Study shows dogs are more visually attuned to actions rather than who or what is performing them 

Have you ever wondered what your dog thinks when he stares at the TV, seemingly fascinated by News At Ten?

Scientists have found that your dog is probably not focused specifically on Huw Edwards, but more on what the people on the screen are doing.

Study dogs at Emory University in Georgia, USA, had their brains scanned by an MRI machine while watching a half-hour video of stimulating content.

This included clips of dogs running around, people interacting with each other, vehicles passing by, and a cat in a house.

Data from the MRI was fed into an artificial intelligence (AI) called Ivis, which correlated brain activity with whether an action or object was shown on the screen.

The results showed that dogs are much more visually attuned to actions in their environment, rather than who or what is performing those actions.

Neuroscientist Erin Phillips said: ‘Although our work is based on only two dogs, it provides proof of concept that these methods work in dogs.

“I hope this paper clears the way for other researchers to apply these methods to dogs, as well as to other species, so that we can gain more data and greater insights into how the minds of different animals work.”

Dogs have only two types of cone cells in their eyes and can only perceive the colors blue and yellow.

This is very different from humans, who have three types of cone cells and can visualize the entire color spectrum.

However, dogs also have a higher density of motion-sensitive vision receptors than we do.

Scientists believe that dogs can perceive the world visually differently than humans in this way because they need to be more aware of threats in their environment.

It could also be because they rely more on their other senses because while humans are very visually oriented, dogs’ senses of smell are much more powerful.

The researchers at the Canine Cognitive Neuroscience Lab wanted to find out if there were any other differences between how dogs and human minds reconstruct what they see.

They recruited Bhubo, a four-year-old male Boxer mix, and Daisy, an 11-year-old female Boston Terrier mix, to participate in a study.

Both puppies had been trained to enter and lie down in an fMRI machine completely uninhibited, so were able to have their brains scanned while awake and alert.

“They didn’t even need treats!” said Mrs Phillips.

For the study, Daisy and Bhubo were each shown specially designed movies in three 30-minute sessions for a total of 90 minutes while relaxing in the fMRI machine.

The films contained video clips that the researchers thought a dog would find interesting enough to watch for an extended period of time.

They were filmed by the researchers using a gimbal — a rotating camera mount — and a selfie stick that allowed them to capture images from a “dog perspective.”

The clips showed dogs running around and people interacting with dogs, giving them pets or treats, or waving a toy at the camera itself.

Other activities included passing vehicles, people cuddling or eating, a deer crossing a path, a cat in a house and dogs on a leash.

While the dogs watched their movies, an MRI scan of their brains was taken that visualized neural activity.

Ms Phillips said, “It was funny because it’s serious science, and a lot of time and effort went into it, but the bottom line was these dogs were watching videos of other dogs and people acting a little silly.”

In comparison, two people were also shown the videos while they were in the fMRI machine undergoing a scan.

The video data was then segmented by timestamps and each clip was given ratings to identify what was currently on the screen.

The classifications include objects, such as dogs, people, vehicles or other animals, or actions, such as sniffing, eating or playing

This information, as well as the MRI data from dogs and humans, was fed into the neural network Ivis and the results were published this week in Journal of Visualized Experiments.

Ivis was trained to map brain activity with the two classifiers, which he was able to do for both with 99 percent accuracy using the human data.

However, it was only successful in finding correlations with the action-based classifiers for the dog data, and it did so with an accuracy of between 75 and 88 percent.

This suggests that dogs’ brains prioritize what’s happening for them over who or what’s involved — a big difference from how the human brain works.

‘We humans are very object-oriented,’ says corresponding author Professor Gregory Berns.

‘There are ten times as many nouns as verbs in the English language because we have a special obsession with naming objects.

“Dogs seem to be less concerned with who or what they see and more with the action itself.”

He added: ‘It makes perfect sense that dogs’ brains will be highly attuned to actions in the first place.

“Animals have to be very concerned about things that happen in their environment to avoid being eaten or to keep an eye on the animals they want to hunt.

“Action and movement are paramount.”

In the future, the researchers want to assign brain activity to olfactory input, since dogs have a much larger part of their brains devoted to processing olfactory information.

They also want to do more detailed research into the vision perception of dogs, and possibly other animals.

Professor Berns said: ‘We have shown that we can monitor the activity in a dog’s brain while he is watching a video and, at least to a limited extent, reconstruct what he is watching.

“It’s remarkable that we can do that.”