A few days ago we learned about Neuralink’s first human brain implant, and the hornet’s nest began to stir again. News about brain implants seems to be coming from only one direction. However, nothing could be more contrary to reality. It’s time to take stock of where we are and how far neurotechnology has come.
The study of the brain is one of the greatest fields of knowledge. Everything we are and do is determined by the activity of the brain. Understanding how we imagine the world, what we feel and why we act is part of this important life-changing issue.
Neuroscience has been trying to unlock the secrets of our governing body for decades. Using animals traditionally, we have made progress in understanding the brain code, that complex sequence of electrical commands that determine action. This code is so complex that even the simplest tasks are difficult to understand.
Decoding brain activity
When we lift our finger, an interconnected network of neurons fires hundreds of milliseconds earlier in the supplementary motor cortex, quickly connecting with other neurons in the prefrontal cortex. These preparatory signals in turn activated neurons in the basal ganglia, providing information about the intensity of movement, and in the cerebellum, where they integrated sensory information and calibrated the possible error. All this activity converges on the main motor neurons, which send direction and force signals to the endings of the index finger.
In cases where the body is disconnected from the brain, such as spinal cord injury, amyotrophic lateral sclerosis, or ALS, understanding these neural commands may allow robotic arms or a cursor to be moved on a device screen. This is exactly the problem that brain implants are trying to solve.
Neural activity associated with movement was first deciphered in the 1960s by neuroscientist Eberhard Fetz of the University of Washington. By recording the activity of several neurons in the monkey’s motor cortex, he was able to report the position of the cursor on the screen.
In an article published in The scienceFetz showed that monkeys can even learn to voluntarily modulate the activity of their motor neurons. That is, not only can we use neural code to perform actions external to the body, but the brain itself can learn to do it skillfully if we effectively connect it to these devices.
Brain-machine interfaces
The interface between the brain and the machine consists of a number of components designed to read, process and interpret neural activity. To read this activity, we need electrodes capable of resolving single neuron firing on the order of a few microvolts. To process it, we need to filter and clean the signal. To interpret it, we use artificial intelligence (AI) algorithms that can learn from data.
The decoded signal is converted into control commands for external devices. For the brain to learn, a feedback signal (or Feedback) visual, tactile or auditory, which allows errors to be calibrated.
One of the first successful human-brain-machine interfaces was developed by John Donohue’s team at Brown University, giving rise to the BrainGate technology published in Nature in 2006 year. Using electrode arrays in the motor cortex, Mark Nagle, the first paraplegic to receive this brain implant, was able to control a computer cursor. Although these early prototypes only allowed simple movements, they marked the beginning of a paradigm shift.
Let us now imagine that we want to understand the motor commands underlying the writing process. The complexity may seem overwhelming. It’s not just about moving a finger, but replicating the dexterity of the hand to draw letters one after another.
Last year, using the same BrainGate technology combined with new artificial intelligence and data science algorithms, it was shown that it was possible to directly read the intent of writing. Patt Bennett, a 67-year-old woman with multiple sclerosis, could write while thinking about it. It is estimated that several dozen patients with similar implants are currently enrolled in regulated clinical trials around the world.
Read more: Realities and problems of merging brain and machine
Much more than a finger
As the above examples show, the race for neurotechnology began many years ago with neuroscience. Without fundamental science it is difficult for applied science to exist. Companies like Blackrock Neurotech have been pioneers, betting on interfaces to help restore motor function in patients with spinal cord injuries or ALS.
Since then, the neurotechnology market has grown at an accelerated pace, with one study estimating that it could reach around $30 billion by 2030. Part of the progress involves AI processing and interpretation of signals, as well as the development of new, less invasive recording systems.
Recently, scientists at the University of California used electrode patches on the cortical region responsible for orofacial movements to decipher speech intent in a paralyzed 47-year-old woman. Here the signal consisted not of isolated impulses from individual neurons, but of electroencephalographic, or EEG, rhythms resulting from their coordinated activation. The decoded signal was projected onto an avatar, which the patient could use to convey a series of simple sentences from a set of 1024 words. This example, published in Nature in 2023, illustrates the potential and limitations of this problem.
In the coming years, we will go further and further in our ability to decode brain activity not only by isolating neurons, but also by understanding electrical field fluctuations or even other physiological signals such as facial or pupil movements. Companies like spin off The Spanish company INBRAIN has developed graphene electrodes that allow recording slower EEG rhythms, and the American Synchron is testing stent to capture brain activity from arteries.
Read more: How far are we from a scientific point of view from reading minds?
Other applications of neurotechnology include monitoring epileptic seizures, treating Parkinson’s disease, or the recent success of brain-spinal interfaces allowing paraplegics to walk again, as published by scientists from the École Polytechnique Fédérale de Lausanne. Nature.
In less invasive areas, the development of non-clinical applications in the field of external device control allows us to approach other types of innovative niches, such as hands-free technology, video game development or the field of education.
These advances demonstrate the innovative potential of neurotechnologies, highlighted in recent reports such as the one published by the Directorate of Science and Technology of the Spanish Congress of Deputies. The work of patients, scientists, doctors, technologists and companies is essential to the enormous challenge of understanding the brain and treating some of its diseases.
If there’s one thing we can thank Neuralink for, it’s the growing public interest in understanding these advances. With this comes the responsibility to publish reports without fanfare, giving a voice to basic science and its many constituents.