We love to tell dogs what to do, but we rarely consider whether they understand what we’re saying. Pet owners assume their dogs comprehend commands like sit, stay, or heel — even play dead and make me Instagram famous, for that matter — but without the ability to read their minds, no one can know for sure. An ingenious new study in Frontiers in Neuroscience, however, finds a way to determine which of our commands they actually understand.
The fact that we can teach dogs tricks makes it obvious that, at a basic level, they can discriminate words from non-words. But in this paper, the objective was to figure out what actually happens in a dog’s brain when it hears its owner’s command.
“This study really highlights that dogs don’t process language as humans do, and that while we train dogs with verbal commands to perform actions, this doesn’t mean that they derive the same meaning from nouns the way humans do,” study co-author Ashley Prichard tells Inverse. Prichard is a doctoral student at Emory University who specializes in studying the neural mechanisms underlying perception and decision-making in dogs using “awake fMRI.”
Before the 12 canine participants went into the fMRI machine, they were trained by their owners, for ten minutes a day, to retrieve either a soft stuffed monkey toy named “monkey” or a rubber pig toy named “piggy.”
At the end of the months-long training session, each dog was instructed to lie in the fMRI scanner while its owner stood directly in front of it. In some of the trials, the owner would say “piggy” or “monkey” then hold up the respective toy. In the other trials, the owner would hold up random objects, like a hat or a doll, and pair those objects with a gibberish word, like “bobbu” and “bobmick.”
When the pups heard “piggy” or “monkey,” there wasn’t much of a change in brain activity. When they heard gibberish, however, there was greater activation in the auditory regions of the brains. That’s the opposite of what happens when humans undergo the same experiment: We demonstrate greater neural activation when we hear words we know.
“The most exciting finding is probably that the greater neural activation to pseudowords [gibberish] over the trained words in dogs is different than what is common in human language studies,” Prichard explains. “In human fMRI, greater brain activation to pseudowords than known words means that humans are likely trying to associate meaning with the pseudowords that sound similar to words they already know.”
Prichards reasons that dogs that heard gibberish could be doing the same thing — trying to understand unfamiliar words. Dogs have been wired by natural selection to want to please us — and to want the cookies we give them when they please us — so it makes sense they would strain to find meaning in the nonsense.
When the dogs in the study heard gibberish, half of them showed increased activity in their parietotemporal cortex, which may be analogous to the part of the human brain that processes lexical differences, and the other half showed heightened activity in their left temporal cortex, amygdala, caudate nucleus, and the thalamus. The researchers think that these differences in brain regions are likely due to the varying range in breeds and sizes, but one thing is more clear — each of these dogs was likely struggling to understand what was going on.
This isn’t to say that the dogs didn’t respond at all to the words they had been trained on at home. Behaviorally, the dogs demonstrated that they had the ability to retrieve the two toys based solely off hearing either “piggy” or “monkey.” Meanwhile, the fMRI data revealed that their brains discriminated between the words by engaging regions that are similar to the human brain regions involved in language processing.
While humans will always default to verbal commands for their pets, this study underscores the fact that language isn’t the best way to communicate with a dog. The more effective way to communicate with a dog is through visual and scent cues.
“Everyone who has a dog may think they are an expert, but there really needs to be more research on how dogs think and perceive the world, not just how we humans think that they do,” Prichard says. “I hope that this research is a step towards better human-dog interactions.”