When my dog, Rocky, hears the word “fetch,” his tail immediately wags with excitement as he scans the room for a tennis ball. I often wonder what images and emotions are racing through his head: is he anticipating the fetch game ahead, remembering games past, or simply focused on the task at hand? As I take the time to question what effect my words have on his thought process, he loses interest and moves on to sniff out a snack.
While he might not have the patience to wait on my exploration of the methods of understanding in his canine brain, the question remains: what constitutes a word to a dog? Rocky’s response indicates that he has some ability to process and distinguish between words, however his depth of understanding remains unclear. A recent study by scientists at Emory University aims to answer this question, looking specifically at the response of canine brain activity to auditory words.
Rocky’s behavioral response to my “fetch” command demonstrates he is capable of matching a word with an action, but the study highlights that this does not tell us much about his understanding of the word itself. The research suggests that his response may have been a combination of my verbal command, emotional expression, and excited intonation. Here scientists seek to better understand how canines process human language, utilizing techniques employed in human studies of word cognition.

Rocky finishes up a game of fetch.
Studying word comprehension in canines can prove particularly challenging because behavioral responses are traditionally required to demonstrate understanding. When I say “fetch,” I can see Rocky effectively processed my command because he showed one of these behavioral responses, bouncing around the room until he found a ball. While more advanced than Rocky’s current skill set, some dogs have learned to pair a command with an object. For example, the study recalls Chaser the border collie who learned over one thousand object-word pairings and was able to retrieve different objects according to his mental dictionary of known objects. This skill requires extensive training, with few dogs on record having demonstrated advanced object-word pairing.
So, how can scientists study word comprehension in dogs without relying on a behavioral response? In this case, the study takes advantage of recent advances in awake neuroimaging – a method of studying the brain that follows changes in the brains of conscious subjects. The study examines canine cognition through approaches similar to those used in human investigations, placing dogs in a functional magnetic resonance imaging (fMRI) scanner. A noninvasive technique to measure and identify brain activity in subjects, the fMRI was used to explore twelve dogs’ cognitive responses to visual and auditory stimuli.
In order to understand how dogs respond to words, scientists must distinguish between their responses to real words and made up words. In human studies, people show a distinct difference of activity throughout the network of regions comprising the semantic system of the brain in discriminating between known and novel words. Researchers in this canine study followed a similar approach, looking for differential activity in the pariental and temporal cortex – areas of the brain important to language processing.
To prepare for the study, owners trained the dogs of varying breeds and sizes to recognize and retrieve two objects according to the object name and, importantly, to voluntarily lie still while in the fMRI machine. In the months leading up to the fMRI sessions, the owners used positive reinforcement such as praise and treats to teach the dogs the word-object pairs. The two objects used were of different texture, one a soft stuffed animal and another a rubber toy.
During the fMRI session for each dog, the respective owner said the names of the trained objects in addition to novel words the dogs had never heard before, referred to as “oddball” pseudowords in the experiment. If dogs use word discrimination methods similar to humans, they should exhibit different brain activity for the known words compared to the novel oddball pseudowords.

Eddie waits patiently in the fMRI. Frontiers in Neuroscience.
The fMRI results demonstrated that, even in the absence of behavioral response, dogs discriminate between words they know and those they do not. The canines displayed a basic brain activity differentiation between words they have heard before and the novel words, showing variation as predicted by human studies.
Unexpectedly, the dogs presented a greater neural activation for novel words than known words. Humans tend to show the opposite: a greater neural activation for known words than novel words. The researchers suggest that this result may stem from the dogs’ desire to learn and understand what the owner is saying, potentially motivated by treats and praise.
The study showed that there were differences in dogs’ brains when it came to increased activation for novel words. While some showed increased activation for the novel words where vocabulary differences may be processed, others exhibited increased response to new words in other regions of the brain. Researchers note that this may be a limitation of the study’s approach, resulting from the inclusion of dogs of varying breeds, sizes, and cognitive abilities.
Though this study shows the importance of novelty detection in canine processing of human language, the question of what words actually represent to dogs persists. Future work is required to reveal the depth of neural responses dogs employ in language comprehension. With much of these canine cognitive mechanisms remaining enigmatic, one thing is clear: Rocky just wants to go play.

Rocky and pal Daisy in Austin, TX.
Humans and canines have been communicating for thousands of years, forming uniquely personal bonds. Along the journey from wolves to my domesticated rescue mutt, dogs have become a familiar part of human lives and, like any partnership, the dog-human relationship requires effective communication, understanding, and the occasional snack.
I sometimes catch myself in a monologue directed at Rocky, chatting with him about how much he is going to enjoy his dinner or a play date he unknowingly has scheduled with his border collie friend, Daisy. Humans often like to believe that our dogs understand us, perhaps because we feel a mutual connection with the animals. The vast majority of my words certainly go over Rocky’s head, but as he sits head-tilted, tail wagging, and eyes engaged, I find myself picturing the brightly colored fMRI results showing the increased activity in his brain as he works to understand my language.
Comments