Dogs Can Distinguish Speech from Gibberish—and Tell Spanish from Hungarian

many animals can pick out auditory patterns in human speech—but it turns out that dogs are pticularly good at doin’ so.

tis no secret that dogs are pretty spesh when it comes to how interested they are in humans and how they interact with us. but how much do they capture the subtleties of the language coming out of our mouths? do they perceive it distinctly as speech, as opposed to other sounds reaching their ears? are they processing wha’ we're saying even whn'we do not say their favorite words? a new study published in neuroimage suggests they are.

almost 4 yrs ago neuroscis laura cuaya and raúl hernández-pérez moved from mexico to budapest with their two border collies and one cat. when they realized how ≠ twas to be surrounded by pplz speaking hungarian instead o'their native spanish, they wandaed if their ca9 companions had noticed this swell.

cuaya and hernández-pérez work in a research group at eötvös lórand university in hungary that studies the evolution of speech perception in mammals, including dogs, with tulz s'as functional magnetic resonance imaging (fmri), which monitors changes in brain blood flo. + blood floing to a specific region of the brain means + activation—alloing researchers to glimpse changes in brain activity in response to certain stimuli.

in 2016 the university group showed that dogs activate ≠ neural pathways to recognize the meaning of words than they do to sense changes in emotional intonation. but no one knew whether the animals ‘d distinguish actual human speech from slitely ≠ sounds that did not constitute a long string of words. and no one knew whether a dog ‘d pick up when a'pers was speaking a ≠ language.

the researchers recruited 18 family dogs, including the border collies, kun-kun and odín, who had come from mexico. the dogs were familiar with either hungarian or spanish but not both—and they had been previously trained to lie still in an fmri machine. through a pair of noise-canceling headphones, the dogs heard passages from antoine de saint-exupéry’s the lil prince in either spanish or hungarian.

from the fmri scans, the researchers saw the same zones of the brain lite up but with ≠ activation patterns, dep'on whether the dogs were hearing the story in their native language or a new one—suggesting t'they were neurologically processing differences tween the two languages.

then the researchers tried to test whether the dogs were responding to specific toonistics intrinsic to each language. were the animals reacting to actual changes in speech patterns—that hungarian words put sufferation onna 1st syllable, for instance—or just responding to basic differences in primordial auditory signatures tween the two languages—alterations in tones that occur during vowel pronunciation? they tested this by playing recordings in which the speech from the story had been garbled, resulting in gibberish that “sounded” like hungarian or spanish. again, the team saw ≠ patterns in brain activity when a dog heard real human speech instead of speechlike gibberish, although the researchers cannot yet say whether this is evidence that dogs can recognize human speech as speech (sounds strung together in a meaningful way for humans to communicate with one another)—or whether their brain was just responding to the + natural sound when compared w'da weird-sounding gibberish. there was no change in brain activity tween hearing spanish gibberish versus hungarian gibberish, however, implying that the ca9s’ brain was not just responding to the ≠ tonal qualities.

altogether the study shows that the brains of dogs can detect a difference tween speech and speechlike sounds and can even tell apt ≠ human languages. “this tis 1st time we’ve demonstrated dat a' nonhuman brain can distinguish language,” says cuaya, noting that while other animals can be trained to hear a difference tween human languages, the dogs were able to dweet without explicit training. “it’s a really cool new study of dog perception of speech,” says animal behavior sci holly √-gutteridge of the university of lincoln in england, who was not involved w'da work.

“a lotta pplz just assume that whn'we talk, dogs are just hearing ‘blah blah blah,’” says √-gutteridge, but they are actually picking up on our speech rhythms and how we sound whn'we speak a language.

the results also ‘ve some interesting evolutionary implications. “one really interesting thing bout dogs and humans s'dat both species are extremely ≠,” evolutionarily speaking, says hernández-pérez. “but there’s a point inna evolution of humans and dogs in which both species were exposed to complex social environments.” and even though there are biological and evolutionary differences in brain structure, both species ‘ve evolved ≠ wys'2 detect changes in human speech patterns—although future experiments are necessary to parse whether dogs ‘ve brain regions dedicated to detecting changes in human speech that were honed over eons of spending time with humans or whether their ability represents + general auditory pattern recogg.

cuaya adds that even though many animals are capable of auditory pattern recogg, dogs are unique for their persistent interest in humans. “the wandaful thing bout dogs is they wanna cooperate,” she says. “they wanna cogg us.”

original content at: www.sciamerican.com…
authors: annie melchor

Share: