Welcome! These forums will be deactivated by the end of this year. The conversation continues in a new morph over on Discord! Please join us there for a more active conversation and the occasional opportunity to ask developers questions directly! Go to the PS+ Discord Server.

Reconstructing visual experiences from brain activity evoked by natural movies

3 posts / 0 new

Last post

Thu, 2011-09-22 16:08

Arenamontanus Arenamontanus's picture

Reconstructing visual experiences from brain activity evoked by natural movies

Here is a cool trick: record the brain activations as a subject sees various images, learning the mapping from brain to image. Then show a clip and deduce what image is being seen: http://newscenter.berkeley.edu/2011/09/22/brain-movies/ http://www.youtube.com/watch?v=nsjDnYxJ0bo https://sites.google.com/site/gallantlabucb/publications/nishimoto-et-al... I love the YouTube comment "That's one bad video codec in that guy's brain." - in EP this actually happens. "Oh, not one of the Skytek Cases... I get virtual migraines from the visual compression artefacts!"

—

Sat, 2011-09-24 22:35

Extrasolar Angel Extrasolar Angel's picture

Re: Reconstructing visual experiences from brain activity ...

Hmm I can't get over the feeling that intelligence agencies will go hot over this one. Possibilities for information gathering and spying? Also I wonder if reverse effect can be deducted from this-stimulate brain activity to generate images in the subject's mind? And one interesting angle in comments: "I want to see this run on the following: 1) A severe schizophrenic with nonstop visual hallucinations. 2) Someone on a massive fucking ton of hardcore psychadelics. 3) Animals and shit."

—

[I]Raise your hands to the sky and break the chains. With transhumanism we can smash the matriarchy together.[/i]

Sun, 2011-09-25 04:59

Arenamontanus Arenamontanus's picture

Re: Reconstructing visual experiences from brain activity ...

Extrasolar Angel wrote:
Hmm I can't get over the feeling that intelligence agencies will go hot over this one. Possibilities for information gathering and spying?

You would need to have the person in a fMRI machine, so it might not be unobtrusive. In theory ou could interrogate somebody and watch the internal imagery that the questions elicit, but this might only work with people with a sufficiently vivid mental imagery. Some people would be immune.

Quote:
Also I wonder if reverse effect can be deducted from this-stimulate brain activity to generate images in the subject's mind?

We need a good brain stimulation method for it. But we kind of understand the architecture of V1 fairly well now; cortical visual prosthetics have been attempted (with limited success) and the mapping of the visual field onto the cortex is known.

Quote:
And one interesting angle in comments: "I want to see this run on the following: 1) A severe schizophrenic with nonstop visual hallucinations. 2) Someone on a massive fucking ton of hardcore psychadelics. 3) Animals and shit."

Bard Ermentrout has done some cool mathematical models of how visual hallucinations work, this would be the way to test them. According to a neuroscientist friend the primary visual cortex is inhibited during dream sleep: we actually just "see" representations of things, not fake visual input. So dreams might be hard to get. Similarly schizophrenic hallucinations might also be in the higher level areas. But migraine ought to show up. And slightly similar research was done in cats a few years ago. Science marches on!

—