The Shapes of Thought
July 7, 2007 8:04 PM Subscribe
The Shapes of Thought is "an exploration of the visualization of emotion as EEG and other bioelectrical signals over time as retrievable data in three-dimesnional forms." It's part of the Einstein's Brain project. [Via Neurofuture.]
For a while, my best friend's mom used to have to guard Einstein's brain when she worked at Princeton.
Huh. It would be pretty sweet if in the future he could be reanimated with that. But, I think it's been preserved in formaldehyde or something, so it will eventually degrade anyway.
posted by delmoi at 9:10 PM on July 7, 2007
Huh. It would be pretty sweet if in the future he could be reanimated with that. But, I think it's been preserved in formaldehyde or something, so it will eventually degrade anyway.
posted by delmoi at 9:10 PM on July 7, 2007
Interesting, but, I think, more art than science. If I understand the technology correctly, the visualization of EEG data over time would include a huge amount of fuzzy data and error. The EEG is useful for instant information - like, what area of the brain is lighting up in general over a period of milliseconds. Unfortunately, its resolution is severely limited by the amount of sensors can be placed realistically on a person's head, given the size they need to be to read the conductivity of that area.
With neural imagining at the moment, there is always a trade-off between resolution and immediacy - the much-touted fMRI is extremely limited because it can only pinpoint information to a certain level of precision - unlike a microelectrode or something, which is extremely precise but limited in its scope. PET and EEG are useful but limited in other ways. Every time any imagining study comes out, somebody is like, "Oh my god! they've discovered the exact location of love!" Or some such... but the fact is these techniques have very specific uses and are not able to "find" things we understand only crudely, like emotions.
Still, this is an interesting idea and could be useful for finding grossly abnormal patterns in sleep brain activity. Thanks for the link.
posted by BlackLeotardFront at 9:18 PM on July 7, 2007
With neural imagining at the moment, there is always a trade-off between resolution and immediacy - the much-touted fMRI is extremely limited because it can only pinpoint information to a certain level of precision - unlike a microelectrode or something, which is extremely precise but limited in its scope. PET and EEG are useful but limited in other ways. Every time any imagining study comes out, somebody is like, "Oh my god! they've discovered the exact location of love!" Or some such... but the fact is these techniques have very specific uses and are not able to "find" things we understand only crudely, like emotions.
Still, this is an interesting idea and could be useful for finding grossly abnormal patterns in sleep brain activity. Thanks for the link.
posted by BlackLeotardFront at 9:18 PM on July 7, 2007
They gave her a gun...
"If it moves, kill it"
posted by arialblack at 10:44 PM on July 7, 2007 [1 favorite]
"If it moves, kill it"
posted by arialblack at 10:44 PM on July 7, 2007 [1 favorite]
Interesting, but, I think, more art than science. If I understand the technology correctly, the visualization of EEG data over time would include a huge amount of fuzzy data and error.
I've taken a couple of AI classes and what you've just said makes no sense. It's the equivalent of everyday phrase "Watermelon only tastes good if you car the apple rotor"
EEGs, or electroencephalographs measure brain activity. Normally they are presented as graphs, and so they are already a visual thing. All this is doing is taking those waves and showing them on a 3d ball of some sort, rather then graph paper.
posted by delmoi at 10:56 PM on July 7, 2007
I've taken a couple of AI classes and what you've just said makes no sense. It's the equivalent of everyday phrase "Watermelon only tastes good if you car the apple rotor"
EEGs, or electroencephalographs measure brain activity. Normally they are presented as graphs, and so they are already a visual thing. All this is doing is taking those waves and showing them on a 3d ball of some sort, rather then graph paper.
posted by delmoi at 10:56 PM on July 7, 2007
EEG data does contain a lot of noise, and visualizations that allow the viewer to more easily filter out that noise could certainly be useful. By making a dramatic visualization, you could end up making the noise even more distracting if you're not careful.
For this to be more than an art project, the developers of the software would have to work with people trained to identify important information in current EEG output and see what types of views and controls they think could assist them in finding out more from the data.
posted by demiurge at 12:32 AM on July 8, 2007
For this to be more than an art project, the developers of the software would have to work with people trained to identify important information in current EEG output and see what types of views and controls they think could assist them in finding out more from the data.
posted by demiurge at 12:32 AM on July 8, 2007
« Older creative fun | Escher prints in 3D LEGO versions Newer »
This thread has been archived and is closed to new comments
They gave her a gun...
posted by Samizdata at 8:43 PM on July 7, 2007