We are all amateur voice actors, whether we know it or not. Even if you cannot make yourself sound like Arnold Schwarzenegger or beatbox like a pro, chances are, you regularly alter and tailor your voice to speak to different audiences, tell a story, or convey a range of emotions. Now, scientists are studying what happens in the brain when we make these voice changes – and for the first time have identified the brain regions and interactions involved in both impersonations and accents.
“Consider the difference between talking to a friend on the phone, talking to a police officer who’s cautioning you for parking violation, or speaking to a young infant,” says Carolyn McGettigan of Royal Holloway, University of London. “While the words we use might be different across these settings, another dramatic difference is the tone and style with which we deliver the words we say. We wanted to find out more about this process and how the brain controls it.”
McGettigan, who grew up in Northern Ireland, has a personal connection to the research. Despite living in southeast England for the last 13 years, she still sounds “very Irish” to people she meets, and they often are “surprised that my accent hasn’t been ‘softened’ or modified to sound more English,” she says. Often, people with strong regional accents will intentionally change their voices to better fit in with new surroundings, and McGettigan finds it fascinating that people can make such a change if they so desire.
So McGettigan and colleagues, including senior researcher Sophie Scott of University College London and a professional voice artist, set out to understand more about the neural processes underlying voice changes. They designed a series of experiments that asked participants, all non-professional impressionists, to repeatedly recite the opening lines of a familiar nursery rhyme either with their normal voice, by impersonating individuals, or by impersonating regional and foreign accents of English – all the while in an fMRI scanner.
The research team asked participants to compile in advance lists of 40 individuals and 40 accents they thought they could attempt to impersonate. They could include any voice with which they were comfortable, from celebrities to family members, and any accent no matter how general or specific. Popular people participants attempted to impersonate included Jonathan Ross (a British TV host often pejoratively referred to as “Jonathan Woss” because of his pronunciation of ‘r’ sounds.), Prince Charles, Queen Elizabeth, Sean Connery, Tony Blair, Elvis, Arnold Schwarzenegger, and Bill Clinton.
“It was entertaining to hear how the impressions came out for each participant,” McGettigan says. “Some of our amateurs were quite talented.” The most memorable one, she says was an attempt at Donald Duck. “Cartoon voices are often quite dramatic acoustically and go outside of the normal range of sounds used in natural speech, and this participant made a very good job of hitting Donald’s high-pitched squawks!”
The results? As published last month in the Journal of Cognitive Neuroscience, the researchers found that deliberately changing your voice to plan a novel articulation recruits the left anterior insula and inferior frontal gyrus (LIFG) – areas of the brain that support speech planning and production. They also found that for impersonations versus accents, areas in the posterior superior temporal/inferior parietal cortex and in the right middle/anterior superior temporal sulcus showed greater responses.
Before the study, the researchers had anticipated that the LIFG might be particularly involved in creating accents versus impersonations, but they were surprised to see no difference in this region. “We thought that participants would concentrate on modifying specific speech sounds to do accents in a more focused way than when doing impersonations, where they might also think about other things like voice quality and pitch,” McGettigan says. And because the LIFG is thought to store representations of speech sounds in a person’s environment, they anticipated that it would be more specifically involved in accents. “However, we found that activity in LIFG was enhanced to an equivalent degree during both accents and impersonations, compared with normal speech.”
Prior to this study, no research had explored the brain regions involved in controlling vocal identity. Past work had shown more generally that listening to voices activates regions of the temporal lobe of the brain above and beyond the response to non-vocal sounds. “Across a set of brain imaging studies investigating the processing of voices and vocal identities, the trend in the results indicates that this is something preferentially handled by the right hemisphere of the brain, and this is supported by the clinical literature, where people with brain injury resulting in a deficit in familiar voice recognition tend to have damage in the right hemisphere,” McGettigan says.
“The cognitive neuroscience of the voice is really still in its infancy, and our aim is to find out more about how the brain controls this very flexible communicative tool,” McGettigan says. “The voice is a powerful channel for the expression of our identity – it conveys information such as gender, age and place of birth, but crucially, it also expresses who we want to be.”
The work could also potentially have implications beyond everyday vocal modulation, as there are clinical settings where changing how the voice sounds is an important goal for speech rehabilitation. Foreign Accent Syndrome (FAS) is a very rare condition where, following stroke, a patient seems to begin to speak with a different accent – for example a person from Madrid who has never traveled begins to sound French.
“There is nothing supernatural about this,” McGettigan says, “but what makes them stand out is how their particular combination of speech output problems is perceived by others. These patients may be able to speak quite fluently, but they find that their vocal identity has been replaced by one that represents a different linguistic background or culture to other listeners.” Better understanding the brain regions that control changes in the way we sound, rather than what we say, could eventually aid individuals who are trying to recover aspects of their own vocal identity following brain injury.
-Lisa M.P. Munoz
The paper “T’ain’t What You Say, It’s the Way That You Say It—Left Insula and Inferior Frontal Cortex Work in Interaction with Superior Temporal Regions to Control the Performance of Vocal Impersonations,” Carolyn McGettigan, Frank Eisner, Zarinah K. Agnew, Tom Manly, Duncan Wisbey, and Sophie K. Scott, was published in the Journal of Cognitive Neuroscience online on May 22, 2013. The research is funded by the Wellcome Trust.
Media contact: Lisa M.P. Munoz, CNS Public Information Officer, cns.publicaffairs@gmail.com