[nik's note: not the first time I have seen this story and I have to say, the headlines are more enticing than the actual reality, but still - interesting developments... ]
A new study published in Science has found a neural footprint for speech recognition, and they can determine not only what was said, but who said it.
The study used a combination of functional magnetic resonance imaging (fMRI) and a data-mining algorithm to take a look at the neural response to speech. Seven study subjects listened to three different speech sounds (the vowels /a/, /i/ and /u/) spoken by three different people. The neural patterns contained both bits of information. The 'neural fingerprint' reflected the sound being made and yet there were specific 'speaker's fingerprints' which were maintained between the different sounds.
[click heading for more]