Researchers show principles of mind-reading machine
Researchers have developed a more sophisticated way to extract visual stimuli from brain signals; they developed a computational model that uses functional MRI (fMRI) data to decode information from an individual’s visual cortex; system may help in decoding dreams, and may offer a more humane interrogation technique
A device which reveals what a person sees by decoding their brain activity could soon be a reality, say researchers who have developed a more sophisticated way to extract visual stimuli from brain signals. Scientists at the University of California, Berkeley, developed a computational model that uses functional MRI (fMRI) data to decode information from an individual’s visual cortex — the part of the brain responsible for processing visual stimuli. “Our research makes substantial advances towards being able to decode mental content from brain activity as measured using fMRI,” Kendrick Kay, a coauthor of the study, told the New Scientist’s James Urquhart. “In fact, our results suggest it may soon be possible to reconstruct our visual experiences from brain activity.”
Previous research has shown that fMRI can pick out brain activity associated with viewing different images, but so far it has only been possible to identify very basic images, from fixed categories, such as a face or a house. The process also depends on prior knowledge of the associated brain activity.
Now the Berkeley team has shown that brain imaging can reveal much more complex and arbitrary images, without prior knowledge of brain activity. The team first used fMRI to measure visual cortex activity in people looking at more than a thousand photographs. This allowed them to develop a computational model and “train” their decoder to understand how each person’s visual cortex processes information. Next, participants were shown a random set of just over 100 previously unseen photographs. Based on patterns identified in the first set of fMRIs, the team was able to accurately predict which image was being observed.
“It is going to be particularly powerful in the field of visual perception and possibly the field of decoding motor responses,” says John-Dylan Haynes of the Max Planck Institute for Human Cognitive and Brain Sciences in Leipzig, Germany. The research also hints that scientists might one day be able to access dreams, memories, and imagery, says Haynes, providing the brain processes dreams in a way that is analogous to visual stimuli. “The difficulty is that that it’s very hard to set up models for other types of complex thoughts, such as memories and intentions,” Haynes says.
-read more in Kendrick N. Kay et al., “Identifying Natural Images from Human Brain Activity,” Nature (5 March 2008) (doi:10.1038/nature06713) (sub req.)
