XPost: misc.news.internet.discuss   
   From: here@is.invalid   
      
   The Mind Readers   
      
   Brains are talking to computers, and computers to brains. Are our   
   daydreams safe?   
      
   Jack Gallant never set out to create a mind-reading machine. His focus   
   was more prosaic. A computational neuroscientist at the University of   
   California, Berkeley, Dr. Gallant worked for years to improve our   
   understanding of how brains encode information -- what regions become   
   active, for example, when a person sees a plane or an apple or a dog   
   -- and how that activity represents the object being viewed.   
      
   By the late 2000s, scientists could determine what kind of thing a   
   person might be looking at from the way the brain lit up -- a human   
   face, say, or a cat. But Dr. Gallant and his colleagues went further.   
   They figured out how to use machine learning to decipher not just the   
   class of thing, but which exact image a subject was viewing. (Which   
   photo of a cat, out of three options, for instance.)   
      
   One day, Dr. Gallant and his postdocs got to talking. In the same way   
   that you can turn a speaker into a microphone by hooking it up   
   backward, they wondered if they could reverse engineer the algorithm   
   they'd developed so they could visualize, solely from brain activity,   
   what a person was seeing.   
      
   The first phase of the project was to train the AI. For hours, Dr.   
   Gallant and his colleagues showed volunteers in fMRI machines movie   
   clips. By matching patterns of brain activation prompted by the moving   
   images, the AI built a model of how the volunteers' visual cortex,   
   which parses information from the eyes, worked. Then came the next   
   phase: translation. As they showed the volunteers movie clips, they   
   asked the model what, given everything it now knew about their brains,   
   it thought they might be looking at.   
      
   The experiment focused just on a subsection of the visual cortex. It   
   didn't capture what was happening elsewhere in the brain -- how a   
   person might feel about what she was seeing, for example, or what she   
   might be fantasizing about as she watched. The endeavor was, in Dr.   
   Gallant's words, a primitive proof-of-concept.   
      
   And yet the results, published in 2011, are remarkable.   
      
   The reconstructed images move with a dreamlike fluidity. In their   
   imperfection, they evoke expressionist art. (And a few reconstructed   
   images seem downright wrong.) But where they succeed, they represent   
   an astonishing achievement: A machine translating patterns of brain   
   activity into a moving image understandable by other people -- a   
   machine that can read the brain.   
      
      
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)