Eye-Tracking Reveals Distinct Visual Attention Patterns in Children with Autism: Systematic Review Finds

The systematic review was published in the journal Medical Sciences in January 2026.

Autism Spectrum Disorder (ASD) is a complex and highly heterogeneous neurodevelopmental condition characterized by persistent social communication deficits and restricted, repetitive patterns of behavior. To overcome clinical barriers arising from this symptom heterogeneity and the specialized expertise required for traditional evaluations, Marcella Di Cara and fellow researchers from the University of Messina aimed to systematize the currently unorganized connections between specific eye-movement markers and clinical features to evaluate eye-tracking as a sensitive and specific diagnostic indicator.

The review was registered in PROSPERO (CRD42023429316) and conducted in accordance with PRISMA guidelines. A comprehensive literature search of PubMed, Web of Science, and Science Direct through March 2025 identified 14 peer-reviewed studies comparing children and adolescents under 18 years with ASD to TD controls. Key eye-tracking metrics extracted included fixation duration and count, spatial distribution of gaze, and saccadic parameters.

Risk of bias was assessed using the QUADAS-2 tool, which revealed substantial heterogeneity across index tests and patient selection methods.

Key Clinical Findings of the Review Include:

• Social Engagement Deficits: The review confirms that children with ASD exhibit significantly reduced fixation on faces and eyes compared to TD peers, with eye fixation serving as a powerful predictor of group membership (Cohen’s d = 1.56).
• Non-Social Preference: It highlights that a preference for dynamic geometric images over social scenes—specifically a fixation threshold exceeding 68.6%—provides a 100% Positive Predictive Value (PPV) for identifying ASD.
• Diagnostic Performance: It reports a high pooled specificity of 92.3% and AUC values ranging from 0.71 to 0.92, confirming eye-tracking as a highly reliable tool for ruling out ASD despite a lower pooled sensitivity of 35.9%.
• Atypical Saccadic Activity: It indicates that autistic children with geometric preferences exhibit significantly fewer saccades per second when viewing preferred non-social stimuli but increased saccade rates when viewing social images compared to typical patterns.
• Spatial Gaze Disorganization: It shows that ASD participants exhibit significantly higher Distance-to-Reference (D2R) values and lower spatial cohesion, reflecting gaze patterns that are more peripheral and less synchronized with social focal points than those of TD controls.

These findings confirm that objective oculomotor markers can effectively capture distinct social attention deficits, supporting more precise phenotypic stratification for children with ASD.

Integrating eye-tracking technology into clinical practice provides an objective and non-invasive means to supplement traditional evaluations by helping clinicians identify and categorize unique visual attention patterns in children with autism spectrum disorder.

The authors suggest that factors such as small sample sizes and methodological variability across diverse age groups may influence the generalizability of the findings, highlighting the need for future research with standardized protocols to support the clinical integration of eye-tracking for ASD.

Reference

Di Cara, M., De Domenico, C., Piccolo, A., Alito, A., Costa, L., Quartarone, A., & Cucinotta, F. (2026). The Diagnostic Potential of Eye Tracking to Detect Autism Spectrum Disorder in Children: A Systematic Review. Medical Sciences, 14(1), 28.