Fractionating autism based on neuroanatomical normative modeling.

Area of Science:

• Neuroscience
• Developmental Neuroscience
• Neuroimaging

Background:

• Autism spectrum disorder (ASD) is a complex neurodevelopmental condition characterized by significant heterogeneity in its presentation, biology, and causes.
• Identifying reliable biomarkers to stratify individuals with ASD into distinct cognitive or biological subtypes remains a significant challenge.
• This heterogeneity can lead to inconsistent findings in case-control studies and difficulties in understanding the underlying neurobiology.

Purpose of the Study:

• To introduce a novel methodological framework for identifying neuroanatomical subtypes within a large cohort of individuals with ASD.
• To leverage neuroimaging data, specifically cortical thickness (CT), to parse the complex neurobiology of autism.
• To explore whether identified subtypes correlate with clinical symptoms and genetic risk factors.

Main Methods:

• Utilized normative modeling of cortical thickness (CT) and spectral clustering on a large, multi-center dataset from the Longitudinal European Autism Project (LEAP).
• Analyzed data from 316 participants with ASD and 206 neurotypical controls, aged 6-31 years, across six European sites.
• Derived five biologically based putative subtypes based on CT patterns.

Main Results:

• Identified five distinct neuroanatomical subtypes within the autism cohort: three with decreased CT and two with increased CT.
• These subtypes exhibited significant morphometric differences from each other.
• The subtypes differentially loaded onto clinical symptoms and polygenic risk scores, suggesting a dilution of effects in heterogeneous samples.

Conclusions:

• The derived neuroanatomical subtypes offer a potential explanation for inconsistent findings in autism research.
• This approach represents a significant step towards dissecting the heterogeneous neurobiology of autism spectrum disorder.
• The findings support the utility of neuroimaging-based subtyping for a more nuanced understanding of autism.