By RICK CALLAHAN – Associated Press Writer
01:26 AM ET 04/29/99
The ability to understand speech is so deeply ingrained that people can decipher recorded sentences that have been chopped into brief segments and played backwards, researchers reported today. Digitally recorded sentences were sliced into very short segments in the study, then reversed. The distorted speech was played to seven test subjects. The participants had no problem understanding the sentences. Their brains were apparently able to perceive the syllables as sounding nearly the same whether heard backwards or forwards.
“When you distort speech, it distorts certain aspects, but other parameters are still able to convey the message,” said Kourosh Saberi, a researcher at the California Institute of Technology’s division of biology.
Saberi and David R. Perrott of California State University in Los Angeles’ department of psychology reported their findings in today’s issue of the journal Nature. Ray Kent, a professor of communicative disorders at the University of Wisconsin in Madison, said the research demonstrates that many areas of the brain are used to handle complicated auditory signals. Not only are both hemispheres of the brain involved in speech processing, but eyesight plays a significant role by allowing people to unconsciously lip read to fill in missing data, Kent said. Anyone who has gone to a party held in a crowded room filled with music and chattering people has tapped those skills to understand what others are saying, he said.
“What this tells us is that speech is quite robust. We can perceive it even when a number of things have been done to distort or muddy the signal,” Kent said. “Somehow the information is preserved or at least recoverable to us even when it’s played backwards.”
Steven Greenberg, a researcher at the International Computer Science Institute in Berkeley, Calif., said the findings could someday lead to improved speech-recognition programs that allow computers to respond to spoken commands. It also adds to a growing body of evidence disputing the notion that individual vowels and consonants are crucial to understanding the spoken word, he said.