Research from Japan has demonstrated the use of eye-tracking technology in the early diagnosis of autism spectrum disorder (ASD), potentially offering a new approach to identify the condition in very young children.
Led by Waseda University's Associate Professor Mikimasa Omori, the study explored whether children with potential ASD would exhibit a preference for predictable movement - a behaviour indicative of neurodevelopmental disorders - longer than typically developing children. For the research, Associate Professor Omori developed six pairs of 10-second videos showing predictable and unpredictable geometric shapes. Each video pair was shown side-by-side in a preferential-looking paradigm to compare how study participants observe them. These observations were captured and analysed using an eye tracker system developed by the Sweden-based company Tobii.
Findings, published in the Nature journal Scientific Reports, showed that children with possible autism "spent significantly more time observing predictable movements," suggesting that they may develop this behaviour over time. According to the researcher, this repetitive behavior characteristic of autism "may be linked to difficulties in learning cause-and-effect relationships between movement trajectories and the anticipation of complete shapes." "Unlike the typically developing children, who did not show a shift in their observation patterns, children with potential ASD demonstrated a gradual increase in their focus on the predictable movements as the stimulus presentation progressed," Associate Professor Omori said.
Until this research, the reasons behind children with autism spending more time observing repetitive movements and how this behaviour evolves were unclear. Previous research has primarily focused on social communication deficits, such as eye contact and language delays. This study demonstrates how predictable movement stimuli can be potentially used as a behavioural marker for early ASD screening, with suggestions that this behavioral indicator could identify autism "as young as three years old" and potentially earlier.
Associate Professor Omori recommends implementing brief video assessment tasks during standard developmental evaluations for children between 18-36 months, with potential adaptations for even younger populations.
This Japanese study adds to growing global efforts to advance ASD diagnosis through technology. In the United States, Georgia-based EarliTec Diagnostics received FDA 510(k) clearance for a device that also utilizes eye-tracking technology to support ASD diagnosis by measuring children's focus and responsiveness while watching short videos .California-based Cognoa also received the US FDA's de novo clearance for its artificial intelligence software that analyses videos of children's behaviour to aid in autism diagnosis. In Australia, researchers at the University of Southern Queensland are developing a cloud-based system that can automatically detect autism from a single brain scan, while South Korean institutions, including Seoul National University Hospital and the National Center for Mental Health, have established a living laboratory to observe children and gather data to discover biomarkers and develop AI models for early autism diagnosis.
Click here for the original news story.