Some relatives of people with autism also display behaviours and brain differences associated with the condition, even though they themselves do not have it. This could make it easier to spot families at risk of having an autistic child. It could also help in the quest to identify the genetic and environmental triggers for the condition, though it seems these triggers might vary from country to country.

Eric Peterson of the University of Colorado in Denver had compared an MRI study of the brains of 40 parents with autistic children to that of 40 age-matched controls. And he told the Society for Neuroscience annual meeting in Washington DC that the parents who had an autistic child shared several differences in brain structure with their offspring.

Looking at the group averages, the differences in parents of the autistic children included an unexpected increase in the size of the motor cortex and basal ganglia, areas important for movement planning and imitation. The somatosensory cortex, neighbouring the motor cortex, by contrast, was smaller than average. This region is important for understanding social information such as facial expressions - one key skill that autistic people often lack. These parents also had reductions in the cerebellum, important for coordinating movement, and in a frontal region thought to be responsible for understanding the intentions and feelings of others - the so-called theory of mind area.

In another study, Brendon Nacewicz of the University of Wisconsin-Madison Medical School and colleagues tested whether brothers of autistic children would avoid eye contact with others, a common feature of autism. While the parents seemed normal in this respect, brothers avoided eye contact just as strongly as their autistic sibling. He is now planning to test sisters too. Nacewicz also showed that the amygdala, a region important for processing emotions, particularly fear, was shrunken in the brothers too, just as it is in autistic people.

One theory laid to rest by these findings, says Nacewicz, is the idea that autism somehow falls on the far end of a shyness spectrum. The siblings showed no signs of autism or shyness, despite avoiding eye contact with others. Although gaze avoidance is accompanied by differences in the biology of the brain, he says, other brain areas must somehow compensate for the differences. Peterson agrees. This suggests that several core brain differences have to be present for someone to show the symptoms of autism, he says.

A further complexity in the underlying biology of autism was reported by Antonio Persico from the University of Rome. He found certain genetic variations linked with autism in North America were not present in autistic families in Italy. It is possible that there are regional differences in the environmental factors that interact with different genes to trigger autism, he suggests.

The differences were in a gene that makes an enzyme called paraoxonase. In North American populations, families with autistic members seemed to share a variant of the gene that makes a less active form of the enzyme. In Italian families with autism, however, that variant was no more common than in families without. One job of the enzyme is to inactivate organophosphates, which are often used in American homes as insecticides. In Italy, they are rarely found in the home. So one possible explanation, claims Persico, is that Americans with the less active enzyme use more of it to clear the pesticide, leaving less free to do another important job in helping neurons migrate to their right places during brain development.

In Italian families, the interaction between environment and genes may be different. For instance, an environmental effect may interact with the gene reelin, which also guides neurons to the right place, and has been linked to autism. "The important thing is that I am not trying to prove that organophosphates are bad," says Persico. These children may become sick anyway. The point is that there may be different genetic risk factors and different environmental interactions which lead to autism, he adds.

Dan Geschwind, an autism expert from the University of California, Los Angeles, who chaired the session on brain differences, is yet to be convinced about the claims of environmental triggers, however. "There is no evidence for one," he says, "but we can't rule it out."