Autism is a complex neurobehavioral condition that includes impairments in social interaction and developmental language and communication skills combined with rigid, repetitive behaviors. Because of the range of symptoms, this condition is now called autism spectrum disorder (ASD). It covers a large spectrum of symptoms, skills, and levels of impairment. ASD ranges in severity from a handicap that somewhat limits an otherwise normal life to a devastating disability that may require institutional care.
Children with autism have trouble communicating. They have trouble understanding what other people think and feel. This makes it very hard for them to express themselves either with words or through gestures, facial expressions, and touch.
We, a team of Computer Science Students came across a research paper published by Stephen Sheinkopf who said “A major challenge in research in autism is how we can identify autism in early infancy,”
“We know that older children with autistic disorders often produce sounds in odd and unusual ways … we wanted to measure this phenomenon in infants in the context of crying,” he said.
Sheinkopf and his fellow researchers examined the naturally occurring cries of about 40 six-month-old infants, representing two subject groups. One group was identified as high-risk for Autism Spectrum Disorder because the subjects had an older sibling with the disease. The second group was identified as low-risk for autism because the subjects had no siblings who had the disorder, Sheinkopf said.
Recordings of the cries were made using a camcorder and microphone embedded in a vest worn by each infant at home, according to the study. Once recorded, the cries were categorized into pain-related and non-pain-related cries. The pitch of the cries was identified using a program specifically designed for the study.
When the researchers compared the pitches of the pain-related cries from both groups they found that the cries of the high-risk infants on average had a statistically significant higher pitch than the cries of low-risk infants, Sheinkopf said. In fact, the three infants in the high-risk group who did develop autism were the subjects with the three highest-pitched cries, he said. No statistically significant difference in acoustic features was found for the non-pain-related cries, according to the study.
“Atypical cry acoustics … may serve as a positive symptom that, when combined with other indices of risk, may have additional value as a means to increase the accuracy of early identification efforts,” the study states.
One of the study’s major strengths was its “careful and systemic approach while using naturalistic observation,” said Kelley Powell ’06, post-doctoral psychology fellow at the Yale University Child Study Center, who was not involved in the study.
Most cry acoustic studies evoke cries systematically to gather data, Powell said. The experimental design of this study, in which cries were observed “naturalistically,” was a new approach to collecting cry acoustic data, Powell said.
She recommended the researchers reevaluate the infants when they are 36 months old to see if the infants with Autism Spectrum Disorder still have a higher-pitched cry.
Identifying autism early in an infant’s development would benefit the autism field of research by allowing specialists to initiate intervention measures earlier, Sheinkopf said.
Unfortunately there is no way we can detect autism in children using the above mentioned cry analysis technique today. All we can do is some MRI scans, which again is not safe for infants. They can do short or long term damage to the infant's mental health
We created a normal and more subtle way of doing the cry analysis.
It is a culmination of 4 projects under the Allisto Organization.
- Allistic Server - The Backend Service & REST API
- Allisto's App - The Cross Platform Mobile App for Autism Detection
- Artz Allisto - The Doctor Panel which is linked to the Allistic Server
- Cry Analysis NN - The ANN code along with the dataset
We make all of these four work together in harmony to record audio of children, send them for analysis to the server api's which does the analysis using the model weights of the cry analysis which shows the results to the doctor on his panel. Making treatment way easier than before.
“Every life must be given the chance to realize its full potential – that every life matters.”
We believe we can cure. We believe you can help us do it better.
Made with ❤️ by Team Allisto