Vocal-analysis software could help spot autism

1 min read

A new automated vocal-analysis system could help screen for autism, according to researchers at the University of Memphis.

The researchers, led by Prof D Kimbrough Oller, demonstrated that the system could automatically analyse recorded sounds from infants and children to predict, with 86 per cent accuracy, which of them might have autism.

To reach that conclusion, they analysed 1,486 recordings from 232 children using an algorithm based on the 12 acoustic parameters associated with vocal development. The most important of these parameters proved to be the ones targeting syllabification − the ability of children to produce well-formed syllables with rapid movements of the jaw and tongue. Infants show voluntary control of syllabification and voice in the first months of life, and refine this skill as they acquire language.

While typically developing children and those with language delays showed statistically significant development of the parameters, the austistic children did not.

Although aberrations in the speech (or lack of it) of children with autism-spectrum disorders has been examined by researchers and clinicians for more than 20 years, vocal characteristics are not included in standard criteria for diagnosis of autism-spectrum disorders (ASDs), said Steven F Warren, professor of applied behavioural science at the University of Kansas, who contributed to the research.

’A small number of studies had previously suggested that children with autism have a markedly different vocal signature but, until now, we have been held back from using this knowledge in clinical applications by the lack of measurement technology,’ said Warren.

Warren predicts that the system − known as LENA (Language ENvironment Analysis) − could significantly impact the screening, assessment and treatment of autism and the behavioural sciences in general.

Since the analysis is not based on words, but rather on sound patterns, the system, theoretically, could potentially be used to screen speakers of any language for ASDs, Warren said. ’The physics of human speech are the same in all people, as far as we know.’

Warren said that children with ASDs can be diagnosed at 18 months, but that the median age of diagnosis is 5.7 years in the US.

’This technology could help pediatricians screen children for ASD to determine if a referral to a specialist for a full diagnosis is required and to get those children into earlier and more effective treatments.’