• Subscribe

Facetime: Catching autism early

Speed read
  • Detection of ASD has historically relied on behavioral markers
  • New research using algorithms and 3D modeling may have unlocked early identification
  • Long term psychological and social health for those with ASD hinge on early intervention

What if you could look into a baby’s face and see if they were likely to develop autism spectrum disorder (ASD)? You’d first be very sad, but then cautiously optimistic because early intervention can lead to behavioral therapies that create a significant difference in the child’s long term health.

That’s the hope motivating Diana Weiting Tan, researcher in the school of psychological science and Syed Zulqarnain Gilani from the school of computer science at the University of Western Australia (UWA) in Perth. Their research focuses on the link between ASD and facial anomalies.

<strong>Face to face. </strong> Researchers from UWA harnessed algorithms to calculate, first, gender-typical facial structures. This metric was then weighed against faces of children with ASD. Analysis points to masculine facial formation as a possible biological marker for ASD. Courtesy Tan; Gilani, et al.

Since the brain and the face develop at the same time during pregnancy, scientists have long been interested in the link between facial features and atypical brain development. Promising research lines indicate that masculinized features correlate with ASD in young children.

Seeing the success 3D photogrammetry has in yielding precise, high resolution facial analyses, Tan teamed up with Gilani to see whether facial recognition algorithms might reveal something about ASD.

“A cross-disciplinary research collaboration is extremely essential in increasing our knowledge about complex and multi-faceted conditions like autism,” says Tan.

The team set about their project by first identifying which features differentiated the faces of boys and girls.

Measuring the faces of 215 children without ASD, Tan’s team annotated some key points on each 3D face through the 3dMDface system.

From 26 linear distances and 26 geodesic distances measured across 21 distinct facial landmarks, the Gender Efficient Feature Selection algorithm settled on 11 distances that differentiated boys and girls.

Taking six of these features, they compared them to the faces of 74 children with ASD to see if either more or less masculine features correlated with ASD.

Analysis of the comparison found a positive correlation in increased facial masculinity observed in the overall facial structure and in individual features of autistic children.

“These research findings have suggested to us that we can use facial features to further our understanding of the biological origins of autism,” says Tan.

<strong>Tan and Gilani </strong> employed the 3dMDface system which produces ear-to-ear facial images captured from stereo camera viewpoints. This technology has helped locate a correlation between masculinized faces and occurrence of ASD in children. Courtesy 3dMD.

This breakthrough is exciting because until now parents and researchers have had to wait on behavioral markers to indicate ASD was present, and this typically meant waiting until a child was three years old.

But by employing the power advanced computation, 3D photogrammetry, and high-performance algorithms, Tan’s team has found a way to detect ASD much earlier.

“Currently there are no biological markers for autism and the outcomes of our research suggests possible facial biological markers which could aid in earlier diagnoses of autism. This could lead to early intervention which has been found to improve long term psychological and social outcomes for individuals with autism.”

Autism can be a devastating disorder, but hope for early detection and possible remedy has been located – by looking into the face of a child.


Join the conversation

Do you have story ideas or something to contribute? Let us know!

Copyright © 2022 Science Node ™  |  Privacy Notice  |  Sitemap

Disclaimer: While Science Node ™ does its best to provide complete and up-to-date information, it does not warrant that the information is error-free and disclaims all liability with respect to results from the use of the information.


We encourage you to republish this article online and in print, it’s free under our creative commons attribution license, but please follow some simple guidelines:
  1. You have to credit our authors.
  2. You have to credit ScienceNode.org — where possible include our logo with a link back to the original article.
  3. You can simply run the first few lines of the article and then add: “Read the full article on ScienceNode.org” containing a link back to the original article.
  4. The easiest way to get the article on your site is to embed the code below.