A study of 9 to 13-year-old children in Finland has revealed that children’s eye movements during a virtual reality memory game can help identify those with attention-deficit hyperactivity disorder (ADHD). The pattern characteristics that allowed the differentiation were dispersed throughout the game, not associated with specific events or game features. The study was published in Scientific Reports.
ADHD is a neurodevelopmental disorder that primarily affects children but can persist into adulthood. It is characterized by persistent patterns of inattention, hyperactivity, and impulsivity that significantly interfere with daily functioning and well-being. Children and adults with ADHD struggle with tasks that require sustained attention, organization, and impulse control. This creates problems in school, at work, and in their everyday social interactions. Treatment options include behavioral therapies, education about the disorder, and in some cases, medication.
Although ADHD can be easily suspected by observing the natural behavior of children, particularly by people spending lots of time with them (e.g. school teachers), diagnostic procedures still mainly rely on formal psychometric assessment instruments administered in a formal psychodiagnostic settings. Due to this, ADHD is not likely to be detected unless events happen that make a child’s caregiver(s) seek help for the child from a professional.
Study author Liya Merzon and her colleagues wanted to explore whether it would be possible to detect children with ADHD in a more naturalistic setting. They have recently developed a virtual reality task that allows them to assess attention and executive function deficits in VR conditions that resemble those where ADHD typically manifests. The task is called Executive Performance in Everyday Living or EPELI. In this study, they used eye-tracking to try to differentiate between children with ADHD and those without this disorder based on patterns of their eye movements while they perform EPELI tasks in VR.
The study included 37 children diagnosed with ADHD (29 of whom were boys) and 36 children without the disorder (21 boys). All participants were aged between 9 and 13. While the two groups were age-matched, children without ADHD generally hailed from more affluent backgrounds and showcased superior cognitive ability scores (WISC-IV Similarities and Matrix reasoning) on average.
Each participant completed a set of 70 tasks in the EPELI virtual environment. The tasks start with an animated dragon character giving instructions regarding the tasks to be done. The child has a maximum of 90 second for each task. The scenario finishes when the child completes all the tasks or when the time limit for a task expires. The total duration of the procedure was 25-35 minutes. The VR headset tracked eye movements of the child using it.
Findings indicated that the ADHD group scored lower on the EPELI tasks. They demonstrated decreased task efficiency, struggled with virtual navigation, were more active with their controllers, and took more actions than their counterparts. Furthermore, ADHD children exhibited shorter and less intense saccades – quick, voluntary eye shifts between points of interest. Children with ADHD also had longer eye fixation durations (periods when eyes look at one point) compared to the control group.
To analyze the eye movement data, the researchers employed a Support Vector Machine (SVM) classifier, a machine learning algorithm ideal for classification tasks. This classifier aims to maximize differences between groups based on available data. Using eye fixation duration, saccade duration, and amplitude as inputs, the procedure successfully classified 84% of ADHD children and 78% from the control group.
“To conclude, our study showed that eye movements recorded in naturalistic setting provide a promising behavioral marker for ADHD assessment, which can be used to predict the diagnosis with excellent accuracy. Demonstrating the performance of VR eye tracking for differential diagnostics and reproducibility of the present results, however, needs further research,” the study authors concluded.
The study makes an important contribution to developing novel ways to diagnose ADHD. However, it should be noted that the sample of participants was small, and the two groups differed on a number of characteristics. Results on a larger sample and with groups that are better matched might not be the same.
The study, “Eye movement behavior in a real‑world virtual reality task reveals ADHD in children”, was authored by Liya Merzon, Kati Pettersson, EevaT.Aronen, Hanna Huhdanpää, Erik Seesjärvi, Linda Henriksson, W. Joseph MacInnes, Minna Mannerkoski, Emiliano Macaluso, and Juha Salmi.
