Do you see a face? You’re actually hallucinating - Susan G. Wardle

譯者: Lilian Chiu 審譯者: Helen Chang

Imagine opening a bag of chips only to find Santa Claus looking back at you.

想像一下，打開一包洋芋片， 卻發現聖誕老人回望著你。

Or turning the corner to see a smile as wide as a building.

或者走過街角，看到 建築物咧嘴對你微笑。

Humans see faces in all kinds of mundane objects,

人類會在各種平凡的 物體上看到臉孔，

but these faces aren’t real—

但這些臉孔不是真的——

they’re illusions due to a phenomenon known as face pareidolia.

它們是「面部錯覺」現象 所造成的幻覺，

So why exactly does this happen, and how far can this distortion of reality go?

那麼為什麼會發生這種現象？

這種現實扭曲能扭曲到什麼程度？

Humans are social animals,

人類是社交的動物，

and reading faces is an important part of our ability to understand each other.

我們了解彼此的能力中， 有很重要的一部分是「讀臉」。

Even a glimpse of someone's face can help you determine if you've met them before,

就算只是一瞥某人的臉孔， 也能幫你判斷是否以前見過他、

what mood they’re in, and if they’re paying attention to you.

他的心情如何， 及他是否有在注意你。

We even use facial features to make snap-judgments

我們甚至會用面部特徵來快速判斷

about a person’s potential trustworthiness or aggression.

一個人是否值得信任 或是否有侵略性。

To capture all this vital information,

為了捕捉到這些重要資訊，

humans have evolved to be very sensitive to face-like structures.

演化讓人類變成對類似 臉孔的結構十分敏感。

Whenever we see something,

每當我們看到某樣東西，

our brain immediately starts working to identify the new visual stimuli

我們的大腦會馬上開始 辨識新的視覺刺激，

based on our expectations and prior knowledge.

其根據是我們的期望以及過往知識。

And since faces are so important,

因為臉孔如此重要，

humans have evolved several regions of the brain

人類的大腦演化出數個區域

that enable us to identify them faster than other visual stimuli.

讓我們辨識臉孔的速度 能比辨識其他視覺刺激更快。

Whereas recognizing most objects takes our brain around a quarter of a second,

我們的大腦大約要花四分之一秒的 時間才能辨識出大部分的物體，

we can detect a face in just a tenth of a second.

但我們可以在十分之一秒 之內就偵測出臉孔。

It makes sense that we'd prioritize identifying faces over everything else.

我們把辨識臉孔放在 第一優先是合理的。

But brain imaging studies have revealed that regions

但腦影像研究指出 這些區域可能太過敏感了，

may actually be too sensitive,

leading them to find faces where they don’t exist.

導致它們在連沒有臉孔的 地方也會找到臉孔。

In one study, participants reported seeing illusory faces

有一項研究拿純噪音影像 給受試者看，受試者

in over 35% of pure-noise images shown to them,

回報在超過 35% 的 影像中看到臉孔幻覺，

despite the fact that nothing was there.

其實影像中什麼都沒有。

It might seem concerning that our brains can be so wrong so often,

我們的大腦會這麼常犯 這麼大的錯，挺讓人憂心，

but these illusory faces might actually be a byproduct

但這些臉孔幻覺其實可能只是

演化優勢的某種副產品。

of something evolutionarily advantageous.

Since processing all the visual input we encounter quickly and correctly

要快速且正確地處理我們 遇到的所有視覺輸入資訊

is an enormous computational effort for the brain,

對大腦來說要用到很大量的運算，

this kind of hypersensitivity might act as a useful shortcut.

這種超高敏感度可能 會是很有用的截徑。

After all, seeing illusory faces is usually harmless,

畢竟，看到臉孔幻覺通常也無害，

while missing a real face can lead to serious issues.

但錯失一張真正的臉孔 就可能導致嚴重的問題。

But for hypersensitivity to be more helpful than harmful,

但要讓這種超高敏感度 能夠利大於弊，

our brains also need to be quick at determining when a face is real

我們的大腦也得快速判定 看到的臉孔是真實的

and when it isn’t.

還是幻覺。

So how fast can our brains tell when they’ve been duped?

我們的大腦多快能發現它被騙了？

To answer this question, researchers used a form of brain imaging

為了解答這個問題，

研究者採用了一種 大腦成像技術叫做腦磁圖。

known as magnetoencephalography.

By measuring the magnetic fields caused by electric currents in the brain,

藉由測量大腦中電流所造成的磁場，

this technique allows us to track changes in brain activity

這項技術讓我們能追蹤 大腦活動的改變，

at the scale of milliseconds.

且可以精密到毫秒。

With this tool, researchers revealed that the brain generally recognizes

用這項工具，研究者 發現大腦一般來說

可以在四分之一秒之內 就發現臉孔是幻覺，

a face as illusory within a quarter of a second—

around the same time that we can identify most non-face visual stimuli.

和我們辨識出臉孔以外的其他 視覺刺激所需要的時間差不多。

However, even after our brain knows the face is fake,

然而，在我們的大腦 知道臉孔是假的之後，

we can still see it in the object.

我們仍然可以在物體上 看到那個臉孔。

And by messing with these brain areas,

透過操弄大腦的這些區域，

we can further impact our ability to differentiate between fact from fiction.

我們可以進一步影響我們 區別事實和虛構的能力。

In one study, researchers stimulated a participant’s fusiform face area

在一項研究中，研究者去刺激 受試者的梭狀臉孔腦區，

while they were looking at a non-face object.

此時受試者正在看著 一個非臉孔的物體。

As a result, the participant reported momentarily seeing facial features

結果是，受試者回報說自己 短暫地看到了臉孔特徵，

despite the object remaining unchanged.

而該物體本身一直都沒有改變。

And while looking at a real face, stimulation of this same area

當受試者看著真實臉孔時， 若刺激這個大腦區域，

created perceived distortions of the eyes and nose.

會讓受試者察覺到 眼睛和鼻子有所扭曲。

These studies suggest that certain features are crucial to face detection.

這些研究點出了某些特徵 對於臉孔偵測而言十分重要，

Just three dots can be enough to represent eyes and a mouth.

單單三個點，就可以 代表雙眼和嘴巴。

People will even assign gender, age, and emotion to illusory faces.

人甚至會將性別、年齡、情緒 賦予他們所看到的臉孔幻覺。

It’s unclear whether a person’s culture or individual history

還不清楚人的文化和過往經歷 是否會影響這些感知，

impacts these perceptions,

but we do know that pareidolia isn’t unique to the human experience.

但我們確實知道， 錯覺不是人的專利。

Rhesus macaque monkeys show eye movements similar to our own

恆河猴在觀察會引發 錯覺的物體和真實臉孔時，

when observing pareidolia-inducing objects and real faces,

會有和我們類似的眼動，

suggesting that this phenomenon is baked deep into our social primate brains.

表示這個現象已經深植在 我們靈長類的社交大腦中。

So, next time you see an unexpected face in a coffee, car, or cabinet,

所以，下次你在咖啡、汽車， 或櫥子上看到未預期的臉孔時，

remember that it’s just your brain working overtime

切記這只是你的大腦在加班工作

not to miss the faces that really matter.

以免漏掉了真正重要的 臉孔，如此而已。