Scientists have shown how a single protein may trigger autistic spectrum disorders by stopping effective communication between brain cells.
The team from Duke University in North Carolina created autistic mice by mutating the gene which controls production of the protein, Shank3.
The animals exhibited social problems, and repetitive behaviour - both classic signs of autism and related conditions.
The Nature study raises hopes of the first effective drug treatments.
Autism is a disorder which, to varying degrees, affects the ability of children and adults to communicate and interact socially.
While hundreds of genes linked to the condition have been found, the precise combination of genetics, biochemistry and other environmental factors which produce autism is still unclear.
Each patient has only one or a handful of those mutations, making it difficult to develop drugs to treat the disorder.
Shank3 is found in the synapses - the junctions between brain cells (neurons) that allow them to communicate with each other.
The researchers created mice which had a mutated form of Shank3, and found that these animals avoided social interactions with other mice.
They also engaged in repetitious and self-injurious grooming behaviour.
When the MIT team analysed the animals' brains they found defects in the circuits that connect two different areas of the brain, the cortex and the striatum.
Healthy connections between these areas are thought to be key to effective regulation of social behaviours and social interaction.
The researchers say their work underscores just what an important role Shank3 plays in the establishment of circuits in the brain which underlie all our behaviours.
Lead researcher Dr Guoping Feng said: "Our study demonstrated that Shank3 mutation in mice lead to defects in neuron-neuron communications.
"These findings and the mouse model now allow us to figure out the precise neural circuit defects responsible for these abnormal behaviours, which could lead to novel strategies and targets for developing treatment."
It is thought that only a small percentage of people with autism have mutations in Shank3, but Dr Feng believes many other cases may be linked to disruptions to other proteins that control synaptic function.
If true he believes it should be possible to develop treatments that restore synaptic function, regardless of which protein is defective in a specific individual.
Carol Povey, director of the National Autistic Society's Centre for Autism, said: "Animal research can help advance our understanding or the role of genetics and their influence on behaviour, however it is only a small part of the picture when it comes to understanding autism.
"Human brains are far more complex than those of other mammals, and it is believed that a variety of factors are responsible for the development of the condition."