Voice technology 'could help detect autism'

A toddler chatting on a mobile
Image caption Children with autism have a very different vocal development pattern

Young children with autism can be identified by listening to the noises they make, say US scientists.

Research suggests the babbling of infants with autism differs from that of children without it. The differences were spotted with 86% accuracy using automated vocal analysis technology.

Vocal characteristics are not currently used for diagnosing autism, even though the link has been suggested before.

The study is in the journal Proceedings of the National Academy of Sciences.

Autism is the name given to a group, or "spectrum", of lifelong developmental conditions characterised by an inability to communicate with or relate to others, a lack of social skills, obsessional traits, and repetitive behaviour.

An estimated 500,000 people in the UK are believed to be affected by autism.

'Child utterances'

The US scientists analysed nearly 1,500 day-long vocal soundtracks from battery-powered recorders attached to the clothing of 232 children aged between 10 months and 4 years.

In total more than three million individual child utterances were used in the research, the study notes.

The study focused on 12 specific sound parameters associated with vocal development.

The most important were those involving "syllabification" - the ability of children to produce well-formed syllables with rapid movements of the jaw and tongue.

Experts believe these sounds form the foundation of words.

In autistic children up to four years old, there was a mismatch between the expected parameter values and age.

Professor Steven Warren, an expert in autism spectrum disorders at the University of Kansas, US, who took part in the study, said: "This technology could help paediatricians screen children for ASD (autism spectrum disorder) 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 new system, called Lena (Language Environment Analysis) could make a big difference to the screening, assessment and treatment of autism, say researchers.

Speech patterns

They point out that since the analysis is based on sound patterns rather than words, it could be used to screen speakers of any language for signs of autism.

"The physics of human speech are the same in all people as far as we know," said Prof Warren.

Dr Gina Gomez de la Cuesta, action research leader at The National Autistic Society, said: "Any tools which could help to identify speech and language difficulties at a younger age have the potential to help families, when used with professional guidance.

"However, they are no substitute for proper assessment by experienced and well-trained professionals."

Dr de la Cuesta added: "The diagnosis of autism is based on a range of behavioural features, not just language development.

"The social aspects of communication must also be considered, and it should be remembered that every child is different and develops at their own pace."

More on this story

Around the BBC

Related Internet links

The BBC is not responsible for the content of external Internet sites