海角大神

Of all the sounds humans are capable of producing, a scream tends to get the most attention. A universal signal for extreme distress. But what鈥檚 the science behind the scream?

According to a new study, screams make use of a sonic quality called 鈥渞oughness,鈥� which activates a neural response relating to fear. Neuroscientists David Poeppel and Luc Arnal also found that screams occupied an , not shared by other human vocalizations. They published their findings Thursday in the journal Current Biology.

On the surface, screams are a simple concept. They are loud, high-pitched, and intended to convey extreme distress or danger. But conventional wisdom aside, the scientific community never settled on a concrete definition, nor did it explain our responses to screaming.

This came as a surprise to Dr. Arnal, a researcher at the University of Geneva, and Dr. Poeppel, a professor of cognitive neuroscience at New York University. They conducted a 鈥渟lightly unconventional鈥� acoustic study, collecting a host of different human and non-human sounds. Another group listened to the sounds, rating them based on how alarming they were.

As it turns out, a true scream isn鈥檛 characterized by volume or pitch. The real key is an acoustic feature called roughness. When the frequency of a sound modulates more quickly than our ears can differentiate, it is considered 鈥渞ough鈥� and we perceive it as unpleasant. Hearing very rough sounds is correlated with activity in the brain's amygdala, a region associated with feelings of fear. Screams, along with dissonant chords and artificial alarm sounds, all fell within the 鈥渞oughness domain.鈥�

鈥淭he more such roughness modulation a sound has, the more scary it seems 鈥� and the more effectively it activates the amygdala,鈥� Poeppel explains.

Poeppel and colleagues also found that screams occupied a 鈥減rivileged acoustic niche鈥� 鈥� completely separate from the frequencies found in speech and gender-associated modulations.

鈥淚t means that that region of sound 鈥� in this case roughness 鈥� is not used by other communication signals,鈥� Poeppel says. 鈥淭hat allows screams to have a high degree of specificity, and therefore efficacy, in eliciting a fear response.鈥�

But humans aren鈥檛 the only animals that scream. It can be remarkably difficult to distinguish caterwauling household feline, for example, from a screaming child. Arnal and Poeppel both agree that other animal sounds warrant similar studies.

鈥淎s a followup of this study we would like to apply our analyses to animal screams to learn how much this trait is conserved across species,鈥� Arnal says. 鈥淥ur primary guess is that screams are largely shared between mammals, but also perhaps birds and other animals. It would be of great interest to see whether roughness features are used to warn conspecifics in other species, and if the brain machinery that is recruited by these signals is also shared between animals.鈥�

But Arnal and Poeppel鈥檚 findings are already significant. Screams aside, their research could have major implications for the study of human vocalization. It may even hint at the origins of vocal communication, Arnal says.

鈥淭here are some fun practical applications,鈥� Poeppel says. 鈥淏y understanding how screams work from an acoustic point of view, we can generate better alarm signals 鈥� and of course, more scary screams.鈥�

鈥淏ut the real insights are in the basic science,鈥� he adds. 鈥淲e are interested in identifying how communication is organized. These studies show us that screams are the kind of communicative signal that occupy a very reserved part of sounds, so they do not get mixed up with other kind of signals.鈥