Language is one of the hallmarks of humanity, and is believed to be unique to our species. However, despite it being a crucial part of the human success story, scientists are just now beginning to understand what speech is, how it works and how the human brain processes it. Case in point, Duke University researchers recently made a stunning discovery: Rhythm is essential for processing speech – particularly when hearing a language other than our own.
From an auditory perspective, speech can be broken into multiple components based on length. Phonemes last an average of 30-60 milliseconds, while syllables are 3-5 times as long (words, obviously, are even longer). Processing all of those small components in real time would require a monumental amount of computational horsepower, so scientists suspect that the brain takes shortcuts, not unlike the way it does when reading written languages: The brain tends to see whole words rather than identifying them piecemeal, letter by letter.
To test this, researchers at Duke University took foreign speech (German, in this case), cut it into chunks ranging from 30 to 960 milliseconds and then patched them back together in a sort of “speech quilt.” They then played these speech quilts for study participants hooked up to an MRI machine, which scanned their neural activity while the sounds played.
Sure enough, the rhythm mattered. A region of the brain called the superior temporal sulcus (STS) lit up like a Christmas tree during the 480- and 960-millisecond quilts, but not nearly as much during the 30-millisecond samples. In other words, when the sounds were cut in such a way that they resembled the manner in which humans are believed to recognize speech (that is, in words and syllables and not individual phonemes), more of the brain became active.
“That was pretty exciting. We knew we were onto something,” said Tobias Overath, an assistant research professor of psychology and neuroscience at Duke.
The STS proved to be uniquely attuned to speech rhythms, as other parts of the brain showed know activity preference based on the changes in rhythms. They also ruled out other explanations for STS activation by exposing participants to a series of control sounds – artificial sounds designed by the researchers to mimic speech. One control shared speech’s frequency but lacked its rhythm; another retained the rhythm but lacked pitch variances. In each case, the STS did not respond the same way.
“We really went to great lengths to be certain that the effect we were seeing in STS was due to speech-specific processing and not due to some other explanation, for example, pitch in the sound or it being a natural sound as opposed to some computer-generated sound,” Overath said.
The next steps will be to see whether or not the same principles apply to languages that are rhythmically distinct from English, like Mandarin Chinese. They also want to know if “speech quilts” made of the participant’s native language would have the same effect.