A CHILD watches her mother pick up a toy. The child smiles: Mum wants to play. A husband watches his wife pluck car keys from a table. He shivers: she really is leaving this time. A nurse watches a needle being jabbed into an elderly patient. She flinches: it must have hurt.

How do these people know what the other person is thinking? How do they judge intentions and feelings, or assign goals or beliefs to the other? It sounds simple, but the child could just as easily have decided that Mum was leaving or the husband that his wife wanted to play. Yet they didn't. They knew.

"Reading" the minds of others is something we take for granted. Yet philosophers, psychologists and neuroscientists alike have been baffled by our ability to anticipate other people's behaviour and empathise with their feelings. Now a team of Italian neurophysiologists may have stumbled on the key to this mystery.

Vittorio Gallese, Giacomo Rizzolatti and their colleagues at the University of Parma have identified an entirely new class of neurons. These neurons are active when their owners perform a certain task, and in this respect are wholly unremarkable. But, more interestingly, the same neurons fire when their owner watches someone else perform that same task. The team has dubbed the novel nerve cells "mirror" neurons, because they seem to be firing in sympathy, reflecting or perhaps simulating the actions of others.

Illustration: Charlie Ward/Photography: Bettina Salomon

Many neuroscientists are starting to think that in higher primates, including humans, these neurons play a pivotal role in understanding the intentions of others. "Mirror neurons may be one important part of the mosaic that explains our social abilities," says Gallese. Vilayanur Ramachandran of the University of California at San Diego goes further. He believes that mirror neurons will answer important questions about human evolution, language and culture--and may take us to the heart of what it means to be human. "I predict that mirror neurons will do for psychology what DNA did for biology," he says. "They will provide a unifying framework and help explain a host of mental abilities that have hitherto remained mysterious."

Gallese and his colleagues didn't set out to find anything so radical when, in the early 1990s, they started recording the activity of neurons in a macaque's brain. They were tapping into the signals emitted from nerve cells in a part of the monkey's brain known as F5. This is part of a larger region called the premotor cortex, whose activity is linked to planning and making movements. Some years earlier, the same researchers had discovered that neurons in F5 fired when an animal performed certain goal-oriented motor tasks using its hands or mouth, such as picking things up, holding or biting them.

They wanted to learn more about F5 neurons--how they responded to different objects with different shapes and sizes, for example. So they presented monkeys with things like raisins, slices of apple, paper clips, cubes and spheres. It wasn't long before they noticed something odd. As the monkey watched the experimenter's hand pick up the object and bring it close, a group of the F5 neurons leaped into action. But when the monkey looked at the same object lying on the tray, nothing happened. When it picked up the object, the same neurons fired again. Clearly their job wasn't just to recognise a particular object.

All fired up

The neurons turned out to be quite fussy about what they reacted to. Those that responded to an experimenter plucking a raisin from a tray, for instance, failed to react when the experimenter dug the same raisin out of a small well with his finger. Some neurons fired when the experimenter held a few slices of apple, but not when he placed the apples on the tray--other neurons fired for that.

Most importantly, the very same action that made a neuron fire when a monkey performed it would almost always make that neuron fire if the monkey saw the experimenter doing the same thing. It soon became clear that the motor system in the brain is not limited to controlling movements. In some way it is also reading the actions of others.

In 1998, Gallese gave a talk about mirror neurons at a meeting on the "Science of Consciousness" in Tucson, Arizona. Alvin Goldman, a philosopher from the University of Arizona, listened with interest. Afterwards, he approached Gallese and they spoke about the potential of these cells for reading the minds of others. "He wasn't familiar with the mind-reading literature in philosophy," says Goldman.

Mind-reading, or theory of mind, is an ability that all healthy humans possess. We are particularly good at representing the specific mental states of others. These can be basic, such as seeing someone crying and understanding that they are sad, or realising that when someone is yelling and gesticulating wildly at you they may be angry and might even mean to harm you. But we intuitively understand more complex mental states too. When a mother loses a baby, other parents get lumps in their throats. When you hear that a colleague has been cheating on their spouse, you share the hurt and shame.

A debate rages over whether other primates, such as chimps, can understand other minds, even in the simplest ways. And even in humans, while almost everybody agrees that some measure of mind-reading goes on, there is little agreement on how it happens. One theory, sometimes called "theory theory", holds that people build up common-sense hypotheses to explain why other people do what they do. Like physicists using rules and laws to explain observable phenomena, we all use our experiences to develop a set of explanatory laws for others' behaviour.

Another dominant theory, championed mainly by philosophers like Goldman, is known as simulation theory. It's based on the idea that people understand what is going through the minds of others by mentally mimicking what the other is thinking, feeling or doing--in essence, putting themselves in the other's shoes. The discovery of mirror neurons backs up this theory nicely.

Monkey see, monkey do: the same cells that fire when an action is performed also spring into life when that action is observed. could this mutual understanding underlie empathy and language?

