Brain-computer interface: Pioneers who put wires into their heads always win Nobel Prizes

Brain-computer interface: Pioneers who put wires into their heads always win Nobel Prizes

1. Black technology or business gimmick?

The layman looks at the excitement, the insider looks at the doorway, and the brain-computer interface is not immune. Many friends in other industries think that brain-computer interfaces are definitely black technology, mysteriously like sci-fi movies, while industry insiders often complain that news is too commercial, and there are more gimmicks than scientific progress. When I encounter this kind of discussion, I often laugh without saying a word, and then use stories to let the audience enter the scene and generate answers by myself.

The first story is a pioneer in measuring human brain electrical activity, German doctor Hans Berg (Hans Berger) .

That was in 1924, and the electrical measurement of the human brain was still a “civilian science” that was not recognized by mainstream science. Berg had to carry someone on his back and hide in the basement to study secretly. The biggest problem at that time was that electronic equipment had not yet appeared, and there was no way to amplify the weak signal that was only one ten thousandth of the cell phone battery.

He tried several methods to display EEG signals, such as using the most sensitive filament suspended galvanometer at the time (filar suspended galvanometer, which uses silk to suspend a small coil in a magnetic field and deflects the coil by the tiny magnetic force generated by the current). A small mirror is used to move the light spot, invented by the famous Sir Kelvin) , so that the movement of the light spot can be recorded with film film. This technology was indeed the black technology in the dark room at the time. It’s a pity that the electrical signal from the computer measured from outside the scalp is too weak. He tried many ways and failed.

Fortunately, Berg later encountered a World War I wounded soldier with half of his skull missing. Today we know that measuring EEG through the skull is like taking a photo through a frosted glass window, and the signal is greatly compromised. The absence of the skull strengthened the EEG signals a lot, which strengthened his confidence in the existence of human EEG. After that, he continuously improved the technique and finally measured the brain electricity outside the scalp of a normal person (Figure 1) .

In the history of neuroscience, the pioneers who put wires into their heads all wanted to win the Nobel Prize. It is a pity that Hans Berg failed to survive the Second World War and was nominated for the Nobel Prize for only a few years before his death.

Brain-computer interface: Pioneers who put wires into their heads always win Nobel Prizes

Figure 1: The EEG measured by Hans Berg is the mark left by the light spot on the film. The upper part is the alpha wave when the eyes are closed, and the lower part is the contrast signal. This is the first electrical signal measured in the human brain.

Today, we can buy a chip with a few cents to magnify the brain electricity by a thousand times, and easily demonstrate Hans Berg’s experiment in the classroom of elementary school students. This kind of demonstration often has a sensational effect. For example, the EEG signals are very different when you open your eyes and close your eyes, which makes the experiencer very surprised. A few years ago, my students brought a simple EEG device I made to the popular science event held on Capitol Hill in the United States, which caused many people to line up to experience it (Figure 2) .

Brain-computer interface: Pioneers who put wires into their heads always win Nobel Prizes

Figure 2: Science EEG recording technology. In the popular science neuroscience activity organized by the Neuroscience Association, my students demonstrated a simple EEG measurement device I made, which can eliminate the interference of the environment on the weak EEG signal, and let the experiencer see it on the iPad in real time. My brain waves change with my thoughts.

So is the science of 1924 still black technology today? The answer is yes. Scientific discoveries can always become advanced technologies in due course. In the face of the vast ocean of literature, I will only cite two examples with news effects. One is that in 2014, the pioneer of brain-computer interface Michael Nicholas successfully used brain-computer signals to command outside machines in order to promote brain-computer interfaces. Bones, let a paralyzed person kick off the FIFA World Cup.

Brain-computer interface: Pioneers who put wires into their heads always win Nobel Prizes

Use the brain-computer interface to let the paraplegics stand up and kick-off for the World Cup. Photo courtesy of the Nicholas team Associação Alberto Santos Dumont para Apoio à Pesquisa (AASDAP).

The second example is that in 2015, the brain-computer interface team of Tsinghua University used brain electricity to drive keyboard typing, reaching a speed of 60 letters per minute. After that, it broke the world record at the 2019 World Robotics Competition, reaching a speed of 60 letters per minute. Type 145 characters (Figure 3) .

