Scientists Closer to Reading Words From Your Brain | Hacker News Reader

Scientists Closer to Reading Words From Your Brain

singularityhub.com

23 pointsjamesbritt11 comments

I really do believe the next big disruptor (on the scale of something like space travel or the internet) is going to be in our ability to harness brain power to directly manipulate our world.

There is so much research going on and a lot of it is quite advanced. Some day we'll put it all together and the results are going to be amazing. Not only in curing disease, but in fundamentally changing how we interact with our machines.

I can't wait.

I really do believe the next big disruptor (on the scale of something like space travel or the internet) is going to be in our ability to harness brain power to directly manipulate our world.

Can't we already do that? I mean, you might have to work a huge machine using a joystick or something, but would eliminating the joystick be that much better?

I see it more as way to prevent carpal tunnel syndrome for programmers.

What's much more interesting for me is the ability to take a full image of someone's brain. It would be an interesting type of backup...

Do you feel that is possible?

50 +/- 50 or so years from now, maybe.

That's some interesting research they're doing. I'd be be more interested to see if somebody can translate it to an electroencephalographic device. I suspect that placing electrodes directly onto the surface of the brain would be a dealbreaker for those of us without an immediate need to have our thoughts processed by a computer in real time (however cool that concept might be).

There's an external headset that "unwrinkles" your brain algorithmically and gets very accurate readings without requiring direct contact with your brain. http://www.ted.com/talks/lang/eng/tan_le_a_headset_that_read...

Hi, neuroscientist here. The stuff about it "unwrinkling" the brain is nonsense. Emotiv is a running joke in the scientific community, unfortunately. EEG has some fundamental problems as a control mechanism, and isn't interesting from a brain-machine interfacing perspective. It's unfortunate because we all want to see neural interfaces, but it will either have to be invasive or come from a major Nobel-worthy innovation.

Ah, thanks! Still, it's a fun video :-)

Hmmm, their demos look good. Are you saying the headset doesn't work at all, or not well enough, or not for a broad range of tasks?

EEG measures mass synchrony between different areas of the brain. That's extremely hard to control consciously and takes a lot of practice. Even when you can do it, it's a 1-dimensional signal. It's not even clear you want to control it consciously; EEG therapy is used in psychiatry every now and then for treatment of a few disorders, and can lead to personality changes. The "inverse EEG" problem, which relates to the distortion caused by the skull, is basically computationally intractable (which is why ECoG is used). EEG is also totally drowned out by EMG (muscles) if you aren't careful, and it's nearly impossible to tell EMG and EOG (eye movements) apart from "real" EEG.

In Emotiv's specific case, though, these fundamentals are the least of their problems. They're three (four?) years behind schedule without anything really shipping even now, and the units that I have seen in the wild are of low build quality.

What if you don't think in words? What if your natural mode of thinking is in terms of fuzzy images and funky metaphors? Looks like I'll have lots of gainful employment in the new dystopian future.

Reply on news.ycombinator.com