Facial morphometric features in autism spectrum disorder: Preliminary findings from canonical discriminant analysis

Abstract

Autism spectrum disorder (ASD) is a neurodevelopmental condition characterized by impairments in social communication and the presence of restricted, repetitive behaviors. Although there is no cure for ASD, early diagnosis and evidence-based interventions can significantly improve developmental outcomes. However, many children are diagnosed later than recommended, limiting timely access to appropriate support services. This proof-of-concept study examines whether facial morphometric characteristics, analyzed through canonical discriminant analysis (CDA), can differentiate children with ASD from their typically developing (TD) peers. The study included 40 children diagnosed with ASD and 40 age- and gender-matched TD controls. Standardized facial photographs were taken in the Frankfurt Horizontal plane in accordance with biometric photography guidelines. Anthropometric landmarks were identified, and inter-landmark distances were measured using the ImageJ software. CDA was then performed in SPSS 28.0 to develop a statistical classification model. CDA was conducted to differentiate ASD and TD groups based on facial morphometric features. While overall facial morphology alone did not significantly distinguish the groups, specific regions—particularly the eyes and lips—showed significant discriminatory power. The nasal profile demonstrated moderate differentiation, and the strongest separation was achieved when combining overall facial and organ-specific features, with a canonical correlation of 0.74 and a significant Wilks’ Lambda (Λ = 0.453, χ²(8) = 58.651, p 〈 0.001). The present findings suggest that specific facial regions, particularly the eyes and lips, may carry morphometric features that significantly differentiate children with ASD from their TD peers. While overall facial morphology alone did not provide sufficient discrimination, combining overall facial and organ-specific measurements improved group separation (canonical correlation = 0.74). These results should be regarded as preliminary, highlighting the potential of facial morphometrics as a supplementary, non-invasive research tool. External validation with larger, ethnically diverse samples remains essential before any clinical or screening applicability can be considered.