Talk:Black Hole (House)
![]() | This redirect does not require a rating on Wikipedia's content assessment scale. It is of interest to the following WikiProjects: | ||||||||||||||||||||||||||||||||
|
Is the cognitive imaging technique to view the images real? If so where can a person learn more about it? --71.200.251.77 (talk) 02:24, 16 March 2010 (UTC)
I saw a segment about it on 60 minutes. It's real. However, Foreman is right in that it's 50 years from doing anything at all useful. —Preceding unsigned comment added by 83.254.144.82 (talk) 01:13, 17 March 2010 (UTC)
Indeed, this technology is still in development, and is highly exaggerated by the show's writers. You can see the reality of such technology at http://www.newscientist.com/article/dn16267-mindreading-software-could-record-your-dreams.html - the output of that technology is seen here: http://www.cell.com/neuron/image/S0896-6273(08)00958-6?imageId=gr2&imageType=large - Importantly, those results were now published over a year ago, so I feel there is little doubt that the technology has improved since then, though there doesn't seem to be any more recent public information on it, and the version seen in House was vastly superior, quite akin to Foreman's claim of the technology not being ready for practical use for another fifty years.
For those curious on how something like this works, it's perhaps most likely done using an Artificial Neural Network, which is a relatively simple piece of technology which has been in use for decades. Perhaps most interestingly, it is commonly used to recognise handwriting, translating it in to distinct letters in order to sort post, handle hand writing on portable devices like the Palm computers, and the japanese written input on the more modern iPhone OS. Quickly, an ANN works by simulating a collection of very roughly approximated 'Neurons'. In this way, the blob of virtual neurons are connected, and taught just as was seen in this House episode, with essentially random data, and once enough education has occurred, are run the opposite direction, and decode their inputs in to desired outputs, based on the understanding they formed during the earlier training.
In this instance, the neural network would be gazing down on to the subject's brain scans, or some other extremely detailed information from their brain. A similar but much crueler experiment was done with a cat which had metallic probes inserted in to it's brain, giving a more detailed reading, and in which the images were much clearer, and the cat could be clearly seen dreaming while asleep, in addition to seeing through the cats eyes while it was awake. Check that out over at http://www.youtube.com/watch?v=FLb9EIiSyG8 - pretty creepy!
So that's that.. the visual glitches seen in House were likely inspired by Datamoshing, and seem to have no real connection to the actual brain scanning technologies. The incredible resolution seen will likely only become possible as we gain better brain scanning technologies. The cat video linked in the previous paragraph is a few years older than the more recent non-invasive form linked at the beginning of my response, and is of higher resolution. One must find it easy to imagine that we'll gain the ability to truly read worthwhile imagery from another person's brain via invasive techniques like those shown, much sooner than via noninvasive techniques, if it is indeed possible at all. The military implications for that are somewhat terrifying. --124.171.18.117 (talk) 07:42, 17 March 2010 (UTC)
- The brain scans are no way near accurate enough (or ever will be) to do this kind of thing. The Berkeley University cat experiment you mention didn't ever claim to show the cat's dreams, merely what its eyes saw in a field of 6 by 6 degrees of its view (though assuming a dream activates visual neurons...). The cat was catatonic so its eyes would not move. They tapped into individual neurons (177) in the thalamus (where signals from the eye go first) and created a matrix of 13 by 13 pixels and tried to teach a computer (neural network) to assign the right neuron [activity] to the right place in the matrix. They probably did that by showing white and black squares to the cat (that's what they did in a similar experiment in Japan). That's a pretty hard "education" with a steep learning curve if you increase the number of neurons. The difficulty increases exponentially. The pictures created were very poor quality even for such a low resolution and I don't think anyone has bettered those results.Simanos (talk) 14:49, 28 September 2010 (UTC)
- This house episode was the worst. Characters were not behaving properly according to their persona and it became science fiction rather than medical drama. It was a Taub episode, pretty much like the previous episode was a Chase episode. The Wilson and Cuddy episodes were more clearly unique though. I've half a mind to go and search and find critics of this episode and add to the article because I hated it so and the article could use it anyway.Simanos (talk) 14:49, 28 September 2010 (UTC)