AUTISM IS NOT A SPECTRUM, IT'S A TYPE.

AUTISM IS NOT A SPECTRUM, IT’S A TYPE
New Princeton research reveals 4 distinct subtypes of autism.
Each has different causes, brain patterns, and traits.
Using machine learning on data from 5,000+ autistic children (ages 4–18), researchers identified four robust subtypes, each with its own patterns of traits, genetics, and developmental timing.

1. Social & Behavioural Challenges (~37%)

  • Classic autism traits (social communication difficulties, repetitive behaviours).

  • Developmental milestones (walking/talking) usually on time.

  • High rates of co-occurring ADHD, anxiety, OCD, depression.

  • Stronger link to common genetic variants (e.g., ADHD/depression) and genes active after birth.

2. Mixed ASD with Developmental Delay (~19%)

  • Delayed developmental milestones (motor and speech), but generally fewer psychiatric issues.

  • Mainly linked to rare inherited mutations, often present from birth.

3. Moderate Challenges (~34%)

  • Milder autism traits overall.

  • Milestones reached on time.

  • Few co-occurring conditions; lower support needs.

  • Genetic profile shows modest rare variant burden in less critical genes.

4. Broadly Affected (~10%)

  • Severe across multiple domains: delayed milestones, communication, social skills, mood disorders.

  • Highest burden of de novo mutations (new mutations not inherited) in developmental genes.

What This Means

  • These subtypes have distinct underlying biology and genetic signatures, not just behavioral differences.

  • The timing of gene-expression effects may vary: some mutations act prenatally, others later in childhood, aligning with clinical features.

  • This reframes autism as multiple “autisms”, rather than a single continuum. It helps explain why past genetic studies struggled: they combined heterogeneous cases that were actually biologically distinct.

Researchers aim to:

  • Expand to adult populations, females (who are often underrepresented), and cross-cultural samples

  • Develop biomarkers or screening tools that classify autism subtypes automatically

  • Explore if subtype-specific treatments (pharmacological or behavioral) have better outcomes

  • Study how gene expression timing maps to neurodevelopmental windows

Autism may not be “one thing” with different severities but many things with different causes.

2 Likes

Well articulated!

Very well written, it gives an insight of the condition making us well aware with the patterns seen in autism.

This is a powerful shift in how we understand autism. Identifying biologically distinct subtypes helps move beyond broad labels toward more precise diagnosis and personalized support. It also explains why one-size-fits-all interventions often fall short.

This topic needs to reach more people. I still see so many confusion among people when it comes to autism.

With help of this knowledge better targeted support customized according to the type of autism could be provided.