Neurolink: Get ready for a new subspecialty of neurological disorders.

This forum made possible through the generous support of SDN members, donors, and sponsors. Thank you.
One of my buddies is in this field. As I think is usual, Elon Musk is resting on other people's shoulders and taking all the credit.

Here's an article about this, a bit sad in a "Flowers for Algernon" way: https://www.newyorker.com/magazine/2018/11/26/how-to-control-a-machine-with-your-brain

Please check out the video in the article, somehow can't link directly.

You'll notice Musk isn't mentioned. Many people are working on this, there are going to be many pitfalls and hopefully MANY rewards. I'd love to see this get developed over my career and even be part of the trials to get it on the market. This has the potential to help many stroke patients and obviously spinal lesions would the first target.
 
One of my buddies is in this field. As I think is usual, Elon Musk is resting on other people's shoulders and taking all the credit.

Here's an article about this, a bit sad in a "Flowers for Algernon" way: How to Control a Machine with Your Brain

Please check out the video in the article, somehow can't link directly.

You'll notice Musk isn't mentioned. Many people are working on this, there are going to be many pitfalls and hopefully MANY rewards. I'd love to see this get developed over my career and even be part of the trials to get it on the market. This has the potential to help many stroke patients and obviously spinal lesions would the first target.
He probably financed it all. That's a very important thing
 
He probably financed it all. That's a very important thing

Dude, it took me less than 10 seconds to find this link: Andrew Schwartz, PhD

About Dr. Schwartz, the dude who's work was highlighted in the New Yorker article: "In collaboration with clinicians at Pitt and UPMC, he is a co-principal investigator on two new federally-funded projects that will put brain-computer interfaces into patients with spinal cord injury to see if they will be able to control assistive devices, such as a prosthetic arm."

At THIS time this particular research is federally funded. As I alluded to above, I know some people in this space. One told me that upon meeting the neuralink folks, one immediately got a sense that they didn't know anything about the brain and didn't care to know.

Funding IS very important. I agree with this. But the funding must go into real science, which sometimes works, sometimes fails. Funding a vanity project won't do anyone any good. Currently these n=1 trials appear perfectly suited for federal funding. A device that rolls out into a commercial trial will need more funding and a device company to assume risk/reward. Perhaps Neurolink will get there first.
 
Dude, it took me less than 10 seconds to find this link: Andrew Schwartz, PhD

About Dr. Schwartz, the dude who's work was highlighted in the New Yorker article: "In collaboration with clinicians at Pitt and UPMC, he is a co-principal investigator on two new federally-funded projects that will put brain-computer interfaces into patients with spinal cord injury to see if they will be able to control assistive devices, such as a prosthetic arm."

At THIS time this particular research is federally funded. As I alluded to above, I know some people in this space. One told me that upon meeting the neuralink folks, one immediately got a sense that they didn't know anything about the brain and didn't care to know.

Funding IS very important. I agree with this. But the funding must go into real science, which sometimes works, sometimes fails. Funding a vanity project won't do anyone any good. Currently these n=1 trials appear perfectly suited for federal funding. A device that rolls out into a commercial trial will need more funding and a device company to assume risk/reward. Perhaps Neurolink will get there first.
Who funded neuralink?
 
Who funded neuralink?

Mostly Musk. Neuralink is "a company in which Mr. Musk has invested $100 million."


Again, The U Pitt work is being done on NIH grants. If and when they do technology transfer, perhaps they will join with Neuralink.
 
As someone who used to work in a brain computer interface lab, I can tell you we are still a long way off from this dream. If you go to conferences with these scientists from different labs, after the jaw-dropping stuff is presented they will later mutter under their breaths to each other in small circles, "how long are your able to get good recordings from your electrodes?". That Utah array that's recording from some 80 separate neurons initially is only able to record from a handful by 6 months. The reason is that the brain treats the electrode as a foreign body and starts to form a glial scar around it. To my knowledge, no one has solved this problem. Until someone comes up with a truly long-term biocompatable electrode this technology is not that useful clinically. It will happen, but we are probably 20-30 years from it becoming a reality. Throwing money at the problem helps, but can only speed things along so much.
 
Watch out for those genius folks at Neuralink!

1599238066970.png


Was on this site before they fixed it. Guess they might want to reprogram that implanting robot.
 
As someone who used to work in a brain computer interface lab, I can tell you we are still a long way off from this dream. If you go to conferences with these scientists from different labs, after the jaw-dropping stuff is presented they will later mutter under their breaths to each other in small circles, "how long are your able to get good recordings from your electrodes?". That Utah array that's recording from some 80 separate neurons initially is only able to record from a handful by 6 months. The reason is that the brain treats the electrode as a foreign body and starts to form a glial scar around it. To my knowledge, no one has solved this problem. Until someone comes up with a truly long-term biocompatable electrode this technology is not that useful clinically. It will happen, but we are probably 20-30 years from it becoming a reality. Throwing money at the problem helps, but can only speed things along so much.

The thing is that we already have these amazing biocompatable electrodes in place that amazingly detect brain signals, code them as electrical signals, and even transfer the code into activators. We call this system the CNS and PNS. Very remarkable tech.

Musk really is today's Barnum. I can't believe people buy Teslas.
 
As someone who used to work in a brain computer interface lab, I can tell you we are still a long way off from this dream. If you go to conferences with these scientists from different labs, after the jaw-dropping stuff is presented they will later mutter under their breaths to each other in small circles, "how long are your able to get good recordings from your electrodes?". That Utah array that's recording from some 80 separate neurons initially is only able to record from a handful by 6 months. The reason is that the brain treats the electrode as a foreign body and starts to form a glial scar around it. To my knowledge, no one has solved this problem. Until someone comes up with a truly long-term biocompatable electrode this technology is not that useful clinically. It will happen, but we are probably 20-30 years from it becoming a reality. Throwing money at the problem helps, but can only speed things along so much.

The Neuralink Demo few days ago was obviously underwhelming for us Neurologists. But Musk has the ability to get many people interested in whatever he delves into. This will encourage more people to work on it and increase investments.

Clinically its not much, but technologically it was pretty impressive, I think they had 1024 channels in that tiny electrode. And I think they demonstrated some biocompatibility too, as one of the pigs had it for 2 years or something. So its a good start.
 
Isn’t this basically an intracranial EEG?

I didn't see the demo described above, but my understanding is they are implanting electrodes directly into cortical tissue rather than having them lie on the brain surface. If that's the case, the answer is no - recording from a single neuron is world's apart from EEG recordings. With EEG or the intracranial version which is called ECog you are recording from a huge number of neurons simultaneously. ECog has higher resolution than EEG and you can make out someone's intention to move their thumb or fingers for example, but directionality and things of that nature tend to be lost. Generally EEG and ECog are extremely crude.

Recordings from a single neuron, on the other hand, are remarkable. Each neuron in the motor cortex is directionally tuned. Recording from many of these individual neurons simultaneously allows you to decipher complex behaviors. Not only that, monkeys can be trained to change the directionality of particular neurons in a matter of minutes. If we had truly biocompatible cortical electrodes many of the things relegated to science fiction could potentially become a reality.
 
Top