Brain patterns used for biometric id

News by Max Cooter

Differences in brain patterns allow a computer to currently achieve 94 percent accuracy in identification, but a lot more work is needed to make it a security id option.

It seems that we can't have too many biometric techniques when it comes to personal identification. A group of scientists from Binghampton University in New York has developed a new method for indentifying individuals by using brain signals.

According to the academic journal, Neurocomputing, they observed the brain patterns from 45 volunteers as they read out groups of letters. The researchers found that the individual's responses were different enough for a computer to pick out each volunteer with 94 percent accuracy, suggesting that the technique could be used for security systems.

Sarah Laszlo, assistant professor of psychology and linguistics at Binghamton University said the method was an attractive one because brain signals could not be stolen by malicious means as a finger could be.

However, there's a long way to go before the technology can be adapted for widespread use. Alan Goode, CEO of biometrics research company Goode Intelligence said that, for a start, 94 percent accuracy was far too low for practical use. “You wouldn't want to rely on that level of accuracy, biometric technology methods are generally of the order of 99.8 percent – and even that 0.2 percent is too high for some companies.” He added, however, that it was early days yet. “I would say that the 94 percent will be improved – algorithms will be tuned to react with other environments. They're probably looking for funding to commercialise it – that will possibly improve the results,” he added.

Accuracy is only part of the problem, however.  There were other issues with the method, just how it could be used. “It would be difficult to integrate the technology with devices,” he said.

According to an academic who has researched biometric identity, but who wished to remain anonymous, there would be practical difficulties with this technology. “It would mean carrying a MRI scanner around with you, they're not exactly small.” However, he said, that's not to say that it wouldn't happen. “All research that offers up new opportunities are good; fingerprint scanners were bulky once and only available for police officers. Just because they're bulky now, they may not be in the future,” he said.

He pointed out some more fundamental issues. “There's a whole load of dodgy observations about the technology, there's an assumption with MRI scans that any change in brain behaviour must be linked to the stimulus, I'm not sure that's true,” he said. He also pointed out that 45 was a very small sample for biometric technology and one that may not be an accurate sample, if they're all middle-class graduates, for example, that's a group that would skew the results.

Alan Goode said there were some possibilities for the technology, “We're seeing different ways in which we interact with machines, things like Oculus Rift, I could see this technology being used in those cases. There are a lot of engineers looking at the way we can control prosthetics, for example. But, he warned “We're looking far into the future, at least ten years.”

But there may be other technologies to explore, the anonymous academic said. “I've seen experiments with a buttock-seating biometric: the way you sit and you fidget is unique enough but of course as soon as you sit it in a different chair, you have to do re-do it.”  He put the brainscan experiment into that light, “It's an interesting bit of science but (not yet) a solution to the world's biometric problems.”

Find this article useful?

Get more great articles like this in your inbox every lunchtime

Video and interviews