AUTHOR=Chen Xiaolu , Wang Sihan , Yang Xiaowen , Yu Chunmei , Ni Fang , Yang Jie , Tian Yu , Ye Jiucai , Liu Hao , Luo Rong 

TITLE=Utilizing artificial intelligence-based eye tracking technology for screening ADHD symptoms in children

JOURNAL=Frontiers in Psychiatry

VOLUME=Volume 14 - 2023

YEAR=2023

URL=https://www.frontiersin.org/journals/psychiatry/articles/10.3389/fpsyt.2023.1260031

DOI=10.3389/fpsyt.2023.1260031

ISSN=1664-0640

ABSTRACT=<sec><title>Objective</title><p>To explore the potential of using artificial intelligence (AI)-based eye tracking technology on a tablet for screening Attention-deficit/hyperactivity disorder (ADHD) symptoms in children.</p></sec><sec><title>Methods</title><p>We recruited 112 children diagnosed with ADHD (ADHD group; mean age: 9.40 ± 1.70 years old) and 325 typically developing children (TD group; mean age: 9.45 ± 1.59 years old). We designed a data-driven end-to-end convolutional neural network appearance-based model to predict eye gaze to permit eye-tracking under low resolution and sampling rates. The participants then completed the eye tracking task on a tablet, which consisted of a simple fixation task as well as 14 prosaccade (looking toward target) and 14 antisaccade (looking away from target) trials, measuring attention and inhibition, respectively.</p></sec><sec><title>Results</title><p>Two-way MANOVA analyses demonstrated that diagnosis and age had significant effects on performance on the fixation task [diagnosis: <italic>F</italic><sub>(2, 432)</sub> = 8.231, <sup>***</sup><italic>p</italic> &lt; 0.001; Wilks’ Λ = 0.963; age: <italic>F</italic><sub>(2, 432)</sub> = 3.999, *<italic>p</italic> &lt; 0.019; Wilks’ Λ = 0.982], prosaccade task [age: <italic>F</italic><sub>(16, 418)</sub> = 3.847, <sup>***</sup><italic>p</italic> &lt; 0.001; Wilks’ Λ = 0.872], and antisaccade task [diagnosis: <italic>F</italic><sub>(16, 418)</sub> = 1.738, *<italic>p</italic> = 0.038; Wilks’ Λ = 0.938; age: <italic>F</italic><sub>(16, 418)</sub> = 4.508, <sup>***</sup><italic>p</italic> &lt; 0.001; Wilks’ Λ = 0.853]. Correlational analyses revealed that participants with higher SNAP-IV score were more likely to have shorter fixation duration and more fixation intervals (<italic>r</italic> = −0.160, 95% CI [0.250, 0.067], <sup>***</sup><italic>p</italic> &lt; 0.001), poorer scores on adjusted prosaccade accuracy, and poorer scores on antisaccade accuracy (Accuracy: <italic>r</italic> = −0.105, 95% CI [−0.197, −0.011], *<italic>p</italic> = 0.029; Adjusted accuracy: <italic>r</italic> = −0.108, 95% CI [−0.200, −0.015], *<italic>p</italic> = 0.024).</p></sec><sec><title>Conclusion</title><p>Our AI-based eye tracking technology implemented on a tablet could reliably discriminate eye movements of the TD group and the ADHD group, providing a potential solution for ADHD screening outside of clinical settings.</p></sec>