒有 一個清晰明確指導

on this that's detailed.

關於各種細節的介紹。

I began to try different mixtures,

我開始嘗試不同的混合，

and I was blown away by how

我已經強烈的感受到

tiny changes in dosages

一點點的劑量變化

dramatically changed my sense of self,

會戲劇性的改變我的 自我感知能力，

my sense of who I was, my thinking,

我對我自己的感受，我的思考，

my behavior towards people.

以及對待別人的行為。

One particularly dramatic case:

我有一個非常戲劇性的例子，

for a couple months I actually tried dosages

在幾個月的時間中我真的試驗了

and chemicals typical of a man in his early 20s,

給 20 歲出頭的男性 所用的藥品及劑量。

and I was blown away by how my thoughts changed.

而我被我自己的想法 因此改變而震驚。

(Laughter)

（笑~）

I was angry all the time,

那時我一直很生氣，

I thought about sex constantly,

我時常會想到性，

and I thought I was the smartest person

並且我覺得自己是世界上

in the entire world, and

是最聰明的人，

—(Laughter)—

（笑~）

of course over the years I'd met guys kind of like that,

當然這幾年我遇到了 一些很類似的傢伙

or maybe kind of toned-down versions of that.

或者說比那種稍微低調一點的人。

I was kind of extreme.

我其實是有點極端。

But to me, the surprise was,

但是對我而言，令我震驚的是

I wasn't trying to be arrogant.

我並不是去嘗試變得傲慢。

I was actually trying,

我實際上是在嘗試

with a little bit of insecurity,

帶著一點不安

to actually fix a problem in front of me,

去解決擺在我面前的問題，

and it just didn't come out that way.

只是結果總是不像我所設想的一樣。

So I couldn't handle it.

所以我並沒有辦法解決。

I changed my dosages.

我改變了我用藥的劑量。

But that experience, I think, gave me

但是那段經歷，我認為，讓我對於

a new appreciation for men

男人以及他們所經歷的一切 有了新的評價，

and what they might walk through,

and I've gotten along with men

所以從那時開始，

a lot better since then.

我和他們相處地更好了。

What I was trying to do

我想嘗試的是

with tuning these hormones

去調節這樣的一些荷爾蒙，

and neurotransmitters and so forth

以及神經傳導物質等等，

was to try to get my intelligence back

去恢復我的智力水平，

after my illness and surgery,

在我生病並且手術之後，

my creative thought, my idea flow.

以及我那些我充滿創意的想法。

And I think mostly in images,

並且我是以圖像思考的人，

and so for me that became a key metric --

所以對我而言這是一個重要的指標，

how to get these mental images

如何獲得這些心智圖像，

that I use as a way of rapid prototyping,

你也可以這樣說，讓我快速建立

if you will, my ideas,

我的想法原型的圖像，

trying on different new ideas for size,

嘗試各種新想法，

playing out scenarios.

推演不同的情節。

This kind of thinking isn't new.

這種想法並不新鮮。

Philiosophers like Hume and Descartes and Hobbes

哲學家像休謨、笛卡兒以及霍布斯，

saw things similarly.

也有著相同的想法。

They thought that mental images and ideas

他們覺得心智圖像以及想法

were actually the same thing.

其實是一樣的東西。

There are those today that dispute that,

現今有很多懷疑這樣的觀點，

and lots of debates about how the mind works,

也有很多關於心智如何工作的辯論，

but for me it's simple:

但是對我而言，這很簡單：

Mental images, for most of us,

心智圖像，對大多數我們而言，

are central in inventive and creative thinking.

其實是發明以及創造性想法的中心。

So after several years,

所以很多年後，

I tuned myself up and I have lots of great,

我調整好了我自己， 並且擁有了很多很好、

really vivid mental images with a lot of sophistication

非常生動，有著各種複雜的心智圖像，

and the analytical backbone behind them.

並且有很多分析支撐著這樣的觀點。

And so now I'm working on,

我現在開始研究，

how can I get these mental images in my mind

我如何才能將我大腦裡的心智圖像

out to my computer screen faster?

更快速的顯示在我的電腦螢幕上？

Can you imagine, if you will,

你能夠想像，這麼說吧，

a movie director being able to use

一個電影的導演能夠只用

her imagination alone to direct the world in front of her?

她大腦的想像力去操縱 她眼前的世界的想法？

Or a musician to get the music out of his head?

或者一個音樂家 直接從頭裡取得音樂？

There are incredible possibilities with this

這種方式創造出不可思議的可能性

as a way for creative people

對於那些有創造性的人們而言

to share at light speed.

