Mind Reading

Read minds? of course not. I love reading. Listen, mind reading may sound like pseudoscience...

Sorry for my wording--

Bull shooting. But its scientific counterpart, the identification of thought,

That's a very real thing. It is based on neuroimaging and machine learning.

And what's really amazing is the mind-reading experiment.

It's not just about eavesdropping on what someone is thinking. It's about knowing how your thoughts are actually structured. In other words, when thinking about something,

What does this mental image actually look like? What resolution? How faithful is the memory?

And how do they change over time? Now, in this episode,

Let's find out how to read other people's minds

Help answer these questions. My journey begins here at the University of Oregon. I'm meeting with Dr. Bryce Kuhl of Kuhl Research Institute. He's a neuroscientist who uses neuroimaging.

Machine learning to know what people are thinking

Without them telling him.

So tell me what you're doing here. Well, I'm in the Cognitive Neuroscience program.

And I study human memory. My laboratory primarily uses neuroimaging methods.

So we use it a lot in our work.

functional magnetic resonance imaging,

Or fMRI. And how do we use fMRI to study memory? Examining patterns of neuronal activity. There are certain patterns when forming memories. and you can record this pattern

and test if this pattern is restored

Or you can reactivate it later, for example when you remember. Does this mean we can study patterns of brain activity?

and guess what people remember and recall,

Or did you just think about it? Yes, that's what is called decoding. So basically you need an input pattern

as patterns of activity that we record.

while remembering something. And make predictions about what you will remember. You can see that this sounds like mind reading. (Laughs) Yes. Sounds like it. So, Bryce, what are you going to do with me today? So what are we going to do today?

This is new territory for us. So I decided to try a new variation.

How the experiment is being conducted on you. Therefore, we cannot guarantee specific results. But it represents where the field is

and where we want to go. Today we will participate in an experiment

A place to study faces. so we will let you learn

12 photos of famous people. These are people you already know. -If you know, yes. -got it. And try to remember these images. It will then be sent to an MRI scanner. Keep this image as clear in your mind as possible. and record your brain activity

Imagine these images. Let's build the face. Basically, I draw what I remember. -picture? -picture. Images that can actually be printed

I could hang it on the wall, so to speak. (Laughs) If you like. [Michael] The first step is for me to memorize

Photos of 12 specific celebrities

I think Bryce will try to recognize me later. I sat down to get this PhD, Max. Whether his predictions come true depends, in part, on:

About my ability to remember these faces

It is shown as clearly as possible with fMRI.

Okay, so... [sigh]

I think I have a pretty good memory of everything. - It feels like a lot is at stake. Hopefully, using the faces of memorable celebrities,

Move on to the next step.

pass through a metal detector

And in fMRI,

A place where Bryce records and monitors my brain activity.

It then feeds it into an algorithm to recreate the face. he will try this for the first time

Reconstructing faces from long-term memory

This is very difficult because we depend on each other

How well do you remember photos of celebrities?

Saw it an hour ago. I love his eyes. Look at this. [woman]

Children might say, ``I'll eat it.'' fMRI monitors activity in the brain

by dividing it into thousands of small cubes

so-called voxels or volume pixels. Each of these voxels contains:

Hundreds of thousands of neurons. With the help of fMRI, this can be detected

blood flow within these voxels,

This means that this part of the brain is active. When shown several photos of people with mustaches,

My brain responds to each facial feature. But there's probably a common area in my brain.

It is consistently committed. Maybe this is just part of my brain reacting to mustaches. So, when I imagine the face later,

Once Bryce notices that this area is occupied,

he can predict my thoughts

About mustaches. So Michael is in the scanner now.

And as I watched the words appear one after another on the screen,

and he tries to imagine that face

Memorize faces in as much detail as possible. What you can see here are the images we took. We get these brain volumes one every two seconds. Therefore, it updates in real time as it collects images. [Michael] Once you have finished part 1 of your fMRI session.

It's time for part 2 featuring Bryce and his team

Learn the language of my brain activity.

So that it can be deciphered later through a brain scan. Hello, Michael. Are you still well? [Michael] Yes. You'll see hundreds of unique faces,

Record how my brain reacts

About specific facial features. They then use this information

reconstruct a celebrity's face

I thought about this at the beginning of the scan. Really, the more we can see Michael's face, the better. I'm basically going to leave it there.

As long as he feels comfortable. [Michael] The longest time he had was 2 hours.

I was able to undergo fMRI. But I was able to see over

00 faces.

That should be enough to get it

There are some pretty interesting results. Hey Michael, you did it. That was great. We will come and kick you out. [Michael] Okay. Yes, these only show part of the photos

Here's what we took while you were there. Some images of your brain. Next, let's crunch some numbers. Max analyzes the data. let's meet again tomorrow

Let's see the results.

Let's actually reconstruct the face image here.

From the brain data we just collected. In order. See you tomorrow. In order. thank you. Thank you very much, Max. I can't wait. You should sleep all night. We want this data to be perfect.

Okay, I'm back to the doctor. Cool's laboratory. His team spent the night evaluating the data.

I can't wait to see what they think, they've seen me think. What are my results?I