The new NS broadcast sound has been tested in a brain scanner. According to experts, that doesn’t say much

Reliable, attractive and predictable. According to marketing agency Neurensics, there are only a few associations that millions of commuters will soon have on trains, when Karen van As, the new broadcaster at NS, announced that unfortunately a train could not run between Utrecht Central and Schiphol. Her voice can now be heard in 37 small stations and from the end of this year in all stations.

According to Neuronsics, how we respond to sounds can be read directly from the brain. Commissioned by Dutch Railways, the agency tested the newly selected broadcast audio on 24 test subjects in a brain scanner. Scans will essentially show that Van Es’s voice “reduces negative emotions.”

This is called neuromarketing: searching inside the brain for consumer research. But does it work? Neuromarketers assume that brain measurements directly reveal how people react to ads, videos, and sounds, better than surveys or outdated experiments.

This is not a bad idea: a disadvantage of survey research is that people can provide socially desirable answers. For example, respondents say they buy more animal-friendly meat than they do, Wageningen researchers have been seeing for years in consumption figures.

And so neuromarketers say: Instead, look inside the brain itself instead of asking questions. That’s exactly what Neuronsics did to find out how travelers experience the sound of NS’s new broadcast, says researcher Walter Lempins. “We didn’t ask people for their opinions on voting,” he says. “Once you do that, they will think about it again. Then often things other than an automatic response will come up, so basically how your spine is reacting to that sound. We measure that in the MRI scanner.”

brain activity

At Neuronsics, scanning is not done overnight. Test subjects have to try or do something. The brain activity recorded by the scanner is only interesting if the researchers compare it to control scans. That’s why the 24 Neuronsics participants listened not only to the sound of the new NS broadcast, but also, for example, to the voice of their mother or friend who read the train announcements. We hypothesized that people respond positively to familiar and liked sounds. That makes a good comparison,” says Lembins.

However, scientists have criticized this type of research for years. Basically, the technology is good, believes Nienke van Atteveldt, professor of neuroscience at VU University Amsterdam. “The automatic response that you see on the brain scanner is very valuable. But that measurement simply cannot be translated into what you want to know. And this is how people react to sound when they are traveling.”

It is precisely this huge leap into the world outside the scanner that makes Neuronsics’ research “really very strange,” says Mark Slurs, who, as a professor of philosophy at Radboud University, studies the assumptions underpinning neuroscience research. When you’re at the scanner, you don’t care whether the train is going to Leiden or not. But if you’re on the podium and hear that, you’ll be really disappointed.”

old studies

It’s the criticism that’s haunted neuromarketing for years. Although it is in principle possible that brain scans predict what people feel and behave when they buy something in a store or when they have to process negative broadcast messages at a railway station, according to neuroscientists, very little scientific research seems to show This link is convincing.

Not that there are no studies. But the group sent by neuroscientist Lembens is largely known to Van Attveldt and Slurs for a long time and It is, for example, about an anti-smoking campaign Which has proven to be effective both in the scanner and in real life.

Slors: “If everything is predicted so well and they’re doing a lot of new research, why are they mainly going back to the old studies?” Additionally, no one keeps track of how many neuromarketing predictions fail, Charles Spence, Professor of Psychology at Oxford University organizational neuroethics: That is, unwanted results may have disappeared in the desk drawer.

According to Limpans, the reason why few new studies are available is much simpler. We rely on our customers for sales numbers and consumer behavior data. This data is often covered in trade secrets, so we can’t just share those kinds of results. According to Van Atteveldt, neuromarketing makes it hard on itself: The technology likely has potential, but it will take years before it becomes clear if it works.

But again: surely brain scans will reveal something about the sound of NS’s new broadcast? This may well be the case, van Atfeldt believes, but the lack of transparency here too is an obstacle. “It’s hard to assess without scientific evidence or publication on exactly how they handled it.”

Neuronsics mails some explanations. Well, van Atfeldt believes the agency is trying to avoid cheating when they look at brain activity themselves. The computer does this for them, with the so-called machine learning. This is much better than if the researcher saw that a brain region such as the amygdala becomes active and then claims that there are positive emotions. It’s not that simple.’

According to Van Attfeldt, it is not clear how the brain patterns were later divided into thirteen emotions, including “trust”, “lust” and “anger”. There he picks Neuronsics: this pattern is “fear” and this pattern is something else. This is fine in itself, but we can’t control how we do it.

basic feelings

Slors are more important. “What is an emotion, ‘confidence’? I can’t imagine it. Even with such basic emotions as fear, According to the agreement of the little slurs among neuroscientists about the pattern of brain activity that best reflects such feelings. One research group interprets fear differently than the other. Neuro marketers actually claim that the entire discussion does not exist and say they know for sure what they are seeing.

Neuronsics says in a later response that they disagree with the criticism, as they base their method on numerous scientific studies and use only “statistically reliable patterns”. However, the agency admits that it cannot share all the information, as is the norm with most companies.

Critics say something else. Perhaps it was a somewhat strange experience for the test subjects, because in control measurements they heard their loved ones or their mother imitate the NS’s public speech system. Slors and van Atveldt suspect that this could distort the results in unexpected ways. Slurs: “I don’t see feelings of alienation anywhere, but it wouldn’t be out of place here.” They think another voice of control would have been more neutral.

Limpens acknowledges that it may have made sense to compare multiple broadcast sounds in the scanner. “But that wasn’t what we got.” Because: NS has already definitively chosen this new vote; Brain research had to confirm essentially why this choice worked so well.