Dollhouse

A flash of laser light can alter the brains of fruit flies so that they learn to fear pain that they never actually felt.

Gero Miesenböck at the University of Oxford and his colleagues genetically engineered fruit flies so that a handful of their nerve cells fired when lit up with a laser.

This allowed them to write false pain "memories" into the fruit flies’ brains. "These memories cause a lasting modification of the flies’ behaviour," says Miesenböck.

It is known that the release of dopamine by neurons in the "mushroom body" – part of the fruit fly brain – is critical to learning. But it was not known whether behaviour can be conditioned by stimulating these neurons directly, without the fly having any real experience. Lessons in pain

To investigate, Miesenböck and his colleagues started by putting ordinary fruit flies into a small chamber while two different odours were pumped in from either end to create two separate odour streams.

The researchers delivered an electric shock each time a fly strayed into a particular odour stream, which taught the flies to prefer the other one: the flies learned to move in the direction of the shock-related odour 30 per cent less often.

Once he had shown that the flies had learned to avoid pain, Miesenböck decided to see if similar conditioning could be created by stimulating neurons without actually hurting the flies. Bright ideas

His team started by genetically engineering a second set of fruit flies so that their dopamine-producing brain cells manufactured a membrane protein called P2X2. When P2X2 binds to a molecule called ATP, the neuron that produced it fires as if zapped by an electric shock.

The team then made these P2X2 neurons light-sensitive by injecting the flies with a form of ATP that is activated only by a laser. By injecting the light-sensitive ATP into different neurons in different flies, they were able to produce flies with different combinations of light-sensitive neurons.

The researchers then put these genetically modified flies into the smell chamber. This time, when the flies strayed into a particular odour stream, the researchers flashed them with a laser beam instead of zapping them with an electric shock as they had with the normal flies.

Many of the flies did not react. But flies that had 12 particular light-sensitive neurons chose to move in the direction of the laser-related odour 28 per cent less of the time – almost exactly the same result as in the unmodified flies that were exposed to electric shocks.

Miesenböck concludes that stimulating dopamine release in these 12 neurons has the same effect as applying electric shocks to flies. In other words, these flies feared that smell as if they had been conditioned to associate an electric shock with it. "Stimulating just these neurons gives the flies a memory of an unpleasant event that never happened," he says. Like fly, like human?

He says that it is likely that humans form memories in a similar way. "I would be surprised if the way humans learn from mistakes turned out to be fundamentally different from the way flies learn from mistakes."

"The scientists have identified a discrete population of nerve cells that are seemingly the source of ’memory’," adds Richard Baines, a neuroscientist based at the University of Manchester, UK. "This represents a further demonstration of the power of using organisms like the fruit fly for understanding how the human brain works."

However, Wayne Sossin, who studies the biochemical pathways of memory formation at the Montreal Neurological Institute and Hospital – part of McGill University, Canada – points out that it will be difficult to show that human memories work in the same way. "It would be unethical to engineer transgenic humans and tell them what memories to have," he says.

He also says that there may be other ways to form memories, apart from stimulating dopamine-producing neurons. "This is an inherently very neat experiment, but further research is needed in some areas," he says. "They showed that activation of a small subset of neurons is sufficient to cause learning, but they didn’t show that these neurons are actually activated during normal learning."

He thinks that Miesenböck’s team should also have looked at long-term memories, which may form via separate biochemical pathways.

The next step is to identify the "upstream" cells that control the activity of these 12 neurons, says Miesenböck. He says this will "point like a finger" to the sites where the flies’ memories are physically stored.