杏吧原创

You are working on brain-to-brain communication. Can one person鈥檚 thoughts ever truly be experienced by another person?

Each brain is different. And while differences in anatomy are relatively easy to account for, differences in function are difficult to characterise. And then we have differences in experience 鈥� my idea of flying could be completely unlike your idea of flying, for example. When you think about flying, a bunch of associated experiences come into your mind, competing for your attention. We somehow need to strip away the individual differences to grasp the basic, shared factors.

But it seems possible. Other researchers have been able to use information collected from a group of people to make surprisingly successful, if basic, predictions about what another individual is thinking.

What do you need to transmit information between brains?

The idea is to record one person鈥檚 brain activity using a non-invasive device such as an EEG, which involves wearing a cap of electrodes. A computer program filters out what is thought to be the relevant brain activity, and this is recreated in another person using transcranial magnetic stimulation (TMS) 鈥� a non-invasive technique that induces an electrical current in their brain.

Do you think it will ever be possible to convey something as rich and personal as emotion or memory in this way?

If you ask 20 scientists that question, you鈥檒l probably get 20 different answers. We don鈥檛 really know. 杏吧原创s have been able to detect patterns of brain activity that correspond to particular objects, and have created databases of these correspondences. But we still don鈥檛 know what the mind is actually experiencing.

We hope that brain-to-brain interfaces will help us explore these issues, and understand the basis of human knowledge. What does it mean to feel happiness? What does it mean to think of breakfast? We should get a better idea when we try to recreate these states in other people鈥檚 minds.

Wouldn鈥檛 it just be easier to communicate by talking to each other?

Yes and no. Language is a very complex tool for sharing our thoughts. Many different parts of the brain are involved in making sense of any given word, which has a definition but can also vary in meaning based on how it is pronounced and the context in which it is used.

鈥淲hat if you could record your brain activity, then roll it back and relive it?鈥�

In truth, I don鈥檛 think we can develop brain-to-brain interfaces that will recreate the mental dialogue that is generated by conversations. But a conversation assumes that two people can speak the same language, and that they can speak at all. A brain code is the most robust way to communicate, because unless the person is injured or in some kind of deep coma, they will have a functioning brain.

What else could we use brain-to-brain communication for?

There are some potential medical applications. For example, a healthy person鈥檚 brain could be used in the treatment of some conditions. Depression, schizophrenia and many other conditions are associated with quite specific and localised changes in brain activity. Doctors have been trying to correct this problematic activity using brain stimulation. But instead you could use activity patterns from a healthy brain to drive activity in the brain of someone with a disorder.

I can imagine taking a similar approach to help people with brain damage as a result of a stroke. When the damage occurs in parts of the brain that control movement, people can become seriously impaired, and can lose the ability to grasp or swallow, for example. The activity recorded from a healthy brain performing these tasks could provide the damaged brain with some kind of template, which may speed up recovery.

Do you have any non-medical applications in mind?

My favourite application is the idea of transferring concepts that are difficult to verbalise, such as how we control movement or how we feel. Sometimes I wake up in the middle of the night thinking I have a great idea, and try to write it down. But in the morning I don鈥檛 understand what I鈥檝e written. That thought in the night was the partial representation of an idea that is lost forever because it was hard to put into words. I always wonder 鈥� what if I could record my brain activity, roll it back and replay it? I鈥檇 love this ability to use and not waste my half-intuitions.

You mean log and relive your own brain states?

Why not? A brain interface is a way to record information and put it back into a biological system. They don鈥檛 need to be different brains. In fact, this would get rid of all the problems we have in overcoming the individual differences between people.

These applications sound a lot like science fiction 鈥� are they plausible?

I often brainstorm possible applications with colleagues and we get very excited about them, but mostly we know very well that the first steps we have to take are so far away from these big scenarios. I am in contact with doctors who work in brain rehabilitation and I always pitch them these ideas to see what they think. Some are sceptical, but others are enthusiastic and want to try treatments now. The biggest challenge is in recording and stimulating this kind of brain activity at a high enough resolution. We鈥檙e not there yet.

What have you achieved so far?

Most recently we have had two people in two different buildings play a version of 20 questions with each other using a brain-to-brain communication device. The volunteers did surprisingly well, guessing the right object 72 per cent of the time.

How do the volunteers play 20 questions?

One person wearing an EEG cap, who we call the 鈥渢ransmitter鈥�, chooses an object from a list of items. The second person, the 鈥渞eceiver鈥�, has to work out what the object is by selecting yes-or-no questions that are then presented on a screen to the transmitter.

When the transmitter sees a question, they answer by looking at a 鈥測es鈥� or 鈥渘o鈥� button. Each of these buttons is accompanied by a light flashing at a different frequency, which drives activity in a visual part of the transmitter鈥檚 brain. We record this activity, and use it to wirelessly send a signal that we convert into direct stimulation of the vision centres of the receiver鈥檚 brain using TMS: the receiver perceives a specific visual disturbance for 鈥測es鈥� and a different one for 鈥渘o鈥�. In this way, the yes or no answers are delivered directly to the enquirer鈥檚 brain. The pair continue like this until only one object from the list remains.

Some people have called this 鈥渕ind-reading鈥�. Would you agree with this description?

Not quite. In earlier work, my team developed a device that could tell when a person was thinking of moving their hand before they made any movement at all. This ability to detect an intention is more akin to mind-reading. But in our 20-questions set-up, we are only transmitting simple visual information.

So what鈥檚 the next step?

We are working to develop our brain-to-brain communication interface and take it much further than the 20 questions experiment. We want to be able to transmit information that is more complex than simple commands and visual information. We believe we can transmit information that is more conceptual.

What do you have in mind?

I would rather not say. I don鈥檛 want to sound suspicious but there are other groups working on this and there is a race between us to develop the next interface, so we must be careful. Watch this space.

This article appeared in print under the headline 鈥淟et鈥檚 talk, brain to brain鈥