[ 3 / biz / cgl / ck / diy / fa / ic / jp / lit / sci / vr / vt ] [ index / top / reports ] [ become a patron ] [ status ]
2023-11: Warosu is now out of extended maintenance.

/sci/ - Science & Math

View post   
File: 145 KB, 477x611, 1270787342089.png [View same] [iqdb] [saucenao] [google]
5058145 No.5058145 [Reply] [Original]

>http://www.medicaldaily.com/articles/12123/20120914/monkeys-made-smarter-prosthetic-device.htm

Beyond the usual transhuman stuff Im sure /sci/ talks about all the time, was anyone else shocked at:
>The device then electrically stimulates the decision-based neural activity by playing the recording back to the same part of the brain.

Since when are we directly stimulating neurons with any meaningful pattern fidelity?

>> No.5058155

For awhile now I believe.

>> No.5058190

>>5058155
Could you give a laymans explanation of how it works? Do they wrap something around the axon or poke into the nucleus or what?

>> No.5058201

>However, when the prosthetic device was turned on during these "drug sessions," the device detected when the monkeys were likely to choose the wrong image and played back the previously recorded "correct" neural patterns for the task.

Uh, this is convincing the monkeys what to choose in a given situation.

Fifty years of hard work might make this mind control.

>> No.5058207

>>5058201
Indeed it does seem like a bigger deal than the news sites are treating it as. Can they record the monkey eating an apple, then replay it later and he'll experience some of the same sensations? How many neurons can they hook up to? Does it only work on thoughts that have a very narrow pathway? Can they counteract signals in real time, but still record them and play them back slightly later and mess up their reaction time?

The really interesting part to me is the weight they added to the decision they wanted, and boosted him up above baseline level. That shit is bangin.

>> No.5058220

>>5058207

Hmmmm.

HMMMM.

The possibility for brain augmentation is... interesting.

It would require a fuckton of complex programming, though, I think.

It might be similar to the programming for face-recognition software, but for brainwaves. It might function a little like that type of texting where it figures out what letter you want to type by context.

But, er, I'm getting a little ahead of myself.

Regardless, this is really interesting.

>> No.5058224

>>5058190
Afaik, they record the electrical signals of a group of neurons in the front of the brain by using multiple recording wires (if you will) that are most likely inserted directly into the brain cell. these wires send the info back to a computer which saves the data. With the wires plugged into the brain still, the pattern can be re-submitted to the brain via the computer. These electrical signals then activate the recorded sequence of events thereby forcing the same actions to repeat. Brain cell signals can be sent by electrical pulses, which are what are being sent via the wires in this case. Thats my best layman explanation via my phone, lol.

>> No.5058229

>>5058224
neuron = brain cell

(in my explanation)

>> No.5058238

>>5058207
The.microelectrode arrays typically only have a hundred or so connections afaik (i may be off) because I cant access the papers from my phone). Therefore they dont commect to a huge amount of.neurons.

Synaptic weighting is indeed very important to computational neurosci.

>> No.5058240

>>5058220
I think one big challenge would be to program what patterns of neurons firing to look for and correct, without having to record them first.

Imagine and augmented reality HUD that has that microsoft paperclip mascot pop up and say
>it looks like you're considering eating at mcdonalds! You specified you are on a diet. Would you like me to change your mind?

>> No.5058242

>>5058238
Well thank you for your time, anon. it sounds like you know about this stuff.

>> No.5058246

What about pre-recorded memories, experiences, or "knowledge" packs crammed straight into your neurons?

Oh gosh I'm so hard right now

>> No.5058248

>>5058242
Its my area of concentration. I hope to get a position in one of the labs next year.

>> No.5058258

>>5058246
The recorded data is based on actual physical systems, and although brain cells can rewire themselves in response to stimuli, I would be quite cautious of any such notions. The brain is very sensitive to change, therefore such a.procedure may be quite dangerous

>> No.5058262
File: 37 KB, 393x264, Untitled-2.jpg [View same] [iqdb] [saucenao] [google]
5058262

>>5058246
>What about pre-recorded memories, experiences, or "knowledge" packs crammed straight into your neurons?
Show me.

>> No.5058273

>>5058246
probably brains vary too much from person to person on the neuron level. However on large scales we all have common brainwave signatures for certain things (ie the P300 wave when we recognize something) but I guess on that macro scale you dont know which neurons are causing the signal

>> No.5058291

>>5058246

If they had time to study your neurons for a long while, maybe.

You might as well just learn it in that time period though.

>> No.5058295

>>5058291
Like training those voice to text programs like Dragonspeak

>> No.5058316

>>5058291

Well it'd obviously have to be computerized...

>> No.5058318

>>5058291
>>5058295


There's actually ways to associate neural activity to specific images and such, I can't remember what the technique is called but it's possible to get "real" images with enough study.