Being able to type with EEG allows people who are completely disabled to communicate with others normally. This is indeed the gospel that technology brings to people with disabilities.

The brain-computer interface using EEG signals is a non-invasive technology without surgery, so the threshold is very low, and healthy people can also use it. This charming charm has attracted countless people outside the industry and entrepreneurs in the IT industry.

Brain-computer interface: Pioneers who put wires into their heads always win Nobel Prizes

Figure 3: EEG-controlled typing. The picture shows the scene of Mao Fangqi, a first-year high school girl in Beijing No. 11 School participating in the 2019 World Brain-Controlled Typing Technology Competition. In this competition she set a world record.

2. Which is the best intrusion or non-intrusion technology?

Since the non-invasive brain-computer interface for recording EEG from outside the scalp is so strong, why are people in the industry studying an invasive brain-computer interface technology that requires surgery to open the brains? This is because non-invasive EEG technology has its insurmountable limits. This is what I mentioned earlier. Measuring EEG through the skull scalp is like taking a picture through frosted glass. No matter how expensive the camera is, it can’t exert its superiority. The EEG measured outside the scalp can only carry very limited information.

I often use some metaphors to explain neural signal measurement technology: “The brain is like a city, nerve cells are like citizens, and the signal processing in the brain is like chatting between citizens. The brain-computer interface is like a reporter using a microphone to interview citizens and understand the city. News here. Measuring EEG is like conducting an interview through a thick wall. You can only hear the voices of people, but you can’t hear what everyone is saying .”

Nerve cells are very small. Measuring the activity of nerve cells between the skull and scalp a few centimeters thick is like using a helicopter to interview people on the ground from the sky. So what can you hear? You can only hear the crowd yelling together. Just like in the sky above the stadium, you can only hear the crazy shouts of the fans when they win, and judge the game based on this.

Therefore, the measurement of EEG can only detect the signal of the synchronized activity of a large number of nerve cells, but cannot measure the specific activity of a single nerve cell. Therefore, if you want to understand the specific activities of a single nerve cell, you need to go deep into the crowd and bring the microphone to the mouth of every citizen. This is the basic principle of the invasive brain-computer interface. Open the brain case and place tiny electrodes between nerve cells to measure.

Friends who have done news interviews know that to interview a piece of news, you need to listen to many people’s speeches to be accurate. The same is true for brain-computer interfaces. It takes hundreds of thousands of electrodes to measure together to derive a relatively large amount of information from the brain. Not long ago , Elon Musk invented the “sewing machine” technology of brain-computer interface. With every stitch, dozens of tiny electrodes are implanted into the cerebral cortex, so that more than 3,000 electrodes can be implanted in tens of minutes. . Listening to more than 3,000 “microphones” together makes it possible to decode the thought activities in the brain more accurately. In the jargon of the industry, intrusive technology has a large amount of information transmission. This kind of high information transmission speed can never be achieved by non-invasive EEG technology.

In May 2021, “Nature” magazine published an article on invasive brain-computer interface, which immediately set off news. It is about implanting more than one hundred microelectrodes in the brain of a patient with high paraplegia, enabling him to write 90 characters per minute, which is about the same speed as an ordinary person typing by hand.

At first glance, the 90-character-per-minute writing speed is not as fast as Tsinghua’s 145-character world record for typing on an EEG-driven keyboard. Why is there a news sensation? Because here are the secrets of the industry. One argument is that the technology used by Tsinghua University requires full dedication, and intrusive technology liberates the eyes.

Specifically, the technology used by Tsinghua is to stare at one of the 40 squares on the screen with your eyes. There is a character in each square. Each square flashes at a different frequency. When you stare at that Characters, your EEG will have the same frequency as the flash of that grid, so by measuring the frequency of your EEG, you will know which character in the grid you want to type. Although this method is simple and effective, the eyes are firmly tied, and when the gaze drifts, typing errors will occur immediately.

The invasive microelectrode in the brain measures the neural commands of the motor area of ​​the cerebral cortex. When writing, the multiple muscles of the wrist, fingers, arm, etc. coordinate movements accurately. It is necessary to have a large group of nerve cells in the motor area of ​​the cerebral cortex to coordinate activities with high accuracy. The commands of the motor cortex are very obvious and certain (imagine everyone The fonts of the signatures at different times are highly consistent) . 

