New tech can turn thoughts into movies.

I don't get what the "breakthrough" is here. They show someone images then scan the brain and compare the scan to youtube clips and the resulting matches are not the same as the images shown...

It shows outlines of people when the bird is shown, and when the little boy is shown it shows three different people, looks like a man then a woman, then a man that looks to be wearing a tie.

Originally posted by MikeNice81
All of the "gee whiz" technology people play with today is already outdated to the people on the inside.

Its due to economies of scale, not technological suppression. It takes a decade or so for the technology to become cheap enough for public consumption.

reply to post by BIGPoJo


What is weird is when there was a black woman the reconstruction appeared white and when there was a white woman at the end the reconstruction appeared black.

reply to post by BIGPoJo

Currently, researchers are only able to reconstruct movie clips people have already viewed.

Ah.. I smell shenanigans. Only movie clips? Why would that be. Seems like if they could get at peoples real thoughts. like the faces of loved ones the images may be better displayed. Only movie clips?

And Pink Panther for God's sake? No wonder it's a blurry mess.. who wants to remember that?

reply to post by watchdog8110


Could never be used against you. People's memories are far too unreliable. We remember dreams, we manipulate our memories, and drum up fantasies.

Now only if they can pick up on audio waves and get these wonderful compositions I have in my brain out into the airwaves for all to, who care, hear.

so... in what journal were the findings published? Cannot find any peerreviews.

ya know a entrepreneur would see a opportunity here.
making new stylish tin foil hats.

reply to post by BIGPoJo


You beat me to it... i just saw this on skynews website and straight away thought 'jeeez, now they can use these machines to read our minds and picture what we are thinking'... this is great news for those who have lost their memory.... actually wait a second.... how can they think of something which happened in the past if their memory of it doesnt exist? I just thought about that.... sounds strange to me... it says people will have to have this scanner on for hours but what i'm thinking is that these machines over time will be able to steal peoples thoughts alot quicker and be put out on the streets!!

Man, this gets worse......

reply to post by BIGPoJo


Might prove to be a bad technology for married couples

Very interesting technology, but I think it's still far from being of any practical use.

If anyone is interested, the original paper can be found on Current Biology.

From what I understand they used blood oxygen level-dependent signals measured using fMRI
to train a number of models they had of the possible brain response to the images.
They then used a different set of data for testing the accuracy of the trained models.
Having some insights on how the brain processes images they were able to reconstruct image data using a statistical approach.

The voxels that were mentioned by someone in this thread are not the ones associated with 3D rendering but
are involved in the sampling of data from the fMRI. Voxels are the output of a technique (Statistical parametric mapping) used to statistically evaluate fMRI results which produces a spatial map of the brain activity.

The paper It's fairly technical in nature, maybe some of you can get a better understanding, the details of the process remain obscure to me but it makes more sense overall now I think.

download.cell.com...

it just makes ya wonder what tech they really have got hidden up there grubby little sleeves

i think this is kinda spooky

Originally posted by GogoVicMorrow
reply to post by BIGPoJo

 

Currently, researchers are only able to reconstruct movie clips people have already viewed.

Ah.. I smell shenanigans. Only movie clips? Why would that be. Seems like if they could get at peoples real thoughts. like the faces of loved ones the images may be better displayed. Only movie clips?

And Pink Panther for God's sake? No wonder it's a blurry mess.. who wants to remember that?

edit on 23-9-2011 by GogoVicMorrow because: (no reason given)

You make a good point here. I think the current limitation is because they have to map images to activity in the brain. They then compare activity to the vast library of images that they have and a computer combines the closest match to make the output. Similar to how your brain decides what you are looking at, you reference memory.

Originally posted by GogoVicMorrow

Could never be used against you. .

Hahaha, nice one.

Denial will only get you so far in the American Police State. Soon you will have to accept that these will be used against you.

Originally posted by BIGPoJo

Originally posted by GogoVicMorrow
reply to post by BIGPoJo

 

Currently, researchers are only able to reconstruct movie clips people have already viewed.

Ah.. I smell shenanigans. Only movie clips? Why would that be. Seems like if they could get at peoples real thoughts. like the faces of loved ones the images may be better displayed. Only movie clips?

And Pink Panther for God's sake? No wonder it's a blurry mess.. who wants to remember that?

edit on 23-9-2011 by GogoVicMorrow because: (no reason given)

You make a good point here. I think the current limitation is because they have to map images to activity in the brain. They then compare activity to the vast library of images that they have and a computer combines the closest match to make the output. Similar to how your brain decides what you are looking at, you reference memory.

The software finds correlation between a subset of video frames (what the test subjects watched over the sample period) and neural activity. The scientists prob. don't even understand the algorithms here, it's not always necessary, like in the case of neural networks.

The software was then allowed to pour over 18 million YouTube clips, considering how their frames would likely correlate to neural activity based on the relatively small sample data set (from the viewing data). This massive video set does NOT include the video that the viewers eventually watch for reconstruction.

Obviously there is no 1:1 matching video or source data for what we dream/visualize in the real world -- that's why the video they show the viewer for reconstruction isn't in the pool of video it can access when making the reconstruction. It is amazing that the reconstruction is so close to the source feed in such an early implementation.

The coolest thing about this is that the more data the software is fed (viewer feedback data + bigger video data set for reconstruction), and the more precise the hardware/biometrics gets, the closer it will get to 1:1. Eventually it won't even use video for the reconstruction (voxels please!), this was just a convenient way to do version 1.0.

reply to post by Exuberant1


There are a few problems with the alarmist paranoia.

First - this requires users to sit still in an fMRI for hours at a time.

Second - it only 'taps' what you are currently viewing. It does not work for thoughts/memories. And likely never will, as thoughts and memories are far too specific to each individual.

Third - the technology, as implemented, will never be precise enough to pull off much better, in terms of pictures. fMRI maps blood-flow, which is simply not precise enough to pull more data. Electrode implantation would be required for more detail.

So.. no... this is not capable of being used against you - unless you're already brain-dead and don't realize you're in an fMRI for hours on end and watching home-videos of self-incrimination.

Great now we really will need to wear tin foil hats....

omg is this real? I could record my dreams?

Finally we can do something about corruption within government and finally.......nobody would find it strange when I wear my alu-foil hat.