Burcu Uslu, İrem Aslan, Fatma Aleyna Çevik
Nutritional variables do not directly or indirectly predict language development in ASD when controlling for core symptom severity. We propose a divergent "V-Type" conceptual model, suggesting that autism severity independently drives both language delays and eating disruptions. Nutritional interventions and speech-language therapy should be pursued as parallel, independent targets.
PURPOSE: To investigate whether eating behaviors and diet quality predict receptive and expressive language development in children with autism spectrum disorder (ASD), independent of autism severity.
METHODS: Seventy-one children with ASD (mean age 55.61 ± 15.88 months) were assessed using the TEIıL (language), GARS-2-TV (autism severity), CEBQ (eating behaviors), and KIDMED (diet quality). Analyses included hierarchical regression, 5000-bootstrap mediation, moderation, and exploratory K-means cluster analysis.
RESULTS: Autism severity strongly predicted both receptive and expressive language (explained variance = 53%). After controlling for severity and age, eating behaviors and diet quality provided no significant incremental variance ( $ΔR² = .020 and .012) and showed no significant partial correlations with language (all p > .15). Furthermore, all 11 mediation and 6 moderation models yielded non-significant effects. Exploratory clustering revealed a paradox: children with the most severe autism and highest food selectivity exhibited the highest Mediterranean diet adherence, likely reflecting rigid parental dietary control.
CONCLUSION: Nutritional variables do not directly or indirectly predict language development in ASD when controlling for core symptom severity. We propose a divergent "V-Type" conceptual model, suggesting that autism severity independently drives both language delays and eating disruptions. Nutritional interventions and speech-language therapy should be pursued as parallel, independent targets.