This kind of obvious and definite movement instruction can be learned by the computer after repeated many times to analyze what kind of neural activity corresponds to what word to write. You don’t need to use your eyes to write in an imaginary way, and you don’t need to use your thoughts to direct specific gestures (muscle memory) . It liberates the eyes and mind, and it is completely understandable to make a sensation. The more important meaning of this work is that it can follow the same routine to analyze other brain movement instructions, such as sitting up, walking, dancing, and making the paralyzed really move.

3. The rise of microelectrode technology

So, is the black technology invented by Musk to insert microelectrodes into the cerebral cortex? The simple answer is: No, but he did make a great contribution . To evaluate Musk’s contribution, we need to understand the principle of microelectrodes and the relationship between scientific original inventions and capital follow-up.

The electrode that measures the signal of a single nerve cell needs to be small in size, so it is called a microelectrode. The size of nerve cells is only one-tenth of that of sesame seeds, so the size of the electrode should be small. It’s a bit like interviewing a person in a busy city. The microphone should be small and close. If the microphone around a person is as big as a bus, then the human voice will be drowned in the surrounding noise.

Electrodes are like wires, they need a conductive core and an insulated shell. Only in this way can the nerve signals detected by the tip be transmitted to the external brain-computer interface. How to make a wire as thin as one-fiftieth of a human hair and add an insulating shell?

Many people in the industry believe that the key figure in the invention of microelectrode technology is a young Chinese student, Yang Zhenning’s friend , Gilbert Ning Ling (1919-2019) (Figure 4, right) . His technique is to melt a thin glass tube on a flame and then pull it hard. At this time, the melted glass is like the syrup on toasted yam. It is first drawn into a filament, and then the filament is broken. Carefully adjust the flame temperature and tensile force, you can keep the fracture of the glass filament in a regular nozzle (Figure 4, left) . In this way, the thin glass tube wall is a good insulating shell, and the salt solution filled in the tube can conduct electricity and become a miniature wire with an opening less than one micron in diameter, which derives the nerve signal measured by the tip. 

This invention by Ling Ning and his mentor helped several Nobel Prize winners and laid several milestones in neuroscience. Of course, they were also nominated for the Nobel Prize. It is a pity that Ling Ning later became lonely because of the opposition to mainstream science. Later, he worked as an ordinary professor in an American university until he retired.

Brain-computer interface: Pioneers who put wires into their heads always win Nobel Prizes

Figure 4: Electrode and his inventor Ling Ning. On the left, a photomicrograph of the port of the glass microelectrode. The port where the fuse is broken is still in a neat tube shape.

However, Lingning’s glass microelectrode technology is not his first creation, but the technology passed down for many years. Hundreds of years ago (1660s) , scientist Robert Boyle explained the method of making exquisite knives by drawing glass. At that time, playing with glass filaments thinner than hair was indeed a black technology. Even Robert Hooke , the originator of cell biology, used and developed this type of glass technology. In the following hundreds of years, scientists have been using glass filaments to make needles, hooks, knives, tubes and other tools at the tiny scale. By 1920, the tiny glass nozzles were small enough to grab individual bacteria from the water.

It is worth mentioning that the skillful scientist Albert Barber used a small gas flame to draw glass microtools, and his craftsmanship has developed to the extreme (Figure 5) . His tools have helped several Nobel Prize original work. Lingning’s microelectrode technology also benefits from the handicraft developed by Gbagbo. Nowadays, Gbagbo’s skillful hands have been completely replaced by machines, and the precise temperature adjustment and tension control technology are also completely stored in the computer, so that a new student can pull out hundreds of identical microelectrodes more accurately than Gbagbo. .

Brain-computer interface: Pioneers who put wires into their heads always win Nobel Prizes

Figure 5: The skillful scientist Gbagbo and his invented glass micro-tool manufacturing method uses a small gas flame to melt the middle section of a thin glass tube, and then draw it horizontally by hand.

Microelectrode technology must be combined with electronic amplifiers to be useful. Historians discovered that glass microelectrodes with a diameter as small as 4 microns were invented in the 1920s, but there was no signal amplifier at that time, and the microelectrodes could not exert their power. They were gradually forgotten by the scientific community until electronic amplifiers became popular after the Second World War. just let Ling Ning electrode about popular neuroscience field, and even spawned a called ” electrophysiology ” (electophysiology) of new disciplines to clarify the function and drug targets on the surface of nerve cells in a variety of molecular machines, until today still It is an important branch of neuroscience.