As the suspicion grew that these neurons might have something to do with the complexities of mind-reading, the burning question became whether human brains had mirror neurons too. But finding out wasn't easy--humans aren't keen on having electrodes implanted into their brains, even for the lofty purposes of science.

Luciano Fadiga, now at the University of Ferrara in Italy, was the first to find some evidence that humans may have a system analogous to that found in monkey brains, when he measured the excitability of particular muscles in the hand. He found that when the volunteers were watching grasping actions, the very muscles that would be needed to copy that movement seemed primed to act--as if they were preparing to make the same movement themselves. "The interesting thing was that the pattern of activated muscles changed according to the observed actions," says Fadiga. But while this suggested that a mirror system might exist in human brains too, it didn't yield any information about where it might reside.

Several brain-imaging studies followed, the first led by Rizzolatti, and another by Scott Grafton, then at the University of Southern California in Los Angeles. Both found that watching an experimenter pick up and handle objects activates two regions of the brain behind the temples on the left side: the superior temporal sulcus and, just above it, a part called Broca's area.

An even more recent study by Marco Iacoboni at the Los Angeles School of Medicine confirmed that Broca's area was active while volunteers either watched images of someone drumming their fingers, and when they also tried to imitate the movement they saw (Science, vol 286, p 2526).

Finding the words

The finding that Broca's area was activated was doubly intriguing. For one thing, F5 in monkeys is considered an analogue for Broca's area in humans. But even more suggestive was the fact that, while F5 is associated mainly with hand movement, Broca's area is traditionally thought of as a speech-production area. This raised questions about what a mirroring system might have to do with language--and language with mind-reading.

Rizzolatti and Arbib think that mirror neurons may have provided the bridge from "doing" to "communicating". The relationship between actor and observer may have developed into one involving the sending and receiving of a message. In all communication the sender and receiver have to have a common understanding about what's passing between them. Could mirror neurons explain how this is achieved? Rizzolatti and Arbib think the answer is yes.

They suggest that it is probably no coincidence that the area which links action recognition and action production in the monkey brain is exactly the same area that in humans has been linked to speech production. They think that the development of human speech was made possible by the fact that F5, the precursor of Broca's area, was endowed with this mirroring mechanism for recognising actions made by others. This, they say, was a prerequisite for the development of communication and ultimately of speech. It made us "language-ready", says Arbib.

Most of the time, a strong spinal cord inhibition prevents you from involving your own motor neurons in activity you are merely observing, according to work by Fadiga soon to be published in the European Journal of Neuroscience. But sometimes the premotor cortex allows a brief snippet of the movement--like the twitchy feeling you get when you're watching someone struggling to open a packet of crisps or untie a knot.

This slight movement, says Arbib, tips off the person carrying out the action that the watcher knows what's going on, in a sort of primitive dialogue. "This dialogue forms the core of language," he says. "Perhaps we evolved some crude form of communication based in sign, then built speech," says Arbib. Imagine an early human chipping away at a stone, he says, and that this person wants to communicate something else while demonstrating this skill. Or perhaps he wants to communicate in the dark or at a distance. In both cases, using sign or gestures doesn't work so well. If the brain could allow the person to develop speech through the same neural apparatus earlier primates were already using to communicate manually or through lipsmacks, so much the better (New Scientist, 8 April 2000, p 30).

Illustration: Charlie Ward/Photography: Bettina Salomon

The exciting news is that mirror neurons may not be limited to these motor regions. Gallese, for one, suspects that they are found in other areas. "My belief is that this may apply also to other modalities, for instance sensory modalities," he says. Gallese points to recent work by William Hutchison, a physiologist at the University of Toronto. He and his colleagues studied humans who were conscious while undergoing brain surgery. They discovered neurons in the anterior cingulate cortex, a region thought to be involved in perceiving pain, which fired both in response to a finger being pricked and also when patients saw the experimenter prick himself (New Scientist, 8 May 1999, p 17).

Gallese sees this as tantalising, preliminary evidence of a far-reaching neural mechanism. Could this explain how we are able to "feel" what others feel? Could it underpin the sensations behind empathy?

Ramachandran also believes that mirror neurons play a bigger role than is generally appreciated. He thinks these exciting nerve cells don't just provide a missing link between gesture and language, but they go a great way towards explaining human learning, ingenuity, and culture in general. "Their emergence and further development in hominids was a decisive step," he says.

He says mirror neurons and the way they facilitate imitative learning help to explain why we only developed things like tool use, art and mathematics about 40,000 years ago, despite the fact that our brains had reached their full size some 150,000 years earlier. These cultural inventions, he contends, probably popped up accidentally, but they were disseminated quickly because of our amazing, imitative, learning brains--made possible by a more sophisticated version of the monkey mirror neuron system.

He admits that mirror neurons probably aren't the whole story--necessary, but perhaps not sufficient--but insists they could be a big part of it. Language, imitative learning and mind-reading, seemingly unrelated human developments, may all be shown to be linked through these intriguing nerve cells. "These are all human qualities. All mysterious qualities," he says. "Mirror neurons may provide the key."