Scientists have identified the first known case of developmental phonagnosia—a likely inborn inability to recognize voices—a finding that may help shed light on how the brain processes sounds as well as offer ways to improve clumsy voice-recognition software.
A study in the January issue of the journal Neuropsychologia outlines the story of K.H., an educated and successful woman in her 60s who cannot identify even familiar voices—such as that of her own daughter—but has no other hearing ailments.
The study’s principal investigator, Brad Duchaine, a neuroscientist at University College London, said K.H. contacted him after reading a news article about prosopagnosia, or “face blindness,” when people cannot recognize the faces of those they know. [See also a 2007 Dana story on prosopagnosia.]
As is common among people with prosopagnosia, K.H. had long known something was wrong but could not identify the exact problem until she read a list of symptoms. To work around her difficulty, she set up her day to answer only prearranged or “booked” phone calls and gave out different variations of her name to co-workers to help differentiate them.
“It’s easy to overlook these kinds of conditions, and that’s especially true for voices, which are less important than recognizing faces,” says Duchaine, who has worked extensively with people who have prosopagnosia. “You simply don’t know what’s wrong.”
Other cases of phonagnosia have been documented, he adds, but in all those cases, the condition occurred after damage from strokes or other traumas to the right hemisphere of the brain. It wasn’t easy for doctors to tell where among all the damage was the brain region that caused this “voice blindness.”
In K.H.’s case, “these selective deficits open up a unique opportunity to inform us about brain function,” says Katharina von Kriegstein, a neuroscientist at University College London not involved in the study.
“Although the connection from deficit to brain function has not been made in [this paper], I am sure that there is a lot of fruitful work ahead which will elucidate the mechanism for voice recognition in the human brain. This is something with which artificial systems still struggle.”
Hunt for others
Because the condition is rare, Duchaine and lead researcher Lucia Garrido had to develop a new series of tests to assess K.H.’s condition.
The results showed that K.H.’s problems involved only voices. She could recognize only 1 of 60 U.K. celebrities—the distinctive Sean Connery—by voice, whereas control groups tended to get more than half correct.
“In tests where she had to discriminate voices she had heard before from those she hadn’t, she also did much worse than controls,” Duchaine says. Her hearing levels, however, were normal, as was her ability to identify specific sounds in noisy environments. She could recognize well-known pieces of music and had no trouble determining the emotional states of speakers.
A magnetic resonance imaging scan showed no structural brain defects, and Duchaine says the condition showed no genetic tendency, unlike many cases of prosopagnosia.
Duchaine plans to run additional tests and hopes to track down other people with the condition. It might show up in a number of ways, he says; prosopagnosia, likewise, is a highly variable condition, in which some people have difficulty only with facial identity, others solely with facial expressions and some just with determining another’s gender.
He already has a couple of leads, suggesting that the condition might not be so rare. Other researchers, though, are taking a more cautious stance.
“Prosopagnosia seems to be much more common than initially thought,” von Kriegstein says. “It is too early to say whether the same will apply to phonagnosia.”