Fourth, tungsten microelectrodes enter the brain

Using glass microelectrodes to study neural activity in the cerebral cortex encountered a technical obstacle: the glass was too brittle to pass through the dura mater that protects the cerebral cortex. This technical obstacle led to another invention, which is the tungsten microelectrode . Tungsten alloy is used to build tanks, and it is harder than steel, so it can pass through the dura mater. But how to make such a hard metal into micrometer-scale microelectrodes?

Smart inventions are generally simple, and the secrets are like a window paper that breaks with a poke. The method is to put the tungsten wire in salt water, and then turn on the electricity. Electric current can deprive the metal lattice of electrons. Once the metal lattice is destroyed, the hard tungsten will dissolve in salt water like sugar lumps. The tip of a metal wire is in three-dimensional contact with the solution, while the other surfaces leaving the tip are only in two-dimensional contact with the solution. Therefore, the tip is dissolved by the current faster than other parts. A tungsten wire naturally forms a tungsten wire when it is energized. The tip, the electrode is generally useful when the tip reaches one fifty-fifth of the diameter of a hair. The finished fine needles should also be put on an insulating coat, and about one micron of metal is left in contact with the brain tissue at the tip to listen to the conversations of nerve cells.

The secret of technology is simple, but it still requires the hands-on people among scientists to fumble day and night like fascinated. David Leisure Expo (David Hubel) is one such scientist. He learned the skills of several seniors in making tungsten microelectrodes and optimized the technology to a practical level. Although playing with these does not seem like a scientist but more like a craftsman, it is enough for him to publish an article on the top academic journal “Science” on his own, devoted to tungsten electrodes. For example, how big the tip is, how much the angle is, whether it is pointed or blunt, how to paint can expose the tip, etc. (Figure 6) .

If a worker wants to do his job well, he must first sharpen his tools. The tip of Huber’s tungsten microelectrode exactly matches the size of nerve cells in the cerebral cortex (Figure 7) , which can perfectly record the firing signals of nerve cells. He and his colleagues used this technology to penetrate the brain and studied the response of the animal’s cerebral cortex to images seen by the eyes, leaving behind an epoch-making classic work. The eyes are the windows of the soul, and their work in the visual cortex has inspired subsequent generations of research on the cerebral cortex. A few years later, Hugh Bo won the Nobel Prize together with his colleagues (Figure 8) .

Brain-computer interface: Pioneers who put wires into their heads always win Nobel Prizes

Figure 6: Tungsten microelectrode of Huebo. This is the original picture in his article “Science” (Science, 125, No. 3247 (Mar. 22, 1957), pp. 549-550)

Brain-computer interface: Pioneers who put wires into their heads always win Nobel Prizes

Figure 7: Artistic imaginary diagram of tungsten microelectrodes and neurons in the cerebral cortex, from the Harvard University Brain Science Popularization website.

Brain-computer interface: Pioneers who put wires into their heads always win Nobel Prizes

Figure 8: Hugh Bo and his laboratory. On the left is Hugh Bo, on the right is the colleague Torsten Wiesel who shared the Nobel Prize with him. The instruments behind them represent the level of brain science instruments in the 1950s: a: micro-manipulator, which uses hydraulic transmission to slowly push the microelectrodes one micron by one micron to get close to the nerve cells being listened to; b: audio monitoring The device is used to hear the response of nerve cells to visual signals. The human ear is very sensitive to sound rhythms and can detect signal rhythm changes that are invisible to the eyes; c: oscilloscope, used to see nerve signals, to estimate the distance between the electrode tip and the listened nerve cell; d: tape recorder, used to observe Recorded the phenomenon (the microcomputer had not yet entered the laboratory); e: Multi-channel recorder (commonly known as a polygraph), used to monitor the physiological condition of the animal after anesthesia.

Five, microelectrodes from one to many

I often use an analogy. Using a microelectrode to study the working principle of the cerebral cortex is equivalent to guessing the plot by observing a pixel on the TV screen. Of course, this is an impossible task.

