- Guess who (Image: Oli Kellett/Getty)
- dn27555-1_300[1].jpg (6.88 KiB) Viewed 5855 times
Blair Armstrong of the Basque Center on Cognition, Brain, and Language in Spain and his team recorded the brain signals of 45 volunteers as they read a list of 75 acronyms – such as FBI or DVD – then used computer programs to spot differences between individuals. The participants' responses varied enough that the programs could identify the volunteers with about 94 per cent accuracy when the experiment was repeated.
The results hint that such brainwaves could be a way for security systems to verify individuals' identity. While the 94 per cent accuracy seen in this experiment would not be secure enough to guard, for example, a room or computer full of secrets, Armstrong says it's a promising start.
Techniques for identifying people based on the electrical signals in their brain have been developed before. A desirable advantage of such techniques is that they could be used to verify someone's identity continuously, whereas passwords or fingerprints only provide a tool for one-off identification. Continuous verification – by face or ear recognition, or perhaps by monitoring brain activity – could in theory allow someone to interact with many computer systems simultaneously, or even with a variety of intelligent objects, without having to repeatedly enter passwords for each device.
What's in a name?
---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
But so far, the noise associated with measurements of all the brain's signals has made such data hard to analyse. Armstrong's approach solves this by focusing on brainwaves from just one region, associated with the task of reading and recognising words. This produces a much clearer signal that can be measured more quickly.
These signals are generated when a person accesses their semantic memories. While episodic memories record our experiences, semantic memories simply record the meanings of particular words. The collection of meanings that we associate with words can subtly differ from person to person, providing an individual pattern. And unlike episodic memories, semantic memories do not change too much over time. If you are stung by a bee, the episodic memory neurons that fire when you next read the word "bee" are likely to be different, but the semantic memory neurons are thought to behave roughly the same as before.
Armstrong thinks that this technique based on semantic memory could be developed into a more personal, harder to compromise alternative to fingerprint recognition or iris scanning in security systems. He refers to a case in Malaysia in 2005, where carjackers cut off the owner's fingertip so that they could trigger the car's fingerprint-activated starter. "You can't have your brain cut off," says Armstrong.
"It stretches the boundaries of how we think about biometrics," says Kevin Bowyer of the University of Notre Dame in Indiana. But he says that Armstrong's method is currently far less accurate than scanning a fingerprint or iris, and – because of the need to place three electrodes on the scalp – it's less convenient too.
Journal reference: Neurocomputing, DOI: 10.1016/j.neucom.2015.04.025