Scientists have developed a computerised mind-reading technique which lets them accurately predict the images that people are looking at by using scanners to study brain activity.
The breakthrough by American scientists took MRI scanning equipment normally used in hospital diagnosis to observe patterns of brain activity when a subject examined a range of black and white photographs. Then a computer was able to correctly predict in nine out of 10 cases which image people were focused on. Guesswork would have been accurate only eight times in every 1,000 attempts.
The study raises the possibility in the future of the technology being harnessed to visualise scenes from a person’s dreams or memory.
Writing in the journal Nature, the scientists, led by Dr Jack Gallant from the University of California at Berkeley, said: "Our results suggest that it may soon be possible to reconstruct a picture of a person’s visual experience from measurements of brain activity alone. Imagine a general brain-reading device that could reconstruct a picture of a person’s visual experience at any moment in time."
Here is the full story. It’s a big step from paragraph two to paragraph four, and paragraph one reads to me like a misrepresentation. Predicting a viewed image from a set is very different from figuring out the image from scratch. But still this is impressive.
Addendum: Elsewhere from the world of science, here is a new article on finger ratios and the length of ring fingers and what it all means.
TC: “Predicting a viewed image from a set of ten is very different from figuring out the image from scratch.”
Where do you see that there was a set of ten pictures?
Your query is correct, I have edited away the incorrect numerical description.
Speaking from experience, it is definitely a quantum leap from selecting from a limited set and selecting from the broad range of possibilities. I got a cochlear implant late in life, and had never had the opportunity to learn how to hear. After five or six years of trying to learn how to hear, I gave up and focused on other things. I was able to identify words out of a set of up to 20 words, but from there to actual conversation was hopeless, even with the aid of limited lip reading ability. I couldn’t imagine trying to do it for visual images, given that there is probably more variety that could be expected for visual images compared to the spoken word.
Hmm, this might raise implications for advertising. For example, one could show a grid of various ads, and see which one jumps out more at the person, by reading which image they’re focusing on.
Rohit, in a similar manner that’s already done– not brai scanning, but eye scanning. Actually a great deal of cognitive neurophysiological work is being done in this area trying to map structures in the brain with functions of cognition.
“The first step is to train the software decoder by scanning a subject’s visual cortex while they view thousands of images over five hours. This teaches the decoder how that person’s brain codes visual information.”
The individualized training (5 hours?) seems to be the real key here. An interesting question is how this would play out over time. Would the training still be useful after a day, week, year, etc. Is the brain being tuned just like the software? Could a person screw up the computer training with random thoughts throughout the process?
You can buy and gain very cheap kamas from our company.
It is enlightening!
Comments on this entry are closed.