When the microelectrode can successfully record the activity of nerve cells in the brain, the next step is to record as many nerve cell activities as possible at the same time . When the cerebral cortex is active, billions of nerve cells are often involved at the same time, so there are hundreds, thousands, or even hundreds of thousands of microelectrodes in the brain-computer interface. The current brain-computer interface technology has used hundreds to thousands of electrodes. So, how can so many microelectrodes be manufactured and applied at the same time?

Neuroscientists thought of the booming semiconductor integrated circuit technology. The so-called integrated circuit is to draw many circuits on a silicon chip to connect a large number of transistor devices. Using the same technology, many electrode surfaces and electrode leads that contact nerve cells can also be made.

The formation of electrode forests on silicon wafers is a bit like rainwater corroding limestone ground, forming stone forest landforms (Figure 9 left) . The silicon wafer is etched in strong acid, and the areas where electrodes need to be formed are shielded, and then an electrode array with ridge-side peaks can be formed as needed (Figure 9 right) . This kind of microelectrode array made of silicon chips is also called ” Utah electrode array “, which is currently approved for use on patients (Figure 10) . The brain-computer interface mentioned in the previous article that allows high-level paraplegics to write 90 characters per minute is the electrode array used.

Brain-computer interface: Pioneers who put wires into their heads always win Nobel Prizes

Figure 9: Shilin and Utah electrode arrays. Rain water corrodes the ground and can form stone forest landforms (left). An electrode forest (right) that can be formed on a silicon wafer with similar etching techniques.

Brain-computer interface: Pioneers who put wires into their heads always win Nobel Prizes

Figure 10: Photo of the actual Utah electrode array placed on the cerebral cortex during the operation. In this example, the two electrode arrays are installed in the body sensory area on the cerebral cortex, so that proper electrical signals can be applied to the electrode array to generate virtual touch. .

The early Utah electrode array had only 100 electrodes (10 by 10) . The subsequent improvement was to increase the number of electrodes on the silicon wafer by another 10 times. The reason is also very simple. People have already mastered the skill of drawing wires on silicon wafers. If you draw 10 wires and 10 electrode surfaces on each electrode rod (Figure 11) , you can create 1,000 (10x10x10) small microphones that listen to nerve signals on a Utah array with 100 electrode rods (10×10) . .

Brain-computer interface: Pioneers who put wires into their heads always win Nobel Prizes

Figure 11: The technique of manufacturing multiple electrode contacts on a single electrode rod.

6. From hard to soft, a big step forward

The brain tissue is as soft as tofu, and it is often deformed or moved during daily activities. Hard microelectrodes, whether glass, tungsten or silicon, cannot move with the brain tissue. The mutual movement between this electrode and the brain tissue can cause damage, like chopsticks mixing mung bean porridge. The micro-damages surrounding the electrodes caused by mutual movement can cause local inflammation and the proliferation of glial cells similar to scars.

Glial cell proliferation will block between the electrodes and nerve cells, causing the signal to gradually weaken, just like putting a blanket between the microphone and the speaker. The signal weakening problem is the main reason that restricts the wide application of Utah electrode arrays -installing the electrode array is a potentially dangerous craniotomy operation, and no one wants to have the electrodes gradually fail months or years after the electrodes are installed. .

Softening the electrode rod can make the electrode fluctuate with the displacement of the tissue like seaweed, which greatly reduces the micro-damage of the surrounding electrode. But it is very difficult to insert a soft electrode into the brain tissue. Imagine how to insert a soft rope vertically into the quagmire?

At this time, Iron Man Musk appeared and solved the world problem of the soft electrode. His method is a ” sewing machine “-using a hard tungsten needle to bring the soft electrode into the brain tissue, then the needle is pulled out, and the soft electrode is left in the brain tissue (Figure 12) . It is not yet known whether this clever idea comes from Musk himself, although he is also a very clever inventor.

Musk has the money to buy and buy, which is an important factor that gave birth to this technology. We all say that the progress of science and technology comes from a flash of light in the brain of scientists, but this kind of flash of light appears too much, like the fireflies on the summer grassland, one after another from far and near. Without strong capital to follow up, most flashes can only fend for themselves, or be reinvented decades later. The irreplaceable role of Iron Man is to use capital to greatly accelerate the process of invention.

In 2016, Musk invested $100 million to create Neuralink. Two years later, the company said that it had invented a black technology, which is a soft electrode that can peacefully coexist with nerve tissue, and a sewing machine that can insert a rope into the mud.

