Brain-computer interface: will the fairy tale become a reality?

Mind reading will become a reality
Marina Muravyeva, STRF.ruIs it possible to learn how to decode human and animal thoughts?

Is it really possible to create devices that transmit thoughts to machine devices and purposefully control behavior by affecting the neurons of the brain? Konstantin Anokhin, a leading Russian neuroscientist, corresponding member of the Russian Academy of Sciences and the Russian Academy of Medical Sciences, head of the Laboratory of Memory Neurobiology and the Department of Systemogenesis of the Research Institute of Normal Physiology, answers all these questions in the affirmative. And although it may seem fantastic so far, serious research is already underway to create technologies that bring these incredible opportunities to life.

For the first time in hundreds of years, people have the opportunity to explore the objective neural mechanisms of the mind, intelligence and thinking. As Christof Koch, one of the leading experts in the field of the neurobiology of consciousness, said, we live in a unique period in the history of science, when technologies that allow us to find out how the activity of the objective brain leads to the emergence of subjective reason become a reality. Brain–machine interfaces are a concrete embodiment of these capabilities. A review of scientific achievements in this direction was presented by Konstantin Anokhin at a lecture at MEPhI "Neurons and behavior: the basics of creating brain-machine interfaces".

The experiments of Brazilian neuroscientist Miguel Nicolelis, initiated in the late 1990s, have gained the greatest fame in the field of the development of brain-machine interfaces. By inserting several electrodes into the monkey's brain, the scientist achieved an amazing synchronization of the movement of the animal's real hand and its robotic counterpart. The implanted electrodes provided two-way communication between the monkey's brain cells and sensors mounted on the robot. As soon as the monkey squeezed the toy in his hand, the robot's hand exactly repeated her gesture. Thus, scientists were able to achieve the result of using the brain activity of a monkey to control a robot. To demonstrate this effectively, Nicolelis conducted an experiment between Japan and the USA: a monkey in the USA thinks that it bends its arm, and a mechanical arm in Japan performs these actions.

In the human brain, an electrode that performs the function of an interface was implanted in 2005. This was done by a group of scientists led by John Donoghue, a well-known physiologist from Brown University (USA) and the founder of Cyberkinetics Neurotechnology Systems. A patient paralyzed after a stroke had microelectrodes inserted into his brain, with the help of which a computer measured the electrical impulses of his brain and converted them into commands to control the cursor. The patient imagined that he was moving his right or left hand, and the cursor on the monitor screen moved in one direction or another. The functionality of the device led scientists to try to make mechanical prostheses controlled by the brain by means of electrodes implanted in it.

Take into account each neuronIf we talk about the use of interfaces for decoding thoughts, then in this case, according to Konstantin Anokhin, it is advisable to divide brain-machine interfaces into two types: "superficial" and "deep".

The first register the total signals from the surface of the skull. The second are signals from single neurons of the brain, which requires the penetration of electrodes (devices with chips) into the depths of the brain.

"For the effective use of surface interfaces, it is practically not necessary to understand the neurophysiological nature of thinking," Konstantin Anokhin notes. – Today's interfaces are a vivid illustration of this thesis. We still don't understand the nature of thinking, but we have working interfaces."

Thought processes and their coding in the nervous activity of the brain are carried out with great specificity: even nearby brain cells can have a completely different connection with the cognitive aspects of behavior. American neurophysiologists and neurosurgeons led by William Frey have proved this phenomenon experimentally. For therapeutic purposes, microelectrodes were implanted into the brain of epileptic patients and the work of individual nerve cells was recorded. When they were shown hundreds of different photos, it turned out that the cells in the anterior region of the hypothalamus are very specialized. For example, one patient had activation of a certain neuron at the moment of recognition of the image of a famous actress. At the same time, the neuron next to him in the same patient was activated only by the image of Mother Teresa.

Surface interfaces record the total activity of large populations of neurons. And one cubic millimeter of the cerebral cortex contains about 45 thousand nerve cells. In this regard, with the help of surface interfaces, no matter how refined they may be, it is impossible to achieve detailed reading, decoding of thoughts, Konstantin Anokhin believes. The future belongs to deep brain–machine interfaces.

What are the rats thinking about?The use of deep brain-machine interfaces opens the way to understanding the mental activity of animals.

Whether the neurons of the animal brain are capable of thinking is a question that has caused lively discussions for many centuries. And a few years ago, scientists were able to get closer to the solution.

American neuroscientists David J. Foster and Matthew A. Wilson from MIT have achieved interesting results. They conducted an experiment: they let a rat run along a two-meter corridor, at the ends of which there were feeders. While the rat was running from one feeder to another, scientists recorded the work of a huge number of neurons. During the calm behavior of the rat (after the next meal), the scientists recorded the same neurons that were activated during the approach to the feeder. It turned out that they are activated together, in a sequence corresponding to their activation during jogging. That is, the rats seemed to be losing the situation in their thoughts – running from the feeder to the feeder. Moreover, the most striking and non-trivial thing was that the animal did it in reverse order (first the neuron that was activated during the meal is activated, then the one that was activated during the approach to the feeder, etc. until the very beginning).

"These data indicate that not only a person, but also an animal is capable of mental activation without performing the process itself," comments Konstantin Anokhin. – This is one of the proofs (there are others) the fact that the animal is mentally thinking through what needs to be done to get food, and it is better to remember it for repeating the same behavior pattern in the future."

If scientists had registered one neuron, they would never have found a certain relationship with the others. This means that the decoding of thoughts will be associated with the registration of a large number of neurons – up to hundreds of thousands, and, accordingly, with the development and creation of appropriate devices.

Mind reading is like flying into SpaceIf we use the registration of individual neurons specialized in relation to certain thoughts, desires, intentions, then it is possible to effectively decode the content of these thoughts due to the phenomenon of cognitive specialization of cells.

So far, this is not a literal mind reading, implemented in the form of practical technology. Nevertheless, the fundamental possibility of such a phenomenon has been proven.

"I would compare the demonstration of mind–reading experiments with the first phase of a space project," says Konstantin Anokhin. – The problem of space travel has not yet been solved: it will take decades before people will move freely from one planet to another. But the human spacewalk shows that this is possible. Similarly with the decoding of thoughts. Experiments show that there is no theoretical prohibition for performing these tasks. Everything else depends on time and technology development."

The most important solutions in this area are the development of technologies that will combine the tasks of building interfaces and studying the deep processes of activity of individual brain cells. This is what will make it possible to read thoughts as an objective process.

In Russia, work on the creation of brain-machine interfaces began in 2003-2004. This direction is actively developing. But mostly, surface interfaces are being developed. According to Mr. Anokhin, not a single Russian laboratory is working with deep ones yet. This direction is planned to begin developing at the Research Institute of Normal Physiology of the Russian Academy of Medical Sciences.

"A really interesting science is one that does not cure any diseases and does not bring any money," says Konstantin Anokhin. – The thought of a scientist always strives upward, beyond the horizons, into those areas into which paths have not yet been trodden and roads have not been laid, which are far from practical implementation. But the work in the field of brain-machine interfaces is just interesting because the distance between technological tasks and deep fundamental research is practically zero. Technology is an incentive for learning how the brain works."