Imagine not being able to tell your son's voice from that of a complete stranger. Welcome to the life of a 60-year-old British woman known as KH. Although a handful of people have reportedly lost the ability to recognize voices after a stroke or other brain damage, researchers believe KH is the first documented case of someone who never developed this ability in the first place.

The case came to light a few years ago when KH read an article in New Scientist magazine about people who can't recognize individuals by face. The article struck a chord, and she contacted the magazine, explaining that she had an analogous voice-recognition problem. For as long as she could remember, the voices of even her closest relatives were indistinguishable. New Scientist contacted Bradley Duchaine, a cognitive neuroscientist featured in the article, and Duchaine invited KH to visit his lab at University College London.

A successful management consultant, KH scored average or above on a variety of memory and reasoning tests. Her hearing was normal and a magnetic resonance imaging (MRI) scan of her brain revealed no obvious defects. She told the researchers her problem was limited to recognizing people's voices, explaining that she sometimes introduced herself to business clients by different names so that when they called she could identify them according to who they asked for.

Additional tests supported KH's account, says Duchaine's collaborator Lúcia Garrido, who presented the findings here 13 April at the annual meeting of the Cognitive Neuroscience Society. In one experiment, the researchers played 96 sentences, half of them spoken by famous actors, politicians, or other celebrities, the other half spoken by unfamiliar voices. KH did little better than chance when asked whether each voice belonged to someone famous. Even when she did correctly categorize a voice as famous, she identified the correct celebrity less than 5% of the time, a dismal performance compared with that of six control subjects who took the same test. KH was similarly hopeless at a second experiment in which she attempted to learn six new voices and then distinguish them from voices she'd never heard before.

"I think it's entirely novel," says Jamie Ward, a cognitive neuroscientist at the University of Sussex in the U.K. He points out that KH performed normally when asked the gender or emotional state of a speaker, supporting the researchers' conclusion that her deficit is specific to voice recognition. The underlying neurological mechanism in KH's case remains a mystery, but Garrido says the team hopes to use functional MRI scans to investigate how KH's brain responds to voices.