Born with a love of speech

Do human newborns develop their preference for speech through in-utero eavesdropping, or is their penchant for speech innate? It's a bedevilling question to test, but one that's central to understanding the origins and dynamics of humans unique propensity for speech. Now a McGill University psychologist believes she's separated out the complicating effects of the uterine sound barrier. And the results, says Dr. Athena Vouloumanos, point to a genetic predilection for listening in on speech above other similar sounds.

"It's well established that neonates have a preference for speech above other sounds, but where does this come from? Is it something that's built in and there's something about the speech signal that they're tuned to listen to without the benefit of experience, or does it come from their prenatal experience in the womb? I think we've shown that there's an experience-independent component to newborns' preference for speech," says Dr. Vouloumanos, an Assistant Professor in McGill's Department of Psychology in Montreal, Canada.

She'll be presenting the findings of her latest research at the annual meeting of the American Association for the Advancement of Science in St. Louis, February 17th.

Neural and cochlear development is such that at about six months gestation the fetus begin to hear a range of frequencies. Thus, there's the possibility that newborns preferentially listen to speech because they're used to it from their in-utero tuning-in.

But, which speech sounds actually get through the uterine sound barrier to the fetus? According to Dr. Vouloumanos, the best evidence indicates that the uterine environment acts as a low-pass filter. This means that only very low, deep sounds – those below about 400 Hertz – make it to a fetus' ear.

Dr. Vouloumanos and colleagues developed a painstaking two part experimental procedure to separate out possible in-utero exposure from innate predisposition.

At the core of the studies is a unique set of complex speech analogues designed to be as close to speech as possible, without being words. This provided a stringent test for a newborn's ability to distinguish speech from other sounds.

The speech-like sounds, which resemble a chorus of Martians, closely mimic the melodic aspects of speech. They also contain a fundamental frequency – the core pitch of a voice, such as baritone or soprano, that helps us clue in to the identity of a voice. These non-speech stimuli were developed by Drs. Sonya Bird and Guy Carden in the Department of Linguistics and Dr. Janet Werker in the Department of Psychology, all at the University of British Columbia, where Dr. Vouloumanos was a doctoral student at the time of research.

In the first stage of the experiments, Dr. Vouloumanos and her colleagues compared the response of 10 to 72-hour-old infants to the speech-like sounds and those of monosyllabic nonsense words, such as "lif". When the neonate sucked on a sterilized pacifier, one or the other sound type was played for a one minute interval. The infants were observed for changes in their sucking behaviour.

"The neonates changed their behaviour to elicit speech," says Dr. Vouloumanos. "What we found is that their sucking behaviour increased to hear speech and it decreased when the sucking would elicit the non-speech sounds."

The study involved 20 neonates at British Columbia's Women's and Children's Hospital. The research, presently in press in the journal Development Science, revealed that the neonates sucked about 15-per cent more in response to the human speech compared with the analogue.

Having demonstrated that newborns could distinguish between these two types of very similar sounds, Dr. Vouloumanos then tested the core question: Had the infants already picked-up an ear for speech in the womb? Could the newborns discriminate between just the low frequency aspects (those that are heard across the womb) of their speech analogue and non-sense words? If they showed a similar preference for the low-frequency non-sense words over the speech analogues it would indicate that they could have developed the preference in their mother's uterus

When Dr. Vouloumanos' tested 16 neonates with just the low frequency sounds the infants showed no preferences.

"We established that the low-pass filter signals weren't distinguishable to the neonates. The low frequency sounds that we used in our study didn't have enough information for the neonates to tell them apart," says Dr. Vouloumanos of the unpublished results. "From that I conclude tentatively that this preference for speech that we observe might be independent of prenatal experience. It's something that's more built-in."

As excited as she is about the results, Dr. Vouloumanos, says that the findings raise as many questions as they answer. Given the similarity between the speech and speech analogues, the $64, 000 question is what is exactly about speech to which newborns are so drawn?

Says Dr. Vouloumanos: "There's something about speech that infants will orient towards, but we don't yet know what it is. What are the aspects of the speech stimulus have we taken away in making our non-speech sounds?"

Source : Natural Sciences and Engineering Research Council