The Science Behind Elon Musk’s Neuralink Brain Chip

Neuralink, Elon Musk’s brain chip, recently pushed back on claims that they violated animal welfare laws a few years ago, while testing on monkeys. This year, the company plans to test on human subjects. What does this mean for brain implant science?

Video Transcript

- Neuralink, Elon Musk's brain chip company recently pushed back on claims that it violated animal welfare laws a few years ago, while testing on monkeys. This year, the company plans to test on human subjects. But when it does, what would this major step mean for brain implant science? PAUL NUYUJUKIAN: Academics, like me, have conducted clinical trials in people with brain implants. - Dr. Paul Nuyujukian is a professor of bioengineering and neurosurgery. He directs the brain interfacing laboratory at Stanford. PAUL NUYUJUKIAN: For about 20 years now, academic research brain implants, up until this point, more or less, have almost exclusively been with wires. The difference that the one has, that the Neuralink-- it's fully implantable. It is battery powered. It is wireless. All of this is being done over Bluetooth protocol. - Let's dive into the science behind Neuralink to understand how exactly human brain chips work. [MUSIC PLAYING] PAUL NUYUJUKIAN: The science behind how these implants work is not that different from how you would go about trying to measure the energy from a AA battery. It's the same principle that we're doing with these brain implants. This is called electrophysiological recording. When you move your arm to the right, certain sets of neurons are activated in a certain pattern, listening in to that activity and that pattern. You can predict, very quickly, which direction the arm is going to move. These are the neurons that are directly wired to your muscle. - Unless that pathway from the brain to the spinal cord to the muscle is damaged, the way it is in patients with paralysis. PAUL NUYUJUKIAN: That pathway is damaged, then the neural signals, the signals from the brain, aren't going to get down to move the muscles. But in many cases, the signals are still present in the brain. They're just not getting out. So if you reach in and put something that listens in to those neurons, then you know what's happening to the muscle. - And that's the goal of a brain implant. Now let's look at a timeline of brain interface breakthroughs over the years. Scholars have long been interested in how the brain works, so it's important to view these new developments at Neuralink as a culmination of breakthroughs by brain machine interface researchers, especially in the last few decades. For example, in 2002, the first demonstration of real time cursor control in monkeys took place. 2008, a monkey controlling a robotic arm in 3 dimensions fed itself. 2012, the first brain controlled robotic arm by a human. 2017, a human controlled a cursor mentally to type out words and sentences. Dr. Nuyujukian was part of the study, as well as the one in 2018, where a human subject mentally controlled a tablet to do things like browse the web, send emails, and play games or music. PAUL NUYUJUKIAN: All that's been done with a couple of electrodes. - But in 2019, Neuralink a private company changed the game when it unveiled a pig named, Gertrude, with a wireless implant that monitored about 1,000 neurons. [MUSIC PLAYING] ELON MUSK: The neurons are like wiring. And you kind of need an electronic thing to solve an electronic problem. PAUL NUYUJUKIAN: That was a very interesting moment because it signaled to the community that they're serious. They're investing. They're building hardware from scratch. And they're putting it in large animals. - For the pig, the electrodes were implanted in somatosensory cortex, allowing them to measure sensory activity, like that of taking a step. PAUL NUYUJUKIAN: Every time that particular neuron they were listening to fired, you would hear this little pop or click from the audio channel. And so the moment I heard it, right? It's like, oh, yeah, cool, they got neurons. You just recognize it instantly. You know what neurons sound like if you've been listening to them for decades. And that's what they were communicating. They were telling the field, we've got neurons. Pay attention. - And overnight, it seemed the industry took notice. Then in April of 2021, Neuralink released the so-called, "MindPong" video. PAUL NUYUJUKIAN: Pager was the name. It's a rhesus macaque, which is the type of monkey that is very commonly used in this field, implanted with two of the N1 devices, the Neuralink devices, performing brain control of a cursor on a screen. That's extremely significant because here, Neuralink is showing their new hardware, their new device, in their hands, works in a monkey. That's the level that's necessary to convince the scientific community, to convince the FDA that you're ready to go into human clinical trials. That's the evidence the FDA is looking for. - The recording power of the N1 device in Pager was eye opening because of the sheer number of individual electrodes that had been implanted. PAUL NUYUJUKIAN: There was definitely a lot of clever engineering that went into that, to build a device that can transmit 2,048 electrodes worth of spiking information, right? Of digital 1s and 0s of spikes over a radio, wirelessly. And when you have that many channels, the performance that you should be able to get should eclipse what we've been able to do in the academic field. The maximum number of electrodes I've ever recorded from is 200 to 300. - So with all those electrodes, how does a device like the N1 get implanted in a subject's brain? [MUSIC PLAYING] PAUL NUYUJUKIAN: Make no mistake, this is neurosurgery. It is not a joke. This requires cutting the skin, getting down to skull, drilling a hole in the skull, exposing what's called the dura, which is this protective layer of tissue that surrounds the brain, cutting the dura, folding it back to expose the brain. And then you get to the surface of the brain, where you can implant the electrodes. The biggest risks with these types of techniques are infection, bleeding, and tissue damage. - So what would it take for the FDA to approve clinical trials in humans? PAUL NUYUJUKIAN: The Neuralink device are called class III medical devices. They are implantable. And they're going into very sensitive body cavities. That is the highest level of scrutiny that the FDA assigns to medical devices. They don't have a predecessor. There's no previous example. That's approved. And so very appropriately, they've got a high bar they have to cross in order to get it approved. - So what Neuralink has to do next is prepare a very long and technical document with all the evidence from animal studies that their device is safe and effective. This document is submitted to the FDA, who has 90 days to review and give them an answer. If the FDA says yes, then their clinical trial is approved, and Neuralink can enroll and recruit human participants. PAUL NUYUJUKIAN: We are on the cusp of a complete paradigm shift. This type of technology has the potential to transform our treatments, not just for stroke and paralysis and degenerative disease, motor degenerative diseases, but also for pretty much every other type of brain disease from Parkinson's to epilepsy to dementias, Alzheimer's, and even psychiatric disease. Seeing Neuralink and the other companies in this space start an industry around neural engineering, brain machine interfaces, neural prosthetics, has been a tremendous amount of validation for neuroscientists and engineers, who have been working in this space for decades. How much happier could the scientific community be than to give birth to an industry? - So will this industry someday lead to the creation of cyborg humans with superhuman intelligence? PAUL NUYUJUKIAN: There's all sorts of wild speculation in our field. I think science fiction is wonderful at telling very creative and captivating stories about all sorts of things, including brain machine interfaces. The reality is we are in such early stages of this space, right? Where we are just barely able to record from neurons that control muscles and try to interpret something, gleaning meaningful information out of that. We're going to be in that space for decades. That's where I will focus much of my career is understanding what's going on with these neurons and the circuits that they are working on. That's where the last 15 years of my work has been. And the coming several decades of my work will focus in on this space because that's going to be the forefront of neuroscience. The rest, I think is fun to think about. But I don't see how that's going to be in the foreseeable future. [MUSIC PLAYING]