Dr Gregory Berns, 53, a neuroscientist at Emory University in Atlanta, spends his days scanning the brains of dogs, trying to figure out what they're thinking. The research is detailed in a new book, What It's Like to Be a Dog.

Among the findings: Your dog may really love you for you, not for your food.

The conversation below has been edited and condensed for space and clarity.

Q: How did your canine studies begin?

A: It really started with the mission that killed [Osama] bin Laden. There had been this dog, Cairo, who'd leapt out of the helicopter with the Navy SEALs.

Watching the news coverage gave me an idea. Helicopters are incredibly noisy. Dogs have extremely sensitive hearing. I thought, "Gee, if the military can train dogs to get into noisy helicopters, it might be possible to get them into noisy MRI [scanners]."

Q: Why?

A: To find out what dogs think and feel.

A year earlier, my favourite dog, a pug named Newton, had died. I thought about him a lot. I wondered if he'd loved me, or if our relationship had been more about the food I'd provided.

As a neuroscientist, I'd seen how MRI studies helped us understand which parts of the human brain were involved in emotional processes.

Perhaps MRI testing could teach us similar things about dogs. I wondered if dogs had analogous functions in their brains to what we humans have.

The big impediment doing this type of testing was to find some way to get dogs into an MRI and get them to hold still for long enough to obtain useful images.

Q: How did you solve that?

A: I worked with an Atlanta-based dog trainer, Mark Spivak, to break down the steps that might make it possible for dogs to go into an MRI.

In my basement, I built an MRI simulator. We introduced Callie, the family terrier and the Newton replacement, to it, acclimatising her to the noise, teaching her to climb the stairs leading to the machine, recline into a head rest and be motionless for increasing periods of time.

After she mastered these tasks, we combined them, as would be necessary when she encountered a real MRI. It took her three months of practicing every day. After perfecting a training system, we sent out a call to local dog owners for volunteers for the study.

We've trained and scanned a total of about 90 dogs. As a matter of principle, we never restrained or drugged any. If a dog wants to get up from the MRI and leave, they can. There's no compulsion.

Q: What did the actual testing look like?

A: Mostly, we did tests analogous to neuroscience tests already done on people. For instance, we trained the dogs to do the go, no-go test. It is similar to the famous marshmallow experiment, which measures the ability of people to delay gratification.

For the dogs, we trained them to nose-poke a target whenever they heard a whistle go. Then, we taught them that arms raised in a cross meant no-go. If they saw raised arms while hearing a whistle, it was still no-go.

In the scanner, we could see that when we went no-go, a part of the prefrontal lobe became active. Dogs who had more activity there did better. It is the same for humans in the marshmallow test.

I don't believe this has been seen before in non-primates. It shows that dogs use corresponding parts of their brain to solve similar tasks as people do.

Q: Do dogs love us more than food? How did you test for that?

A: We did an experiment where we gave them hot dogs some of the time and praise some of the time. When we compared their responses and looked at the rewards centre of their brains, the vast number of dogs responded to praise and food equally.

Now, about 20 per cent had stronger responses to praise than to food. From that, we conclude that the vast majority of dogs love us at least as much as food.

Another thing that we've learned by showing pictures of objects and people to the dogs is that they have dedicated parts of their brain for processing faces. So dogs are in many ways wired to process faces.

This means that dogs aren't just learning from being around us that human faces are important they are born to look at faces. This wasn't known before.

Q: Are there practical uses to your research?

A: It can be useful for training service dogs.

For two years, we collaborated with Canine Companions for Independence to study puppies slated to become service dogs. Most service dogs cost between $US20,000 ($A24,800) and $US60,000, because they need extremely intense training to be able to do their future work.

Even though these puppies are specifically bred for the task, a great many turn out to be inappropriate. Canine Companions wanted us to try to identify which puppies were most likely to succeed.

So we scanned their puppies and followed up on them later. We found that the dogs who were the best candidates had more activity in the brain region that has the most dopamine receptors, the caudate nucleus.

They also had less activity in the part of the brain associated with fear and anxiety, the amygdala.

Q: You've done brain-scanning of sea lions. What has that taught you?

A: In recent years, record numbers of sea lions have been washing up on California beaches, having seizures and unable to function.

With other researchers, we scanned the brains of stranded animals, looking to pinpoint the damaged parts. It turns out to be the hippocampus. This is what is damaged in people with temporal lobe epilepsy.

The sea lions taught me that consciousness disorders in animals can look very similar to consciousness disorders in people. In fact, the aggregate of my research has made me realise how similar many animals are to us.

Of course, it's hard to know what animals are thinking, because they can't speak. But when you look at their brains, you realise how similar some of their processes are. You recognise that they are not just things.

Q: Are there policy implications to these insights?

A: There might be some for animals in shelters. We might be able to use this research to help shelter dogs who have aggression problems. If we could learn what's going on in their brains, we could find alternatives to euthanising them.

The main thing that these studies have given me personally is serious questions about how we treat animals.

Think about how we farm animals in large industrialised centres, where they are confined for much of their lives and then slaughtered, often cruelly.

If the animals are aware of their suffering, and I think they are, we ought to reconsider their treatment.

On a personal level, I've been a vegetarian, to varying degrees, since college. This research makes it clear that animals have brains with the capacity to feel many of the emotions we do.

That sharpens my resolve to be a better vegetarian, although I have also learned not to beat myself up about it when I fall short.

New York Times