Investigating the iPhone application for young children, which reveals signs of autism, has shown that it is easy to use, is welcomed by caretakers and works well with reliable scientific data.

The study, described on June 1 in the journal of open access to Digital Medicine, points the way to a wider, easier access to the diagnosis of autism and other neurodevelopmental disorders.

The application first manages the consent forms for care and polling, and then uses the phone's "selfie" camera to collect video reactions of young children's reactions while they watch films designed to identify the risk of autism, for example, emotions and attention on the device screen .

The child's video reactions are sent to the research servers, where the automatic behavioral coding software tracks the movement of video landmarks on the child's face and quantifies the emotions and attention of the child. For example, in response to a short video of bubbles floating around the screen, the video encoding algorithm looks for the movements of the person that will point to joy.

In this study, children whose parents rated their child as having a large number of autism symptoms showed less joyful emotions in response to the blisters.

Prevention of autism in young children is currently carried out in a clinical setting, rather than in the child's natural environment, and people with a high degree of preparation are required to learn and analyze the results. "It's not scalable," says Helen Egger of New York University, one of the co-leaders of the study.

This study, based on informed consent on data collection and preliminary analysis, was conducted using an application available free of charge from the Apple Store and based on Apple ResearchKit's open source development platform.

During the year, more than 10,000 downloads of the application were downloaded, and the study involved 1,756 families with children aged one to six years. Parents completed 5,618 polls and downloaded 4,441 videos. Useful data was collected on 88 percent of downloaded videos, demonstrating for the first time the possibility of using this type of tool for observation and coding in natural environments.

"This demonstrates the feasibility of this approach," says Geraldine Dawson, Ph.D., director of the Center for Autism and Brain Development and co-chair of the study. "Many educators were ready to participate, the data was of high quality, and the video analysis algorithms produced results consistent with the results we produce in our autism program."

The application-based approach can significantly improve in areas with insufficient protection and greatly facilitate tracking of changes in each child over time, said Guillermo Sapiro, Edmund T. Pratt, Jr. Professor of the Department of Electrical and Computer Engineering Duke and one of the co-chairs of the study .

"This technology has the potential to transform how we screen and control the development of children," said Sapiro.

The project reported consisted of a 12-month study. The whole test took about 20 minutes, and only a few minutes with the child's participation.

The application also included a widespread questionnaire that screens the autism spectrum disorder. On the basis of the questionnaire, participating families received some feedback from the appendix on what could be an obvious risk of autism in a child. If parents reported a high level of signs of autism in the questionnaire, they were asked to seek further medical advice.


Materials provided by Duke University.

Journal Reference:

Helen L. Egger, Geraldine Dawson, Jordan Hashemi, Kimberly L. Kh. Carpenter, Stephen Espinosa, Kathleen Campbell, Samuel Brotkin, Jana Shaich-Borg, Qian Qiu, Mariano Tepper, Jeffrey P. Baker, Richard A. Bloomfield , Guillermo Sapiro. Automatic analysis of emotions and attention in young children at home: a feasibility study of Autism ResearchKit. npj Digital Medicine, 2018; 1 (1) DOI: 10.1038 / s41746-018-0024-6