Like a classic movie monster, neuroscientist David Poeppel’s fascination with bone-chilling screams was, as he puts it, “born out of an accident in the lab.”
The moment of serendipity – one that would expose the secret to a spooky scream – began in his New York University (NYU) office over a casual discussion of creatures who shriek with such fervor they bring grown adults to tears.
It started with newborn babies.
“A couple of postdocs in my lab had just become parents and unsurprisingly the topic came up,” says Poeppel, a AAAS Fellow. “My God these kids scream all the time. What the hell?”
As a professor of psychology and neural science, Poeppel’s research focuses on all aspects of hearing, including how our brains make sense of words and sounds.
He was surprised to discover minimal research around this facet of speech so core to the human experience it transcends languages and cultures.
“Everyone has screamed, knows what a scream is, can interpret a scream and probably screams him or herself,” Poeppel says.
Screams vary in meaning. There are screams of anger and intimacy. You might let out a scream of joy if your team scores or you win the lottery.
Poeppel opted to study the acoustics of “alarm screams,” indicators of fear. The project, he says, kicked off with some “very sketchy YouTube research” collecting movie clips and other audio samples.
Then, Poeppel’s team gathered screams from volunteers willing to provide their best blood-curdling sample in a sound booth. Analysis of Poeppel’s scream database yielded a common trait called “roughness,” a measure of how fast the loudness changes. Typical speech modulates between 4 and 5 hertz – or times per second. For screams, Poeppel says, the range is 30 to 150 hertz, remarkably faster than other types of sounds.
Next, Poeppel took benign everyday sentences and synthesized them to have different degrees of roughness. Study participants were asked to rate the “scariness” of each phrase. This yielded a correlation between roughness and interpretation of something as frightening. Poeppel then turned to brain imaging to nail down the link, capturing stimulation as subjects listened to screams and other alarms. Screams triggered increased activity in the amygdala, the part of the brain linked to fear response.
“Using the physics of sound, we can show in brain activity that a very simple acoustic feature like roughness correlates with your experience of finding something alarming or scary,” Poeppel says. “That's interesting because you can manipulate it.”
In fact, he says, roughness has been used implicitly for a long time in car alarms and ambulance sounds, which are – for lack of a better term – obnoxious. Famous movie screams – think Jamie Lee Curtis in “Halloween” or the iconic shower scene in Alfred Hitchcock’s “Psycho” – typically land in the higher end of the roughness spectrum.
“It's low-hanging fruit for people to take advantage of,” Poeppel adds. “You can take a whole movie track and make it rough and go from comedy to drama. In a day, you could take an episode of Larry David and make it scary.”
Poeppel’s work on alarm screams was published in 2015. Since then, his colleagues have taken the topic full circle, aiming to decipher the acoustics of loud cries emitted by babies – from distress to whininess.
“Parents share that intuition. You can tell when your kid is genuinely afraid or genuinely a pain in the neck,” Poeppel says. “As a caregiver, you make a very nuanced interpretation of their vocalization.”
In addition to scream studies, Poeppel is widely cited for other achievements across the fields of language, speech and hearing. In 2007, he and cognitive scientist Gregory Hickok proposed the theory of a “dual-stream model” for explaining how the brain experiences and perceives speech.
The theory suggests there are two concurrent streams of information used in language processing. Originally controversial, the concept challenged previously popular views on how the brain processes auditory input and has implications for treating language disorders, including those caused by stroke.
While this model is now more universally accepted, it's the scientists Poeppel inspires and mentors, he says, that are among his proudest accomplishments.
“I’m a mediocre scientist,” he jests, “but I think I'm a pretty good motivator. I'm a pretty good spotter of talent.”
In fact, Poeppel never wanted to be a scientist, aspiring instead to be a theater director.
While studying at MIT, he gravitated toward philosophy and found himself in the classroom of famed linguist Noam Chomsky. Poeppel felt connected to the topic, his exposure to languages running deeper than most. He grew up in Massachusetts, later attending high school in Germany where he took years of Latin and Greek.
“I had an intuitive sense of languages, but to see the formal study of language as a science was an amazing experience,” he says. “There’s something you know something about, and someone holds the curtain open. That's what the experience was like.”
Along with his role at NYU running the Center for Language, Music, and Emotion, Poeppel is also the director of the Ernst Struengmann Institute for Neuroscience in Frankfurt, Germany, where he explores questions at the intersection of systems neuroscience and cognition. The way a human brain processes language presents an endless puzzle, he says, each discovery yielding more questions than answers.
“What does it mean to have knowledge in your head?” Poeppel asks, describing the speed at which we’re able to pull words from our mental dictionaries. “My hunch is that we’re missing something gigantic.”