London: Autism could be detected in children by analyzing their voice, says a new study.

Toddlers with the developmental disorder pronounce words differently to their healthy peers which can be picked up by a new automated vocal analysis system created by scientists.

The device called LENA (Language Environment Analysis) could lead to the screening for autism spectrum disorders (ASD) for which early intervention is important.

It works by recording a child's speech for a whole day and then feeding the data into a special computer programme that compares the noises with those of other youngsters already known to have the condition, reports the Telegraph.

The researchers said early speech of infants with autism - particularly the way they pronounce the syllables of words - are distinct from those of typically developing children.

The system which costs about 130 pounds picked up those with the condition with 86 percent accuracy, according to the findings published online in Proceedings of the National Academy of Sciences.

It also differentiated normal children and those with autism from children with language delay based on the automated vocal analysis.

Early diagnosis and treatment of autism can have a dramatic effect on the development of children.

The problem is that it is hard to detect and by the time it is usually detected a lot of damage is already done.

Steven Warren, professor at Kansas University, said, "This technology could help pediatricians screen children for ASD to determine if a referral to a specialist for a full diagnosis is required and get those children into earlier and more effective treatments."

The researchers analyzed 1,486 recordings from 232 children aged between 10 months and four years - more than 3.1 million identified utterances. IANS