Cart

  • A new study has discovered a new technique that ensures autism spectrum disorder in kids is accurately and quickly detected by doctors.
  • Previously discovered and used autism detection techniques aren’t child-friendly. This study aimed to develop an effective and efficient method for kids.
  • The researchers during the study evaluated 23- neurotypical kids and 17 children with ASD.
  • The most effective approach in identifying ASD eye-gaze sequence is betweenness centrality
  • The newly discovered technique can be used for all accurate ASD detections, but it is specifically efficient for kids.

Quick Summary

This study was carried out by researchers from the University of Waterloo and was recently published in the journal Computers in Biology and Medicine. The researchers analyzed how kids diagnosed with ASD usually scan individuals face differently from a neuro-typical kid. Utilizing their findings, the study team were able to develop a new technique that considers how kids with ASD gaze transitions from one section of an individual face to another.

Goals

This study aimed to characterize the temporary and spatial attributes eye gaze of a kid with autism. Another goal of this research is to identify an appropriate approach (analytic) that can more efficiently unveil the variations between the face scanning patterns of TD and ASD children.

Previous techniques which are currently utilized by doctors for determining autism in kids aren’t child-friendly, this new development ensures that kids are quickly, effectively, and efficiently diagnosed with less error possibility.

Investigation

While developing the new method, the study evaluated 17 kids with ASD and 23 neuro-typical kids. The average chronological ages of these kids grouped in neuro-typical and ASD were 5.5 and 4.8 respectively.

Each kid was shown 44 pictures of faces on a 19-inch screen which was linked to an eye-tracking device. This device identified and interpreted the areas on stimuli at which each kid was gazing through reflection and emission of a wave from the iris.

The pictures were separated into seven areas of interest (AOI) in which each kid-focused his stare: the left eye, the right eye, beneath the left eye, beneath the right eye, mouth, nose and other areas of the screen.

The researchers from the University of Waterloo aimed to understand not just the average time each kid stared at each key AOI, but also, the pattern of their eye movement to scan the faces. To actualize that information, the study group utilized four different concepts from network analysis to determine the varying degree of importance each child placed on the seven key AOIs during the exploration of the facial features.

The initial concept evaluated the number of AOIs that the child directly stared to and from a particular area of interest. The second concept determined how often a specific AOI is involved when the kid moved their eyes between two other different AOIs quickly. The third concept relates to how quick each kid can navigate their eyes from one AOI to other AOIs. The last idea which is the fourth evaluates the importance of an AOI, in the aspects of face scanning and eye movement by the number of other important AOIs which shares direct transitions with it.

Result

The result of the study is summarized in these three points:

  • ASD kids spend more time staring at the mouth than TD kids.
  • ASD individuals have lower saccades than ASD individuals.
  • The most effective approach in identifying ASD eye-gaze sequence is betweenness centrality.

The new technique determines not just if the child is focused on the eyes or mouth, it is about in general, how the child sees everything.

Limitations

  • The study was carried out using only kids in a locality, extensive study needs to be carried out in various localities.

Source

https://doi.org/10.1016/j.compbiomed.2019.103332

https://www.sciencedirect.com/science/article/pii/S001048251930201X

Author Details
Information coming soon!
×
Information coming soon!

Author Conductscience

More posts by Conductscience