I recently received a note from a dog trainer that I first met at a meeting of the International Association of Canine Professionals (IACP). She wrote, in part:
"I was teaching a beginners dog obedience class and one woman, who was trying to train a young spaniel mix dog, seemed to be having difficulty. She was getting frustrated with the dog and was scowling at it, and the dog was showing signs of stress and seemed to be trying to avoid looking at her face.
"I remembered that there was some data which showed that dogs can recognize human faces and they react to the emotions that they see on them [click here to read more about that]. Furthermore, I recalled that dogs don't seem to respond well to negative expressions [click here for more on that]. So I said to her 'Dogs are really good at reading human faces so you might get a better response from your dog by smiling, or at least not using an obviously angry expression.'
"The woman acted quite annoyed at me and told me, 'That doesn't make any sense. My dog is just six months old. How would he know anything about human faces? To start with he's a dog, so he might know something about dog faces, but not human faces. It's not like I've been teaching him to understand expressions on my face—maybe by giving him a treat every time I smile and hitting him every time I have an angry face. I just don't believe that a young dog like this has the ability to understand anything about a human face.'
"I did remember that some researchers seemed to think that the ability that dogs have to understand faces was inborn rather than learned but I couldn't recall any details and didn't know exactly how to respond to her. Can you tell me what the facts really are?"
Actually, the situation is really simple—the brain of a dog is hardwired to recognize human faces. Of course, demonstrating this is not necessarily easy, but several pieces of research all lead to this conclusion.
Two separate lines of early research showed that face recognition is special. The first one looked at human beings and found that there is a region of the human brain that seems to be specifically designed to recognize human faces. That region is located in the temporal lobe of the brain which is an area located on the side of the brain at the bottom middle part of the cortex which would be behind, and slightly above your ears. The first evidence of the importance of this region came from patients who suffered damage to this part of the brain and developed prosopagnosia, which is a form of "face blindness". People with this problem have difficulty recognizing familiar faces and even their own face in a mirror, while other aspects of their mental functioning, such as the ability to recognize objects, are not affected.
The second line of research involved measuring the activity of single neural cells in the temporal regions of the brain. In monkeys, it was found that there were some cells that were specifically tuned to respond when the animal was viewing a monkey face. When the research was extended to look at the same area in the brains of sheep, they found cells that responded specifically to typical sheep-shaped faces. There is even some research which shows that similar cells can be found in the corresponding region of the brain of crows, and these are tuned to respond when the crow looks at bird-shaped faces.
Thus researchers began to conclude that a number of animal species have a region in the brain which was expressly designed to recognize the faces of conspecifics (the technical term for animals of the same species). It was then that the research took an interesting turn. Scientists began to ask themselves "If behavioral tests show that dogs are so good at recognizing human faces and expressions is it possible that they might have additional face-tuned cells in their brains—not just cells tuned to dog faces but cells tuned to respond to human faces as well?" One of the first examples of research exploring this question came from a team of investigators headed by Laura Cuaya of the Institute of Neurobiology at the National Autonomous University of Mexico.
The methods used in this research take advantage of new advances in brain scanning, particularly functional magnetic resonance imaging (fMRI). This is a technique that measures the level of activity in specific brain structures by detecting changes in blood flow and oxygen levels when a group of nerve cells in the brain begin to respond more vigorously. Increased neural activity tends to alter the nature of the magnetic field in defined regions of the brain and this can be measured. So if a large number of cells in a brain area were tuned to specific facial characteristics and many of them began to activate at the same time this can be easily detected by the scanner.
Actually getting dogs to participate in a fMRI-based study is not easy. To be tested, the dog is required to remain motionless in a confined space in the MRI device for a period of time. The dog must also ignore the many loud sounds made by the MRI's operation which includes the sounds of spinning gears, loud clanks, bangs, and so forth, any of which might be expected to startle a dog and cause him to move. So a period of intensive training (which might involve months of working with the dogs) is required before you can actually collect data from them. In this case, seven dogs were trained to endure the fMRI measurements required by the study.
Once inside the scanner and settled into position the dogs were shown blocks of photographs of human faces (with neutral emotional expressions) or pictures of common objects (such as a coffee cup or a clock, etc.) while their brain activity was monitored.
The results were relatively unambiguous. When presented with human faces the area of the brain which responded most vigorously in the dogs was the ventral temporal cortex (the front part of the temporal cortex), and this area was relatively unresponsive to images of objects which were not faces. This roughly corresponds to the same region of the brain that is involved in the recognition of same-species faces in the other animals that have been tested using single-cell measures. So dogs are responding to human faces in the same region of the brain where they respond to dog faces.
Why should this be happening? One possibility is a process called co-evolution, which is where the evolution of one species affects the historical evolutionary changes of another. Since dogs have been evolving into a species that shares the human environment, becoming more efficient at responding to the people in that environment would be an adaptive advantage and recognizing faces would be one aspect of this.
Another possibility is that during the process of domestication of dogs which spanned some thousands of years, humans might have systematically selected and nurtured dogs that responded accurately to human faces and emotional expressions. This certainly would make communication between people and their dogs easier. The dogs which had this ability would be preferred, cared for, and nurtured. While all of that selective breeding based on the dogs' behavior was going on it might well be the case that at the neurological level humans were unknowingly selecting dogs with brains that had specific mechanisms with the ability to recognize human faces in the same way that they might recognize the faces of animals of their own species. If so, this would be an innate ability that does not require education or experience to show itself.
The authors of this study conclude that, "The recognition of human faces by dogs could be an essential factor for establishing attachment with humans. This is supported by the evidence found so far, that dogs, and no other canids, are able to recognize and attend to human faces without training."
So perhaps telling a novice dog trainer to keep smiling and avoid angry emotional expressions might be a reasonable bit of advice since it is likely that the dog is processing and interpreting their face. His brain certainly seems to have the apparatus to do so.
Cuaya LV, Hernández-Pérez R, Concha L (2016). Our Faces in the Dog's Brain: Functional Imaging Reveals Temporal Cortex Activation during Perception of Human Faces. PLoS ONE 11(3): e0149431. doi:10.1371/journal.pone.0149431