他們可以以光速來分享他們的想法。

And the truth is, the remaining bottleneck

但事實上，有待突破的瓶頸是，

in being able to do this

想實現這樣的效果

is just upping the resolution of brain scan systems.

必須提高腦部斷層掃描的解析度。

So let me show you why I think we're pretty close to getting there

現在讓我展示爲什麽我認為 我們幾乎已經成功了，

by sharing with you two recent experiments

通過我們最近的兩個實驗，

from two top neuroscience groups.

來自我們兩個頂尖的神經學小組。

Both used fMRI technology --

他們都採用了 fMRI 技術，

functional magnetic resonance imaging technology --

也就是功能性核磁共振造影

to image the brain,

去給大腦成像，

and here is a brain scan set from Giorgio Ganis

這裡有一組從哈佛的喬治·甘尼斯

and his colleagues at Harvard.

以及他的同事得到的大腦掃描圖。

And the left-hand column shows a brain scan

左手邊的一列顯示了大腦的掃描圖，

of a person looking at an image.

是一個人正在看一張圖片的時候掃描的。

The middle column shows the brainscan

中間這一列顯示的大腦掃描圖

of that same individual

是同樣的一個人

imagining, seeing that same image.

去想像那張圖片的時候掃描的。

And the right column was created

右邊的這一列圖則是

by subtracting the middle column from the left column,

用左邊的圖減去中間的圖的結果，

showing the difference to be nearly zero.

你可以發現幾乎沒有差別。

This was repeated on lots of different individuals

已經在很多人身上做了這個實驗

with lots of different images,

得到大量不同的圖片，

always with a similar result.

都得到了相同的結果。

The difference between seeing an image

看一張圖以及

and imagining seeing that same image

想像同一張圖

is next to nothing.

大腦幾乎沒有變化。

Next let me share with you one other experiment,

接下來讓我分享另外的一個實驗，

this from Jack Gallant's lab at Cal Berkeley.

來自柏克萊的傑克·格蘭特的實驗室

They've been able to decode brainwaves

他們能夠解密腦電波，

into recognizable visual fields.

轉化成可以辨認的視場。

So let me set this up for you.

讓我來為你解釋這個。

In this experiment, individuals were shown

在這個試驗中，我們給每個人

hundreds of hours of YouTube videos

播放了上百小時 YouTube 的影片

while scans were made of their brains

同時掃描著他們的大腦，

to create a large library of their brain reacting

並且根據大腦的反應 建立一個龐大的數據庫，

to video sequences.

依照影片的順序。

Then a new movie was shown with new images,

然後我們播放有著新圖像，

new people, new animals in it,

新人物，新動物的新電影，

and a new scan set was recorded.

並且記錄了新的一系列掃描結果。

The computer, using brain scan data alone,

而電腦則使用之前的 大腦掃描數據庫

decoded that new brain scan

對這個新的大腦掃描圖解碼，

to show what it thought the individual was actually seeing.

並且展示電腦認為 人們看見的圖像。

On the right-hand side, you see the computer's guess,

在右手邊，你可以看見電腦的猜測，

and on the left-hand side, the presented clip.

左手邊則是我們提供的影片片段。

This is the jaw-dropper.

這樣的結果是令人震驚的。

We are so close to being able to do this.

我們就要成功的實現這樣的技術，

We just need to up the resolution.

我們要做的只是去提高解析度。

And now remember that when you see an image

現在請記住當你看到一張圖片

versus when you imagine that same image,

同時當你想像同樣的圖片的時候，

it creates the same brain scan.

它會產生相同的大腦變化。

So this was done with the highest-resolution

這是用現今可得的最高解析度

brain scan systems available today,

大腦掃描系統得到的結果

and their resolution has increased really

它們的解析度最近幾年都是

about a thousandfold in the last several years.

以千倍的速度增長。

Next we need to increase the resolution

下一步我們需要再一次

another thousandfold

以千倍的速度提高解析度，

to get a deeper glimpse.

以獲得更深層次的變化。

How do we do that?

那我們應該如何做呢？

There's a lot of techniques in this approach.

現在有很多科學方法 可以用於實現。

One way is to crack open your skull and put in electrodes.

一種方法是打開 你的腦殼並放入電極。

I'm not for that.

我並不支持這種方法。

There's a lot of new imaging techniques

現在有很多新的成像技術

being proposed, some even by me,

正在開發，有些則是由我負責，

but given the recent success of MRI,

想到最近 MRI（核磁共振成像） 的成功，

first we need to ask the question,

首先我們要問的是，

is it the end of the road with this technology?

