Detecting Autism Spectrum Disorder Using Networks
Autism spectrum disorder (ASD) is a developmental disorder that affects how a person might interact with other people. The article emphasizes one particular effect of ASD: it affects how, in a group setting, a person looks at other people. A person with ASD tends to look more at the mouths of people that they interact with, rather than the eyes.
In a study, researchers have been beginning to use images of people to identify how a participant “gazes” at the image. A participant who tends to look at the mouth or under-eye area of the person is more likely to have ASD or perhaps a separate developmental issue. They study also highlights how a participant examines the image over time.
Researchers are beginning to study the gazes of multiple people in a group setting. They are able to identify how exactly each participant gazes at the other participants. Graph theory, as we discussed in class, is very relevant to this study. Eye-to-eye gazes are represented by a weighted graph: each person is represented by a node, the gazes are represented by edges. The weights on each edge vary depending on the number of times the two people make eye-to-eye contact. A person with ASD would on average have fewer connections to other people, with lower weights on each of these connections (or edges). Their local network would be significantly smaller than that of a person without ASD.
In the following graph, one person clearly has fewer, lower-weight connections–they make less eye-to-eye contact than anybody else in the graph. Graphs like these can help diagnose ASD.
