JMIR Publications

ABSTRACT

Background:

Attention-Deficit/Hyperactivity Disorder (ADHD) is a prevalent neurodevelopmental condition characterized by inattention, hyperactivity, impulsivity, and emotional dysregulation. Beyond these core symptoms, children with ADHD frequently exhibit executive function deficits, including working memory, inhibitory control, cognitive flexibility, and processing speed, which compromise the ability to regulate attention, plan and organize tasks, and modulate emotional responses. These cognitive and affective difficulties are interrelated and contribute to academic, social, and emotional challenges, highlighting the need for assessment approaches that integrate multiple dimensions of child functioning.

Objective:

To develop and evaluate an automated, multimodal screening approach capable of identifying individualized cognitive–emotional–attentional profiles in children, supporting earlier detection and personalized interventions.

Methods:

In this cross-sectional study, 549 children aged 6–15 years were assessed using a multimodal approach integrating neuropsychological tests (two- and three-symbol cancellation tasks from the Coimbra Neuropsychological Assessment Battery; Trail Making Test Parts A and B), eye-tracking metrics, and automated facial emotion recognition during Rapid Naming and Emotion Recognition tasks. A machine learning pipeline, combining XGBoost and Random Forest classifiers, analyzed these data to differentiate children with attention deficits from typically developing peers. SHAP analyses identified the most influential features for model predictions.

Results:

Children with attention deficits exhibited heightened emotional reactivity, unstable and fragmented visual attention, impulsive eye movements, and frequent neutral facial expressions during emotionally salient tasks. Anger and Fear stimuli were particularly discriminative between groups, indicating that affective processes modulate attentional performance. Typically developing children demonstrated stable fixation patterns, consistent saccadic control, and higher accuracy in emotion recognition, supporting the validity of multimodal markers for differentiating attentional and emotional profiles.

Conclusions:

This study demonstrates the feasibility and utility of integrating eye-tracking, emotion recognition, and neuropsychological assessment within an interpretable machine-learning framework for early ADHD screening. The approach provides objective, real-time insights into attentional, cognitive, and emotional functioning, offering a scalable tool to guide personalized interventions in educational and clinical settings. The findings underscore the multidimensional nature of ADHD and emphasize the importance of simultaneously considering cognitive, attentional, and affective processes for a comprehensive assessment.