What is a brain-computer interface?

Stand outside a football stadium during a big match. You cannot see the pitch at all. Even so, you learn a lot. A roar means a goal, and a long groan means a near miss. Silence means very little is happening.

Now buy a ticket and take a seat inside. Suddenly you can hear single voices, and you can see exactly who kicked the ball.

That is the gap between the two kinds of brain-computer interface. One listens from outside the skull, and the other listens from inside it.

What a brain-computer interface is

A brain-computer interface is a device that reads activity in the brain and turns it into a command for a machine. Most people shorten the name to BCI.

Brain cells talk with small electrical pulses. When you plan to move your hand, one patch of your brain fires in a pattern. A BCI picks up that pattern and passes it to software, and the software then moves a cursor, a wheelchair or a robotic arm.

Some devices work the other way round, because they send small currents into the nervous system to create a sensation. A cochlear implant does this for hearing.

How it works, step by step

1. A sensor records electrical activity from the brain.
2. A processor cleans the signal and strips out noise from muscles and mains power.
3. Software hunts for the pattern that goes with one intended action.
4. The system turns that pattern into a command.
5. The device moves, types or speaks, and the user sees the result.
6. The user adjusts, and the software learns from the correction.

People forget the last two steps. A BCI is trained together with its user, and both sides improve over days and weeks. Therefore, a short demonstration video hides months of patient practice.

The three ways to reach the signal

Where you put the sensor decides everything else.

• Outside the head. A cap of electrodes reads the surface of the scalp. It is cheap and safe. However, the signal is blurred, like the roar heard outside the stadium.
• Under the skull, on the surface of the brain. A thin sheet of electrodes rests on the tissue. The signal is far clearer, but surgery is needed to place it.
• Inside the brain. Tiny needles sit among the cells and hear single ones. This gives the sharpest signal and carries the most risk.

A newer approach avoids opening the skull at all. Doctors thread a mesh of electrodes up through a blood vessel and park it against the vessel wall, next to the target area. The signal is weaker than a direct implant. However, the operation is far less serious.

What it is good at

The clearest use is giving movement and speech back to people who have lost them.

A person paralysed by injury or by motor neurone disease can still form the intention to move. The brain signal is there, but the path down to the muscles is broken. Therefore, a BCI carries the message around the break.

With an implant, users have moved a cursor, controlled a robotic arm, and produced sentences from the part of the brain that plans speech. Research systems now turn intended speech into words much faster than older letter-by-letter methods.

Simple caps do a narrow job well too. They can spot when a driver is drowsy or when attention drops, which is useful in safety work.

What it is bad at

Implants meet the body’s defences, because scar tissue builds up around the electrodes. The signal can fade over months or years, and some devices must be taken out and replaced.

Caps outside the head have the opposite trouble. They are safe and they are noisy, because hair, sweat and a small shift of the cap all change the reading. A blink can look like a command.

Both kinds need calibration, and many need it again at every session. Prices are high, surgery is scarce, and most advanced use still happens inside clinical trials.

Can it read your mind?

No, and this is the most common mistake people make.

A BCI reads the patterns it was trained on, in the small region it can hear. Train it on the plan to move a right hand, and it will spot that plan. It cannot then tell you what somebody thinks of their neighbour.

Think of one microphone at a market gate, and it hears the traders nearest to it. It does not hear the whole city, and it understands nothing until somebody teaches it a few words.

The honest summary is narrow and still useful, and a BCI decodes a small set of trained intentions well. IT DOES NOT READ THOUGHTS.

Even so, the data deserves care. Brain data is personal, and a few governments have started writing rules that treat it as its own protected class. Gene editing raises the same kind of question, and our explainer on CRISPR works through that side of it.

Where you already meet it

Some parts of this field are already ordinary medicine.

Cochlear implants have restored a form of hearing to many thousands of people, including young children. Deep brain stimulators calm the tremor of Parkinson’s disease with steady pulses. Both are neural devices, and neither one reads intention, so strict definitions do not call them true BCIs.

On the consumer side, headbands claim to track focus, calm or sleep. Treat those claims carefully, because a cheap sensor outside the skull cannot do what a research implant does.

What is coming next

Watch three things, and ignore the rest.

The first is how long an implant keeps working. A device that lasts ten years is a medical product, but a device that lasts one year is an experiment.

The second is whether the system works at home without a technician present. Daily life is the real test, and laboratories are gentle places.

The third is the rules, and consent, ownership of brain data and the right to switch a device off all need law behind them.

So when you next read a headline about mind control, check three things. Ask how many people used the device, for how long, and whether they used it outside a laboratory. Those answers separate a product from a promise. The same three questions work well on gene editing news.

Just Out Tech explains new research in plain language. This article was drafted with AI assistance and checked by a human against the original source.