Quantifying individual differences in brain morphometry underlying symptom severity in Autism Spectrum Disorders.

Area of Science:

• Neuroscience
• Developmental Neuroscience
• Neuroimaging

Background:

• Autism spectrum disorder (ASD) is a heterogeneous neurodevelopmental disorder with unclear neurobiological underpinnings.
• High individual variance in phenotypic expression complicates research findings and reproducibility.

Purpose of the Study:

• To develop and implement a novel subject-level, distance-based method to quantify individual differences in brain morphometry in ASD.
• To identify specific neuroanatomical features that predict symptom severity in individuals with ASD.

Main Methods:

• A large, multi-cohort sample of 100 individuals with ASD and 100 matched controls was used.
• A novel subject-level, distance-based method analyzed MRI brain morphometry and Social Responsiveness Scale (SRS) scores.
• Regularized machine learning with feature selection, cross-validation, and permutation testing was employed.

Main Results:

• Subject-specific structural morphometry features significantly predicted individual variation in ASD symptom severity.
• Nineteen cortical thickness features (p=0.01) and 10 surface area features (p=0.006) were identified.
• Findings were robust across subjects and replicated in validation samples.

Conclusions:

• Subject-level neuroanatomical markers, particularly in the salience and default mode networks, are associated with social impairments in ASD.
• This approach highlights the importance of individual differences in understanding the neurobiology of heterogeneous disorders like ASD.
• The study provides a step forward in identifying neurobiological features relevant to ASD heterogeneity.