go to content

No, RBS Can't Read Your Mind And Tell You If You Ought To Work In Banking

Weird that we have to say that, really.

Posted on

There's a story going around today that Royal Bank of Scotland is using brain scans at graduate job fairs to "read" students' minds and tell them which roles they should take at the bank.

The Huffington Post reported that RBS takes "portable brain scans" to campuses around the country. Apparently, students wear a headset with electrodes on it and are shown a series of images and videos. “An algorithm is then used to look for spikes in brain activity,” an RBS spokesperson told HuffPo. “It then automatically generates which courses you might be most interested in.”

The story's also been reported in Metro.

Unfortunately, both the story and the science behind it are "bollocks", according to neuroscientists.

Oh for God's sake... what fresh bollocks is this? https://t.co/X7xeCuikfz

Hey @JasminGray2, please please delete this article please please please please. It makes science cry. @m_wall https://t.co/yGxCaaKEwg

The idea that RBS's "MindWave" machine can do this is "total rubbish", Matt Wall, a neuroscientist at the Imanova Centre for Imaging Sciences at Hammersmith Hospital, told BuzzFeed News.

The basic idea behind the science isn't 100% ridiculous. It is possible, in a very limited sense, to "read" people's minds with brain scans using a very high-tech machine called a functional magnetic resonance image scanner, or fMRI.

To do this, "You put people in a scanner for 12 or 20 hours over several days or a week, an hour or so at a time" Wall said. "You show them lots of pictures of videos, record the data, and feed it into a kind of semi-intelligent algorithm.

"The algorithm learns that when you see a picture of a cat, you get this brain pattern; when you see a picture of a house, you see this brain pattern, et cetera. And then you can later look at a scan and the algorithm can say 'you were looking at a picture of a cat,' or a house.

"And that's, kind of, reading their brain. It's very difficult, but it works, more or less, on average, over a large group."

But that's not what RBS's machine does. It's a simple EEG machine – a couple of electrodes measuring electrical activity – and the students use it for only a few minutes.

And you just can't get any good information out of that. For one thing, brains are all different, and you have to train an fMRI scanner for hours and hours to get a sense of how each brain works. "The idea that you can just stick this headset on someone and read it off is ridiculous," Wall says.

And even if that weren't true, the scanner itself won't do the job, because there's too much else going on in your head. "These cheap one- or two-sensor EEG headsets are silly," Wall says. "If you raise your eyebrow [while you're wearing it], or smile or move your eyes, you'll see a massive spike in the data.

"You're trying to record electrical activity in the brain, but every time you move a muscle you get electrical activity in the muscle. The signal from that is much stronger than the signals coming from the brain, through your skull and scalp." You can see that happening in this video:

View this video on YouTube

youtube.com

The headset will give you lots of data, but that data will be meaningless.

"It’s mostly noise," Wall says. "If you torture a dataset long enough it’ll tell you what you want it to tell you, but it’s like torturing a human. The information you get isn't reliable.

"As a publicity stunt it’s not bad, but you just can’t meaningful stuff out of it."

The problem of "neuromarketing" – the use of neuroscience-y buzzwords and often meaningless brain scans to sex up market research and other commercial activity – has been growing in recent years.

It has also been given the less flattering nickname "neurobollocks".

RBS wouldn’t comment on the record, but BuzzFeed News understands the bank feels the HuffPo piece overstated the scan’s potential, and that the scans will only be used to start conversations with students at graduate fairs, although they do think it is of value.

Tom Chivers is a science writer for BuzzFeed and is based in London.

Contact Tom Chivers at tom.chivers@buzzfeed.com.

Got a confidential tip? Submit it here.