CAITLIN O'NEILL is a happy, gurgling baby, an adored third child to her parents and the centre of her brother Gabriel's world.
She is also one of the most closely monitored infants in Australia, as researchers attempt to define a screening test that could point, even before birth, to a likelihood of autism.
Its developer says such a test would raise ethical challenges if it prompted parents to terminate possibly affected foetuses.
But for Caitlin's mother, Julia, it represents a chance to give children born with the condition, which affects Gabriel, 6, a head start.
''If they could [diagnose autism prenatally], I would want to know that,'' said Mrs O'Neill, of Perth. ''I'd want to be prepared to help her as soon as she's in the world.''
Throughout her pregnancy, Mrs O'Neill undertook blood tests for hormone and chemical levels and had three scans with an ultrasound machine adapted to record structures within Caitlin's developing brain. Umbilical cord blood drawn at birth may also yield clues to how the condition arises.
Mrs O'Neill said the intensive early intervention therapies Gabriel began after his diagnosis, at age three, ''delivered such tremendous results'' but ideally would have begun even earlier.
Andrew Whitehouse, who heads the autism and related disorders program at the Telethon Institute of Child Health Research in Perth, said there would probably never be a cast-iron diagnostic test during pregnancy for autism like those available for Down syndrome. Instead, a test would only point to an elevated risk of autism, but some people might still use it to decide whether to continue a pregnancy.
''These are factors we need to start discussing out in the open,'' said Professor Whitehouse, who will present the continuing project at an Autism Queensland symposium at Bond University on Thursday. ''It's a broader debate that we'd like to see happen.''
There was ''a critical period between eight and 24 weeks of pregnancy when we think brain development goes off course'', he said.
The world-first foetal brain measurements - in 100 pregnancies where there was an older autistic child and 100 controls where siblings developed normally - could be used to track retrospectively any link between growth patterns and a later autism diagnosis.
Autism predominantly affects boys, and higher testosterone levels in some pregnancies could be ''shaping the brain to grow in a very masculine way'', Professor Whitehouse said. The study would also monitor pesticide and food packaging chemicals that could alter hormone function.
Professor Whitehouse said he had ''started the trial with some hesitancy. There's a huge amount of emotion when people are having another child after a child with autism.'' But participants had responded positively.
Caitlin is too young for formal assessment but ''my gut tells me she's not autistic'', said Mrs O'Neill, whose four-year-old daughter, Alice, does not have the condition.
''We're not anti-autism,'' she said. ''Gabriel's such a sweet soul. He takes such joy in what he loves and he doesn't filter it.''