Neuralink: Can Connecting our Mind to Computers Really Solve all of our Problems?
Elon Musk has made a lot of eyebrow raising claims about the intended capabilities of Neuralink. How much of it is rooted in the real science?
The Intro
Human beings are naturally intelligent creatures who learn and adapt to the environment in which they live. One of the most significant changes in our lives is how we are adapting to the technology that surrounds us. Over time, we’ve developed a symbiotic relationship with technology; it has become an integral part of our lives, with a profound impact on how we process information, communicate with one another, and organize the world around us. As technology continues to advance at an exponential rate — much faster than humanity’s evolutionary biology can keep up — there is concern that we might be entering an era where people are no longer as able to solve problems independently because they have become reliant on technology.
This is the current state of the world’s dependence on technology, and we are at a precipice where more advanced technologies might be introduced to further alter our brains. Neuralink, a company by Elon Musk has recently entered this arena, and they are developing brain-computer interfaces which allow neural information to be uploaded and downloaded from the brain. They aim to develop this technology in order to connect our minds with computers within a few years to enhance our brains and even solve problems such as depression and anxiety.
The Technology
These brain-computer interfaces use electrodes attached to the neurons of the brain, which are either implanted in the brain or inserted via needle. These electrodes are connected to a recording device, which captures electrical activity from the neurons. Then this activity can be decoded into a readable pattern that is inputted into a computer or other electronic device. The information contained in these patterns can then be used to move a cursor or control the movements of an artificial limb, for example.
Reading information from the brain is not anything new. Technologies, such as electroencephalography (EEGs) and magnetoencephalography (MEGs), have been used in clinics for decades because of their ability to monitor brain activity. EEGs measure the electrical signals on the scalp by using electrodes attached to the head, while MEGs measure these signals inside the brain by using sensors attached to the scalp. With these technologies, we have been able to map out the brain and understand which areas of the brain are responsible for certain functions, such as movement and vision. We’ve also been able to learn about how different parts of our brains communicate with one another, and even develop better prosthetic limbs for amputees. In fact, computers that interface with the human brain through electrical signals have been around as early as 2006 and a study that allowed monkeys to control a robot through a device implanted in their brain dates back as early as the 2000.
The Breakthrough
What Neuralink has done to advance this field is developing an implant that’s less invasive and collects more data. They’re also adding more functionality to these types of devices such as wireless connectivity. They’re essentially streamlining the technology that interfaces between humans and computers and making it smaller, safer and easier to insert. This is a monumental step, because it will allow for greater adoption and the collection of more accurate data about neural activities on a scale that has never been realized before.
As with any machine learning model, more inputs and information mean more training and a better model. With enough training data, Neuralink may eventually be able to map and predict intended behaviours with their respective brain signals with a high degree of accuracy. In their most recent August 2020 demonstration, they showed off how accurately a pig’s brain activity was able to predict it actions. One can only imagine the types of implications this has for technologies like prosthetics, wearable computers and even mind-controlled autonomous vehicles.
The Problems
These advancements alone are nothing short of breakthroughs in the field. Musk’s promises, however, don’t stop there. Memory loss, brain damage, depression, anxiety and addiction are among the many other problems that Musk hopes Neuralink could solve with this new technology. He hopes that this technology would help us create a symbiotic relationship with AI.
Here’s where I get skeptical about these promises. Neuralink’s fundamental shortcomings don’t lie with their ability to read information to the brain. Rather, the issue lies with the ability to write data (something we would have to figure out before we solve problems such as depression and anxiety). While the electrical activity of the brain can be read in real-time, it’s not clear how to write information into the brain. This is because the brain is not a hard drive. It is not made from digital memory, but rather it is organic and as such, it can’t be overwritten like a computer. The closest that this technology has come to writing information into the brain are various experiments dealing with deep brain stimulation which has been used to treat Parkinson’s disease and other disorders for decades. Even then, the exact mechanism behind the success of deep brain stimulation isn’t fully understood.
On top of this, science is still unclear as to what the true causes of diseases such as depression and anxiety are. It’s clear that they aren’t caused by a chemical imbalance in the brain, because there is no evidence that implicates a single chemical. Rather, it is more likely that these conditions are caused by an imbalance of many different chemicals. This makes neural treatments even more difficult to achieve individually because these compounds may not all be affected by the same electrochemical processes in the brain.
Take one look at a psych textbook and you’ll find that the answers to a lot of our fundamental questions about mental health are a lot less clear than one might think. Correlating brain activity with physical actions with machine learning is one problem. Scientists call machine learning techniques such as neural networks “block box solutions” because all you have to do is provide enough training data to get a desired output without having to understand the complex relationships between the input variables. Interacting with brain chemicals at a neurotransmitter level to solve complex problems is a completely different problem. It requires a much more nuanced and fundamental understanding about the mechanisms behind some of our most complex brain interactions. The problem is, we simply don’t know that much about the brain.
The Reality
The fact of the matter is, the field of neuroscience has not caught up to Neuralink’s ambitions. While Elon Musk has made promises such as merging one’s brain with an artificial intelligence, the reality of the matter is that even simple brain-computer interfaces are still a long way off from being a viable product. Neuralink’s new surgical implantation devices are nothing short of an engineering marvel, but can’t engineer your way out of every problem. The engineers at Neuralink have come face to face with the problems that psychologists and neuroscientists have been facing for decades: learning about the mind takes long periods of study and observation, and we have only just begun to scratch the surface.
Neuralink has set their sights incredibly high, and many aspects of their plan are far ahead of what scientists have discovered about the brain. But that’s a good thing: ambition is what drives neuroscience discovery, and if you don’t reach for the stars you can’t get anywhere. Unfortunately, however, there are no shortcuts to mental health treatment. It will require a combination of technological advancement and multidisciplinary collaboration between physicists, chemists, biologists and psychologists to finally crack many of these complex problems.
