Project Description
Challenge
Children often wait a long time for an FASD assessment, which means it takes longer for them to access supports. A team led by Dr. James Reynolds (Queen’s University) is working to develop an assessment tool based on eye movements that could screen children for FASD so that interventions can be accessed sooner.
Project Summary
A device called the Eyelink 1000 tracks a child’s eye movement behaviours while they watch a series of video clips. This data is stored and used to train a machine-learning classification algorithm (known as a classifier). In other words, the algorithm learns to differentiate between children with FASD and those without the disorder based on data collected from the eye movement tracker.
Initially, the classifier had an overall accuracy of 79%, meaning it was able to accurately identify children with FASD 79 percent of the time based on three structured eye movement tasks. The team also tested out the classifier by combining data from another project initiated in 2017, which collected eye movement data simultaneously with electroencephalography (EEG) data in adults with FASD and controls. EEG is a neuroimaging technique that measures electrical activity in the brain and can highlight differences between individuals with FASD and those without the disorder.
By combining eye movement data with EEG data, the classifier became much more accurate—it was able to correctly identify individuals with FASD 88% of the time, compared to 79% of the time with just eye movement data.
Result
The end goal of the project is to fine-tune the classifier so that it can accurately differentiate children with FASD from typically developing children with greater than 90% accuracy. The team also plans to test the screening tool in a community setting to show that it’s widely applicable and portable. While the initial focus is on FASD, validating the tool will open the door to improve screening for all neurodevelopmental disabilities.
Funding
Partners – $400,000
Team
Principal Investigators
James N. Reynolds, Queen’s University
Tim Oberlander, University of British Columbia
Christine Loock, University of British Columbia
Partners
Previous Cycle I Initiatives
Eye Tracking Device to Aid in Early FASD Screening
