By Charles Choi, Inside Science -- Bloodcurdling screams in horror movies often send tingles down people's spines, even though they know such shrieks are fake. Now scientists have discovered the key ingredient of screams that activates the brain's fear circuitry. Inventors may have unknowingly copied this hair-raising acoustic feature into alarms found in cars and houses.

Screaming is apparently innate in origin. Many animals do it too, which suggests that the harsh sound is ancient in origin, and it may be an ancestor of more complex forms of communication. Screams can also be key to survival. They can either call for help or warn others of a life-threatening emergency.

However, surprisingly little research has investigated human screams, said study lead author Luc Arnal, a cognitive neuroscientist at the University of Geneva in Switzerland.

"I started being interested in screams when a friend of mine told me that the sound of his newborn's screams were hijacking his brain," Arnal explained.

Conventional thinking suggested that what sets screams apart from other sounds are their loudness or high pitch. However, many sounds that are loud and high-pitched do not raise goosebumps like screams can, the researchers noted.  

To find out what makes human screams unique, Arnal and his colleagues first used their lab's sound booth to record four different kinds of sounds from 19 volunteers: screams, screamed sentences, meaningless sounds such as "ahhh," and sentences spoken in a normal tone of voice. "The recording part was a lot of fun," Arnal recalled. "One part of the recording session consisted of screaming sentences like 'Get the f--- away!'" He added the screams sounded realistic enough to spur unsuspecting students nearby to want to call the police.

The scientists found that screams and screamed sentences had a trait called roughness, which means the sound changes very quickly in loudness. The rate at which loudness changes in speech is typically between four and five hertz, or cycles per second. The researchers measure a low level of roughness in English, French and Mandarin sentences that were spoken normally, and also in a cappella singing. But the rate was much higher with screams, between 30 and 150 hertz.
It remains uncertain why screams use roughness. One possibility is that the brain finds roughness extremely unpleasant, ensuring that screams win our attention, Arnal said. Roughness is essentially the audio equivalent of super-fast strobe lights, which can intensely stimulate the visual part of the brain and feel similarly unpleasant. It may be that generating roughness is difficult for vocal cords, which encourages people to limit how much they scream, to avoid a cry-wolf effect, Arnal said.
The scientists also compared screams and speech with a wide range of sounds downloaded from the Internet: musical instruments such as keyboards, alarm signals such as horns and buzzers, and pure tone intervals such as perfect fifths. Intriguingly, alarm sounds possessed roughness — the rougher the alarm, the scarier it seemed. In addition, notoriously dissonant pairs of notes such as the so-called "devil's interval" also contained screamlike rough frequencies.
Finally, the researchers used functional magnetic resonance imaging (fMRI) to monitor the brain activity of 20 volunteers as they listened to a wide variety of sounds. Both screams and alarm sounds triggered increased activity in the amygdala, the brain region linked with processing and remembering fear. They also found that volunteers were faster and more accurate at figuring out whether a sound was coming from their right or left side if it was either a scream or a normal vocal sound artificially modified to have high roughness.
One potential application for this research might be to add roughness to alarm sounds to make them more effective, "the same way a bad smell is added to natural gas to make it easily detectable," Arnal said. Warning sounds could also be added to electric cars, which are particularly silent, so they can be efficiently detected by pedestrians, he added
Rough frequencies could also be removed from sounds to make them more pleasant. "Tracking these frequencies in our acoustic environment may contribute to reducing noise disturbances and the effects of noise stress on our health," Arnal said.
The scientists plan to study the screams of human infants to see if they are particularly rough. "Screaming is one of the earliest sounds that everyone makes," Arnal said. "It's found across cultures and ages, so we thought maybe this is a way to gain some interesting insights as to what brains have in common with respect to vocalization."
Future research should also analyze animal screaming to learn how similar this behavior is across species, said behavioral ecologist Daniel Blumstein at the University of California at Los Angeles, who did not take part in this research. "How does roughness in human vocalizations compare to those of other animals?" Blumstein asked. "Perhaps the brain circuits that respond to screams are shared between animals."

The scientists detailed their findings online July 16 in the journal Current Biology.

Charles Q. Choi is a freelance science writer based in New York City who has written for The New York Times, Scientific American, Wired, Science, Nature, and many other news outlets. He tweets at @cqchoi. Reprinted with permission from Inside Science, an editorially independent news product of the American Institute of Physics, a nonprofit organization dedicated to advancing, promoting and serving the physical sciences. Top image: Tree Leaf Clover, CC license