Give me a hug.

Yes, I know. Good girl.

Wanna do some training?

I work with dogs to train them to go into an MRI scanner, and try to figure out what makes them think, and what they're thinking.

My name is Gregory Berns.

I'm a neuroscientist at Emory University in Atlanta, Georgia.

I've grown up with dogs, and I've lived with dogs pretty much my whole life.

One of my favorite dogs had passed away.

In the back of my mind, I started wondering, did that dog love me in the way that I loved him?

And that was really the beginning of it.

All of these basic things that we've begun to sort out in humans, at least in human brains, we want to do in dogs.

If dogs can be trained to go into an MRI machine, then maybe we could figure out what they're actually thinking.

So, we have a dog in the scanner.

That's Katie, and she is a veteran of the project.

This is probably about the fifteenth time she's done this.

Just having the idea of, "OK, well, let's train dogs to go on the scanner," it wasn't immediately apparent how you would go about doing that.

And MRI scanners are not the most pleasant environments.

Good girl!

Callie is the little black terrier who was the first dog to train for the dog project.

High five.

I built a simulator of an MRI.

So this is the simulator that I built, here in my basement.

Then we started adding in recordings of the scanner noise, which is actually quite loud.

We just started working with, you know, treats and positive reinforcement to see if we could get her to go into this thing.

That actually proved to be pretty easy.

And then so what we did is we kind of put out word of mouth.

Do you wanna join this project?

Do you wanna train your dogs for an MRI and you know, maybe figure out what they're thinking?

The purpose of this is to go through a number of exercises including basic obedience, as well as how they react to the obstacles.

When we hold the try-outs, it lets us weed out the dogs that we don't think would enjoy doing this.

One of the benefits, then, of having more participants is it gives you statistical power.

You can start to average dogs' brains together and get a better sense for what's happening.

I started wondering very seriously if we could really finally answer this question:

Do dogs essentially like us just for the social bond? Not about the food.

And what we found was that in looking at the reward system, that almost all the dogs had equal responses both to food, as well as praise.

Even a few dogs like the praise more.

The things that we were finding about the dog's brain, in many ways confirmed, I think, what people know in their hearts about how dogs behave, and why they behave.

And the way I think about dogs is, in many ways, they're the ambassadors to the animal world.

They're not that different from many of the other mammals out there.

And so I suspect that a lot of what we find in dogs probably holds true for pretty much any mammal.