Seven, the sewing machine on the cerebral cortex

This sewing machine is a robot that automatically implants soft electrodes. The sewing machine needle is a tungsten alloy needle with the diameter of a human hair, and its tip can be easily inserted into the brain tissue (Figure 12A-B) . There are a few cameras watching it, automatically avoiding the blood vessels on the surface of the brain, so that the process of implanting the electrodes becomes very safe, and there is very little bleeding.

The soft electrode is made of plastic film. On the polyimide plastic film, small patterns can be drawn by photolithography technology for manufacturing chips, and then coated with a conductive polymer or metal film to form thin wires and electrode surfaces in contact with brain tissue (Figure 12C) . Each needle of the sewing machine takes a narrow strip of the film into a soft strip and inserts it into the brain tissue. On the surface of each strip are 32 electrodes in contact with nerve cells (Figure 12C) . In this way, 3072 soft electrodes can be implanted in one operation with one needle by needle. This is the current world record.

Brain-computer interface: Pioneers who put wires into their heads always win Nobel Prizes

Figure 12: Musk electrode array. A, the working diagram of the sewing machine. Determine the position of the soft electrode according to the depth of the needle. B, Partial enlargement of sewing machine and needle. C. Soft electrode strips constructed on plastic film, each narrow strip contains 32 electrodes. D, the implantation process of the soft electrode strip, i is the one that is about to be implanted, and ii is the one that has been implanted. E. On the animal brain, automatically avoid multiple soft electrodes implanted in blood vessels.

8. The future: the last micron project

Predicting the future is a big-open fantasy. Fantasy is not random thinking, but also based on the history of neuroscience development in the previous century. Based on my superficial knowledge and limited imagination, the author believes that the key to the brain-computer interface is the “last micron”, which is the contact interface between electrodes and nerve cells. For this reason, I conceived the future directions of several brain-computer interfaces and the technical bottlenecks that need to be broken.

Due to space limitations, this article is limited to historical stories, and the future prospects are left in the next narrative.


[1] The story of Hansberg Haas LF. Hans Berger (1873-1941), Richard Caton (1842-1926), and electroencephalography. J Neurol Neurosurg Psychiatry. 2003 Jan;74(1):9. doi: 10.1136/ jnnp.74.1.9. PMID: 12486257; PMCID: PMC1738204.

[2] Tsinghua’s EEG typing article Chen, X. et al. High-speed spelling with a noninvasive brain–computer interface. Proc. Natl Acad. Sci. USA 112, E6058–E6067 (2015).

[3] The article Willett FR, Avansino DT, Hochberg LR, Henderson JM, Shenoy KV. High-performance brain-to-text communication via handwriting. Nature. 2021 May;593(7858): 249-254. doi: 10.1038/s41586-021-03506-2. Epub 2021 May 12. PMID: 33981047; PMCID: PMC8163299.

[4] The development history of glass micropipette electrode Bretag AH. The glass micropipette electrode: A history of its inventors and users to 1950. J Gen Physiol. 2017 Apr 3;149(4):417-430. doi: 10.1085/jgp. 201611634. Epub 2017 Mar 15. PMID: 28298356; PMCID: PMC5379916.

[5] The Harvard University Brain Science Popularization website introduces Hubel’s work

[6] Hubel DH. Tungsten Microelectrode for Recording from Single Units Science, 125, No. 3247 (Mar. 22, 1957), pp. 549-550

[7] Musk E; Neuralink. An Integrated Brain-Machine Interface Platform With Thousands of Channels. J Med Internet Res. 2019 Oct 31;21(10):e16194. doi: 10.2196/16194. [7] Musk E; Neuralink. An Integrated Brain-Machine Interface Platform With Thousands of Channels. PMID: 31642810; PMCID: PMC6914248.

This article comes from WeChat public account : Mr. Sai (ID: mrscience100) , author: Wu Jianyong

Posted by:CoinYuppie,Reprinted with attribution to:
Coinyuppie is an open information publishing platform, all information provided is not related to the views and positions of coinyuppie, and does not constitute any investment and financial advice. Users are expected to carefully screen and prevent risks.

Like (0)
Donate Buy me a coffee Buy me a coffee
Previous 2021-07-13 12:07
Next 2021-07-14 08:59

Related articles