In the future, brain computer interface is moving from science fiction to reality.
In the sci-fi movie “Mechanical Warrior,” the hero Murphy suffers an accident, leaving only the brain and part of the trunk. Scientists build a complete mechanical skeleton for him. Murphy controls the steel body through his brain and becomes a mechanical warrior.
In the movie The Pacific Rim, humans can fight monsters through a cranial nerve connection manipulator armor, setting off a frenzy of machine armor.
Humans have been dreaming that one day the brain can directly control machinery and replace physical defects. If someone tells you that this is about to become a reality, is it crazy? In fact, this day has arrived. This technology is the brain computer interface to be discussed in this paper.
What is brain computer interface?
Brain-computer interface (BCI) constructs direct information transmission pathways between the brain and the outside world by decoding the information of brain nerve activity in the process of human thinking. BCI has a wide range of applications in the fields of neural prosthesis, neural feedback training, and brain state monitoring.
To put it simply, it is to control machines with ideas. This means that the main mode of human-machine interaction, in addition to manual input, and in recent years the rise of artificial intelligence voice interaction, can also directly send instructions to the machine through the brain.
Current brain-computer interface technologies can be divided into two categories, one is intrusive, such as implanting chips into the brain, and the other is non-intrusive, such as wearing a helmet or hat to collect brain waves.
At the opening ceremony of the 2014 World Cup in Brazil, a teenager who was paralyzed below the waist got up from a wheelchair through a BCI-controlled exoskeleton and took a few steps forward to kick the first football of the World Cup.
The brain computer interface tool currently in use
Functional Magnetic Resonance Imaging (FMRI)
Scale: large (showing information of the whole brain).
Resolution: low spatial resolution and extremely low temporal resolution.
Invasive: non-invasive
FMRI is not a traditional brain-computer interface tool, but it’s a typical recording tool — it tells you what’s going on inside the brain.
FMRI can scan the whole brain and generate three-dimensional models.
The main disadvantage of FMRI is its resolution. The image it scans produces does have a resolution, just as there are pixels on a computer screen, except that its pixels are three-dimensional, that is, the so-called voxel.
Next is the time resolution. The blood flow tracked by FMRI is not accurate, and there will be about one second delay.
EEG (Electroencephalography, EEG)
Scale: big
Resolution: extremely low spatial resolution and high temporal resolution.
Invasive: non-invasive
Electroencephalography (EEG) has been used for nearly a century.
A completely non-invasive brain-computer interface tool that records electrical activity in different areas of the brain and produces such results:
Electrocorticogram (Electrocorticography, ECoG)
Scale: big
Resolution: low spatial resolution and high temporal resolution.
Invasive: partial invasiveness
Electrocorticogram works like electroencephalogram, and electrodes are used as well — but this technique puts them under the skull, the surface of the brain.
It’s a bit scary, but it’s very effective, at least better than EEG. Without skull interference, electrocorticogram can obtain higher spatial resolution (about 1 cm) and temporal resolution (5 ms) information.
O local field potential (Local Field Potential, LFP)
Size: small
Resolution: low spatial resolution and high temporal resolution.
Invasive: highly invasive
The microelectrodes involved in the brain are between 10 and 30 microns in diameter. One of the latest developments in the field of local field potential is the multielectrode array, which works just like the local field potential, except that it inserts 100 microelectrodes simultaneously in a certain area of the cortex. The following figure shows 100 tiny silicon electrodes on a small square of 4 mm x 4 mm, with a tip of just a few microns. :
O single cell recording (Single-Unit Recording)
Scale: minimal
Resolution: extremely high
Invasive: highly invasive
Single-cell recording also uses microelectrodes, but the tip of the electrode becomes unusually small, so its resistance increases dramatically. This shields most of the noise, but it also causes the electrode to detect almost nothing — unless it is particularly close to a neuron (perhaps 50 microns away), the neuron sends a signal strong enough to pass through the resistance of the electrode.
Because this method detects the unique signals of a single neuron without interference from background noise, we can now monitor the individual activity of a single neuron. This method has the smallest record size and the highest resolution.
Another electrode, called a patch clamp, can detect neurons more closely. The tip of the electrode is removed, leaving a tiny glass straw (1/100 of the diameter of a human hair) that holds one part of the cell membrane. It can be absorbed into the glass tube to achieve more precise measurements.
The last electrode, called sharp electrode recording, actually pierces the cell membrane and enters the neuron completely. If the tip of the electrode is sharp enough, it doesn’t destroy the neurons — because the cell membrane closes around the electrode.
This way can easily stimulate neurons, or record the voltage difference inside and outside the neurons. But this technique doesn’t interfere with neurons long enough — because the punctured neurons don’t survive long enough.
Brain computer interface on draught (BCI)
With the maturity of BCI technology, many top players compete to enter the BCI field:
Neuralink
As you may know, Silicon Valley Iron Man Musk used Tesla to drive the explosion of electric cars and SpaceX to speed up space exploration, but you may not know that he founded Neuralink, a brain-computer interface company, to transform humans themselves. Since the era of artificial intelligence, Musk has been trying to promote man-machine integration and turn human beings into artificial intelligence to counter the uncontrollable risks of strong artificial intelligence. This may be the ultimate path that human species can continue.
I think the fact that people don’t like is that they’ve become semi-robots. Compared with 20 years ago or even 10 years ago, people are not the same creature, for example, many people have been inseparable from the mobile phone for a moment. I think people have been integrated with mobile phones, notebooks and applications to some extent. Neuralink’s goal is a brain-computer interface that transmits nerve signals to any region of the brain. It will become a comprehensive “digital third cortex,” and humans will become semi-robots.
Facebook
In 2017, Facebook unveiled the mysterious department Building8 at the F8 Developers Conference, and released the black technology Brain-Computer Voice Text Interface, which, in human words, means that what you want in your head is displayed directly on your monitor.
At the conference, Facebook showed an experimental video from Stanford, in which subjects were able to input eight English words per minute with a computer cursor controlled by electrodes implanted into the brain.
Intelligent brain-computer interaction is one of the ten most promising technologies in the field of new generation artificial intelligence (2018–2019) published by China Electronics Society in 2008.
BCI research bottleneck
Brain: The human brain science exploration is still in infancy, how to develop BCI under the cognitive limitation of brain science?
Computer: bandwidth bottleneck — how can children skillfully accommodate large amounts of information?
Interface (Interface): how to solve the problem of biocompatibility? The human immune system excludes the “invading” object outside the body. How can BCI be integrated into the organism?
Brain Computer Interface (BCI): The ethical question — in the experimental phase, if the craniotomy, you can do a lot of biological experiments? In the promotion stage, if the reading mind is realized, where is the privacy?
The future of BCI
So, the question is coming. When will BCI be able to enter ordinary people’s homes? Xiao Bian can not answer this question.
“I think that in the next 8 to 10 years, brain computer interface technology can be applied to ordinary people. It is worth noting that the marketing time of the product is subject to regulatory approval and the effect of the use of patients with various diseases.
— musk said