What language does your dog speak?: How dog brains process language

By Alexander Albury

If you have a pet, chances are you talk to them, though you may not expect them to actually understand you. But have you ever stopped to think about just how much your furry friends might be listening? Researchers in the Neuroethology of Communication Lab at Eötvös Loránd University in Hungary set out to find out just how much dogs understand from human language. To do this, they conducted a study examining what happens in a dog’s brain when they hear different languages.

Although language acquisition has been extensively studied for decades, we are still far from understanding how language learning happens in the brain. Some researchers have taken more of a comparative approach to this question by investigating language cognition in species other than humans, including macaques, parrots, and man’s best friend, dogs.

Language learning is believed to rely heavily on exposure; that is, we learn by hearing. A popular theory of language acquisition is statistical learning. Under this theory, the brain is viewed as a pattern detection machine that is constantly learning the regularities of the world around us. In language, these regularities include aspects of speech such as tone, rhythm, and word boundaries.

Like humans, many dogs are exposed to language in their everyday environments. And just like we learn language by hearing it, dogs also gain something from this constant language exposure. Research has shown that dogs can identify speaker identity, emotional cues, and recognize specific words. And although dogs show a preference for the language they hear most,  there’s a big difference between recognizing sounds, and actually understanding language.

Laura Cuaya and colleagues set out to better understand how our dogs’ brains process language and whether dogs are tuned-in to their own “native” language – the language they’ve been exposed to the most.

The researchers trained dogs to lie still in an MRI scanner and recorded their brain activity while they listened to audio clips in a familiar or unfamiliar language. As a control, they also played the dogs scrambled versions of both languages, so they could test whether the dogs only respond to speech-like sounds or to actual language. 

They found different patterns of brain activity for scrambled and unscrambled speech in parts of the primary auditory cortex, an area of the brain important for processing speech sounds like tempo and pitch. This finding suggests that dog brains are attuned to language.

The authors are careful to point out that this finding could have multiple causes. Along with the scans of brain activity, they also had human raters listen to the audio recordings and rate how natural they sounded. The activity in the brain areas responsible for distinguishing between scrambled and unscrambled speech correlated strongly with these ratings of naturalness. This means that while the results could indicate that parts of the primary auditory cortex are responding to speech sounds specifically, it could also be the case that they are just sensitive to the naturalness of the sounds.

Next, the researchers compared brain activity between the audio clips in dogs’ familiar language and an unfamiliar language. There were unique patterns of activity for the familiar versus unfamiliar language in parts of the secondary auditory cortex, and temporal cortex – areas of the brain responsible for identifying emotion, semantic meaning, and speaker identity. Even more surprising was that this difference in activity was greater in older dogs, suggesting that the amount of exposure a dog has with a language changes how they respond to it. This lines up well with statistical learning theories of how humans learn language.

Although fMRI results can’t really tell us what dogs are thinking, these findings hint at the possibility that not only can dogs differentiate between low-level acoustic features like tone and emotional valence, but they may also be able to recognize distinct languages.

These findings may seem intuitive at first; after all, dogs have a long evolutionary history of close contact with humans, and so it’s no surprise that they might be sensitive to language. But these results provide some of the first evidence of distinct brain activity patterns for two languages in a non-human species. And maybe they’ll make you second-guess what you say when you think no one’s listening.

Original Research:

Cuaya, L. V., Hernández-Pérez, R., Boros, M., Deme, A., & Andics, A. (2022). Speech naturalness detection and language representation in the dog brain. NeuroImage, 248, 118811. https://doi.org/10.1016/j.neuroimage.2021.118811

The paper is the recipient of the 2022 NeuroImage Editor’s Choice Award