Have you ever felt a tingling sensation sweep across your head and neck as someone brushes your hair or speaks softly to you? If so, you probably fall among the 20 percent of people who experience Autonomous Sensory Meridian Response (ASMR).
Because it is pleasurable and relaxing, some people watch YouTube videos that induce the sensation through sounds and visuals. But despite its popularity online, researchers still do not know how it works. Now, a team claims to have generated ‘artificial’ ASMR audio for the first time, made entirely without natural sound. This, they say, could help us understand how sound causes tingles.
A team of researchers analyzed recordings of classic ASMR ‘trigger’ sounds to begin the study. Keyboard typing, or breathing are examples of these triggers. The clips all showed a cyclic frequency pattern, which the team replicated using machine learning to form completely electronic, non-natural sounds.
“We suddenly realized, okay, there is a way to extract these features and [generate random sounds with] the same acoustic feature,” says author Bin Han, a signal processing engineer based at the Technical University of Kaiserslautern in Germany.
Both he and one of his co-authors, psychologist C. Clark Cao from Lingnan University in Hong Kong, say that they themselves got tingles when they listened to the artificial clips. “That was pretty exciting,” says Cao.
A survey of 250 online listeners confirmed that they were not alone, although slightly fewer reported getting ASMR from the artificial clips than from the recorded ones.
The study currently only exists as a preprint and has not yet been reviewed by other scientists. However, the idea of a common acoustic signature in ASMR triggers intrigues psychologist and neuroscientist Nick Davis at Manchester Metropolitan University in the U.K.
“While there seems to be a few universal triggers for ASMR, many people are triggered by things that seem to be unique to them, things like metallic sounds or certain types of voice,” he says. “The approach [in this preprint] is really interesting, as it looks for universal features in the sounds that ‘work’ for people.”
Stephen Smith, a psychologist at the University of Winnipeg in Canada, says the work is clearly still preliminary but that down the road, the technology could be useful for designing new experiments.
“[Highlighting] one acoustic characteristic of ASMR videos – such as repetitive sounds or soft, emotional tones – would allow researchers to examine how that precise characteristic is related to people’s subjective ASMR experiences.”
Synthetic or Human-Generated?
Han and Cao were surprised that ASMR can arise from sound that is meaningless, or it does not sound identifiably like anything. On YouTube, ASMR enthusiasts often seek out intricately produced videos meant to recreate the feeling of, say, having a haircut. Therefore, one might expect that trigger sounds work by making such scenarios feel more real.
Ryota Shimokura, an acoustics researcher at Osaka University in Japan, subscribes to this idea and does not think spectral features alone can control ASMR. The most important factor, he reckons, is whether the sound can produce a realistic, immersive experience.
In his recent study, he found that human-generated sounds, like hair-brushing or page-flipping, induce ASMR compared to nature sounds. The data suggests this is because the human-generated sounds are recorded closer to the microphone which “can make a listener feel as if [they are] there,” he says.
Physiologist and ASMR researcher Craig Richard from Shenandoah University in Virginia says Shimokura’s work shows that ASMR is about more than sound properties and supports the idea that ASMR “is a feeling of comfort and safety induced by the proximity of a kind or helpful person.” However, he does not think this is impossible to recreate digitally.
For example, he highlights the 2018 Synthetic Trigger Project by Marcus Nystrand, in which people reported getting tingles from watching animated objects scratch, rub and drag against each other.
But Richard asks, “can computers create better ASMR stimuli and scenarios than the top ASMR artists on YouTube? Not now, the genuine human factor seems crucial for the ASMR response. But as science reveals the mysteries of ASMR, it could be a possibility someday.”
The next step for Han and Cao is to play their artificial ASMR to people in a brain scanner, because there is evidence that tingles produce unique activation patterns in the brain. If the brains light up the right way, it will be a strong indication that ASMR can arise from a meaningless acoustic feature. If not, well, humans will remain in charge of ASMR – at least for now.