On Tuesday, 16th July Neuralink made an announcement about their progress and N1 chip. Elon Musk and his amazing team explained in a lot of depth exactly how Neuralink’s neural lace works. I recommend that you watch the 1:44:41 long video before reading our coverage of the Neuralink event.
The goal of Neuralink
Though the launch event was for the purpose of recruitment, Neuralink itself is much more than that. According to Mr Musk, Neuralink’s aim is first to help those with debilitating brain or spinal conditions and later to create a digital superintelligence symbiosis to stave off the existential threat of AI. While this is the ultimate goal, it will apparently take many years to happen, with FDA approval likely being the slowest step.
How does neural lace work?
As shown in the brief summary (at the start of the video and at 8:31), neurons communicate by electrical pulses. The pulses create electromagnetic waves which can in turn induce electricity in an electrode. This means that by inserting many thousands of microscopic conductive wires into the brain, a machine can detect the activity inside the brain.
As with most signal receivers, there is a lot of noise so it’s hard to tell what is happening without some sort of algorithm. Thankfully, scientists have discovered what is known as a neuron ‘spike’. Using algorithms to separate these spikes from background noise, scientists are able to attribute spikes to specific neurons and use the data to decode what that neuron is attempting to do. Together with other neurons you can gain a map of a person’s cognitive function and use that to control devices such as a mobile, keyboard or mouse. Neural lace also allows for information to be inputted. For example, inputting a haptic feedback sensation would make synthetic typing feel more real, which helps the patient learn how to control the device.
Neuralink’s first product – N1
Neuralink has been working secretively for two years. In that time, they have come up with 15 chips, with 8 of them being produced. Their latest design, the N1 sensor, is what will be used in their first clinical trials and (probably) commercially available product. The N1 System 1 chip consists of 1024 threads that can both read and write the brain, 1024 ‘analog pixels’ that decode each neuron and 64 stimulation engines. Each pod fits in a 8mm diameter X 4mm height cylinder. First clinical trials will have four N1 sensors, with the commercial version having ten.
From the above image, we can see how the N1 implant will be installed. The four N1 sensors are connected by a thin wire to a conductive plate behind the ear. This plate is in turn connected to a wearable device, called the ‘Link’. This link is the only part of the system with a battery or communication capabilities, meaning that removing the Link turns off the system. This does raise some security concerns but is a good emergency shut-off.
Clinical trials are set to begin in 2020. For now, Neuralink will continue to do animal testing, even hinting at monkeys that can already control computers with their brains. What we did learn from the Neuralink event is that progress will be slow in comparison to Elon Musk’s other ventures (though fast compared to the rest of the medical world). Neuralink will begin by helping disabled people and eventually create AI-human symbiosis. There is a lot in store for Neuralink, even 3D simulated reality is on the menu.
What the event does highlight however is that there are plenty of capabilities, concerns and cornerstone technologies in the pipeline. Anything from digital security to gaming and simulated intercourse is up to speculation, exactly what Generation Byte is here for. On such a note, do come back for a super in-depth analysis of the event, publicly released research paper and more as we have barely scratched the surface of Neuralink so far.