這項科技已經走到了盡頭嗎？

Conventional wisdom says the only way

傳統意義上講，只有採用更大的磁體

to get higher resolution is with bigger magnets,

才能產生更高解析度的圖片，

but at this point bigger magnets

但是所謂更大的磁體

only offer incremental resolution improvements,

只能夠持續改進成像解析度，

not the thousandfold we need.

而不是我們所需要的千倍提高。

I'm putting forward an idea:

我們有這樣的一個想法：

instead of bigger magnets,

不要使用大型的磁體，

let's make better magnets.

而是採用更好的磁體。

There's some new technology breakthroughs

現在有一些技術 能夠取得突破性的進展，

in nanoscience

在納米科技中，

when applied to magnetic structures

當對磁體的結構進行改造，

that have created a whole new class of magnets,

就能產生一個全新級別的磁體，

and with these magnets, we can lay down

當使用這些磁體的時候，我們可以

very fine detailed magnetic field patterns

在大腦中產生 非常精細的磁場圖案

throughout the brain,

and using those, we can actually create

並且通過這樣的技術，我們可以產生

holographic-like interference structures

如全像攝影般的干涉結構，

to get precision control over many patterns,

便能精確地控制不同的圖案，

as is shown here by shifting things.

就像這裡通過移動 而展示出來的效果。

We can create much more complicated structures

我們可以創造更多複雜的結構，

with slightly different arrangements,

通過一點點不同的排列安排

kind of like making Spirograph.

就像製作呼吸運動記錄器。

So why does that matter?

爲什麽這個這麼重要？

A lot of effort in MRI over the years

這些年來在 MRI（核磁共振成像）的努力，

has gone into making really big,

已經取得了巨大的成果，

really huge magnets, right?

非常大的磁體，對不對？

But yet most of the recent advances

但是近些年來在解析度上的進步

in resolution have actually come from

實際上來自

ingeniously clever encoding and decoding solutions

在 MRI （核磁共振成像）系統中

in the F.M. radio frequency transmitters and receivers

對調頻信號頻率的傳輸和接收

in the MRI systems.

巧妙的加密和解密的方法。

Let's also, instead of a uniform magnetic field,

同樣的，我們不採用標準的磁場，

put down structured magnetic patterns

而是記下有結構磁性的圖案

in addition to the F.M. radio frequencies.

以及所對應的調頻信號頻率。

So by combining the magnetics patterns

通過結合磁性的圖案

with the patterns in the F.M. radio frequencies

以及調頻信號頻率中產生的圖案，

processing which can massively increase

一次掃描所提取的資訊 取得了巨大的進展

the information that we can extract

in a single scan.

在僅僅的一次掃描中。

最重要的是，我們可以將

And on top of that, we can then layer

our ever-growing knowledge of brain structure and memory

我們持續增長的 大腦結構和記憶分層，

to create a thousandfold increase that we need.

產生我們所需要的成千倍的增長。

And using fMRI, we should be able to measure

使用 fMRI（功能性磁共振成像）， 我們不僅僅能夠測量

not just oxygenated blood flow,

血液含氧量，

but the hormones and neurotransmitters I've talked about

還有我之前說的荷爾蒙 以及神經傳導物質，

and maybe even the direct neural activity,

甚至是直接的神經活動，

which is the dream.

就是我們所說的夢。

We're going to be able to dump our ideas

我們能夠實現將我們的想法

directly to digital media.

直接轉換為數位媒體資訊。

Could you imagine if we could leapfrog language

你能想像我們能夠不使用語言

and communicate directly with human thought?

而直接通過人類的思想進行交流嗎？

What would we be capable of then?

那我們還能再做甚麼？

And how will we learn to deal

我們要如何學習處理

with the truths of unfiltered human thought?

未篩選的人類想法中真實的內容？

You think the Internet was big.

你們認為網路的問題很大。

These are huge questions.

這些都是很大的問題。

It might be irresistible as a tool

這可能是一個不可抵擋的工具

to amplify our thinking and communication skills.

用來放大你的想法以及交流的技能。

And indeed, this very same tool

確實，也就是這個相同的工具

may prove to lead to the cure

可能是解決

for Alzheimer's and similar diseases.

阿茲海默症以及 相似的疾病的方法。

We have little option but to open this door.

我們別無選擇，只能打開這扇門

Regardless, pick a year --

無論如何，隨便選一年，

will it happen in five years or 15 years?

5 年後還是 15 年後?

It's hard to imagine it taking much longer.

很難想像這會花費太多時間。

We need to learn how to take this step together.

我們需要學習 如何一起走過這一步。

Thank you.

謝謝。

(Applause)

（掌聲~）